objectDetection.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "id": "146BB11JpfDA"
   },
   "outputs": [],
   "source": [
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "id": "42hJEdo_pfDB"
   },
   "outputs": [],
   "source": [
    "CUSTOM_MODEL_NAME = 'my_ssd_mobnet' \n",
    "PRETRAINED_MODEL_NAME = 'ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8'\n",
    "PRETRAINED_MODEL_URL = 'http://download.tensorflow.org/models/object_detection/tf2/20200711/ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8.tar.gz'\n",
    "TF_RECORD_SCRIPT_NAME = 'generate_tfrecord.py'\n",
    "LABEL_MAP_NAME = 'label_map.pbtxt'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "id": "hbPhYVy_pfDB"
   },
   "outputs": [],
   "source": [
    "paths = {\n",
    "    'WORKSPACE_PATH': os.path.join('Tensorflow', 'workspace'),\n",
    "    'SCRIPTS_PATH': os.path.join('Tensorflow','scripts'),\n",
    "    'APIMODEL_PATH': os.path.join('Tensorflow','models'),\n",
    "    'ANNOTATION_PATH': os.path.join('Tensorflow', 'workspace','annotations'),\n",
    "    'IMAGE_PATH': os.path.join('Tensorflow', 'workspace','images'),\n",
    "    'MODEL_PATH': os.path.join('Tensorflow', 'workspace','models'),\n",
    "    'PRETRAINED_MODEL_PATH': os.path.join('Tensorflow', 'workspace','pre-trained-models'),\n",
    "    'CHECKPOINT_PATH': os.path.join('Tensorflow', 'workspace','models',CUSTOM_MODEL_NAME), \n",
    "    'OUTPUT_PATH': os.path.join('Tensorflow', 'workspace','models',CUSTOM_MODEL_NAME, 'export'),  \n",
    "    'TFLITE_PATH':os.path.join('Tensorflow', 'workspace','models',CUSTOM_MODEL_NAME, 'tfliteexport'), \n",
    "    'PROTOC_PATH':os.path.join('Tensorflow','protoc')\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "id": "LwhWZMI0pfDC"
   },
   "outputs": [],
   "source": [
    "files = {\n",
    "    'PIPELINE_CONFIG':os.path.join('Tensorflow', 'workspace','models', CUSTOM_MODEL_NAME, 'pipeline.config'),\n",
    "    'TF_RECORD_SCRIPT': os.path.join(paths['SCRIPTS_PATH'], TF_RECORD_SCRIPT_NAME), \n",
    "    'LABELMAP': os.path.join(paths['ANNOTATION_PATH'], LABEL_MAP_NAME)\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "id": "HR-TfDGrpfDC"
   },
   "outputs": [],
   "source": [
    "for path in paths.values():\n",
    "    if not os.path.exists(path):\n",
    "        if os.name == 'posix':\n",
    "            !mkdir -p {path}\n",
    "        if os.name == 'nt':\n",
    "            !mkdir {path}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "OLU-rs_ipfDE"
   },
   "source": [
    "Download Tensorflow Object Detection API"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "K-Cmz2edpfDE",
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "import wget"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "iA1DIq5OpfDE"
   },
   "outputs": [],
   "source": [
    "if not os.path.exists(os.path.join(paths['APIMODEL_PATH'], 'research', 'object_detection')):\n",
    "    !git clone https://github.com/tensorflow/models {paths['APIMODEL_PATH']}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "rJjMHbnDs3Tv"
   },
   "outputs": [],
   "source": [
    "# Install Tensorflow Object Detection \n",
    "if os.name=='posix':  \n",
    "    #!pip3 install protobuf==3.19.6\n",
    "    !cd Tensorflow/models/research && protoc object_detection/protos/*.proto --python_out=. && cp object_detection/packages/tf2/setup.py . && python3 -m pip install . \n",
    "    \n",
    "if os.name=='nt':\n",
    "    url=\"https://github.com/protocolbuffers/protobuf/releases/download/v3.15.6/protoc-3.15.6-win64.zip\"\n",
    "    wget.download(url)\n",
    "    !move protoc-3.15.6-win64.zip {paths['PROTOC_PATH']}\n",
    "    !cd {paths['PROTOC_PATH']} && tar -xf protoc-3.15.6-win64.zip\n",
    "    os.environ['PATH'] += os.pathsep + os.path.abspath(os.path.join(paths['PROTOC_PATH'], 'bin'))   \n",
    "    !cd Tensorflow/models/research && protoc object_detection/protos/*.proto --python_out=. && copy object_detection\\\\packages\\\\tf2\\\\setup.py setup.py && python setup.py build && python setup.py install\n",
    "    !cd Tensorflow/models/research/slim && pip install -e . "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2022-10-21 13:57:04.547729: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "Running tests under Python 3.8.8: /Users/ritvikseth/opt/anaconda3/envs/asl_new/bin/python3\n",
      "[ RUN      ] ModelBuilderTF2Test.test_create_center_net_deepmac\n",
      "2022-10-21 13:57:13.082268: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "/Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/object_detection/builders/model_builder.py:1102: DeprecationWarning: The 'warn' function is deprecated, use 'warning' instead\n",
      "  logging.warn(('Building experimental DeepMAC meta-arch.'\n",
      "W1021 13:57:13.428624 4547945984 model_builder.py:1102] Building experimental DeepMAC meta-arch. Some features may be omitted.\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_create_center_net_deepmac): 0.65s\n",
      "I1021 13:57:13.721164 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_create_center_net_deepmac): 0.65s\n",
      "[       OK ] ModelBuilderTF2Test.test_create_center_net_deepmac\n",
      "[ RUN      ] ModelBuilderTF2Test.test_create_center_net_model0 (customize_head_params=True)\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_create_center_net_model0 (customize_head_params=True)): 1.17s\n",
      "I1021 13:57:14.890915 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_create_center_net_model0 (customize_head_params=True)): 1.17s\n",
      "[       OK ] ModelBuilderTF2Test.test_create_center_net_model0 (customize_head_params=True)\n",
      "[ RUN      ] ModelBuilderTF2Test.test_create_center_net_model1 (customize_head_params=False)\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_create_center_net_model1 (customize_head_params=False)): 0.31s\n",
      "I1021 13:57:15.206486 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_create_center_net_model1 (customize_head_params=False)): 0.31s\n",
      "[       OK ] ModelBuilderTF2Test.test_create_center_net_model1 (customize_head_params=False)\n",
      "[ RUN      ] ModelBuilderTF2Test.test_create_center_net_model_from_keypoints\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_create_center_net_model_from_keypoints): 0.26s\n",
      "I1021 13:57:15.469478 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_create_center_net_model_from_keypoints): 0.26s\n",
      "[       OK ] ModelBuilderTF2Test.test_create_center_net_model_from_keypoints\n",
      "[ RUN      ] ModelBuilderTF2Test.test_create_center_net_model_mobilenet\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_create_center_net_model_mobilenet): 3.35s\n",
      "I1021 13:57:18.821704 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_create_center_net_model_mobilenet): 3.35s\n",
      "[       OK ] ModelBuilderTF2Test.test_create_center_net_model_mobilenet\n",
      "[ RUN      ] ModelBuilderTF2Test.test_create_experimental_model\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_create_experimental_model): 0.0s\n",
      "I1021 13:57:18.826543 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_create_experimental_model): 0.0s\n",
      "[       OK ] ModelBuilderTF2Test.test_create_experimental_model\n",
      "[ RUN      ] ModelBuilderTF2Test.test_create_faster_rcnn_from_config_with_crop_feature0 (True)\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_create_faster_rcnn_from_config_with_crop_feature0 (True)): 0.03s\n",
      "I1021 13:57:18.854384 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_create_faster_rcnn_from_config_with_crop_feature0 (True)): 0.03s\n",
      "[       OK ] ModelBuilderTF2Test.test_create_faster_rcnn_from_config_with_crop_feature0 (True)\n",
      "[ RUN      ] ModelBuilderTF2Test.test_create_faster_rcnn_from_config_with_crop_feature1 (False)\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_create_faster_rcnn_from_config_with_crop_feature1 (False)): 0.02s\n",
      "I1021 13:57:18.876270 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_create_faster_rcnn_from_config_with_crop_feature1 (False)): 0.02s\n",
      "[       OK ] ModelBuilderTF2Test.test_create_faster_rcnn_from_config_with_crop_feature1 (False)\n",
      "[ RUN      ] ModelBuilderTF2Test.test_create_faster_rcnn_model_from_config_with_example_miner\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_create_faster_rcnn_model_from_config_with_example_miner): 0.02s\n",
      "I1021 13:57:18.897636 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_create_faster_rcnn_model_from_config_with_example_miner): 0.02s\n",
      "[       OK ] ModelBuilderTF2Test.test_create_faster_rcnn_model_from_config_with_example_miner\n",
      "[ RUN      ] ModelBuilderTF2Test.test_create_faster_rcnn_models_from_config_faster_rcnn_with_matmul\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_create_faster_rcnn_models_from_config_faster_rcnn_with_matmul): 0.15s\n",
      "I1021 13:57:19.044718 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_create_faster_rcnn_models_from_config_faster_rcnn_with_matmul): 0.15s\n",
      "[       OK ] ModelBuilderTF2Test.test_create_faster_rcnn_models_from_config_faster_rcnn_with_matmul\n",
      "[ RUN      ] ModelBuilderTF2Test.test_create_faster_rcnn_models_from_config_faster_rcnn_without_matmul\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_create_faster_rcnn_models_from_config_faster_rcnn_without_matmul): 0.31s\n",
      "I1021 13:57:19.358613 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_create_faster_rcnn_models_from_config_faster_rcnn_without_matmul): 0.31s\n",
      "[       OK ] ModelBuilderTF2Test.test_create_faster_rcnn_models_from_config_faster_rcnn_without_matmul\n",
      "[ RUN      ] ModelBuilderTF2Test.test_create_faster_rcnn_models_from_config_mask_rcnn_with_matmul\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_create_faster_rcnn_models_from_config_mask_rcnn_with_matmul): 0.12s\n",
      "I1021 13:57:19.479732 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_create_faster_rcnn_models_from_config_mask_rcnn_with_matmul): 0.12s\n",
      "[       OK ] ModelBuilderTF2Test.test_create_faster_rcnn_models_from_config_mask_rcnn_with_matmul\n",
      "[ RUN      ] ModelBuilderTF2Test.test_create_faster_rcnn_models_from_config_mask_rcnn_without_matmul\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_create_faster_rcnn_models_from_config_mask_rcnn_without_matmul): 0.12s\n",
      "I1021 13:57:19.602823 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_create_faster_rcnn_models_from_config_mask_rcnn_without_matmul): 0.12s\n",
      "[       OK ] ModelBuilderTF2Test.test_create_faster_rcnn_models_from_config_mask_rcnn_without_matmul\n",
      "[ RUN      ] ModelBuilderTF2Test.test_create_rfcn_model_from_config\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_create_rfcn_model_from_config): 0.12s\n",
      "I1021 13:57:19.726526 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_create_rfcn_model_from_config): 0.12s\n",
      "[       OK ] ModelBuilderTF2Test.test_create_rfcn_model_from_config\n",
      "[ RUN      ] ModelBuilderTF2Test.test_create_ssd_fpn_model_from_config\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_create_ssd_fpn_model_from_config): 0.03s\n",
      "I1021 13:57:19.758937 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_create_ssd_fpn_model_from_config): 0.03s\n",
      "[       OK ] ModelBuilderTF2Test.test_create_ssd_fpn_model_from_config\n",
      "[ RUN      ] ModelBuilderTF2Test.test_create_ssd_models_from_config\n",
      "I1021 13:57:19.986828 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:145] EfficientDet EfficientNet backbone version: efficientnet-b0\n",
      "I1021 13:57:19.986964 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:147] EfficientDet BiFPN num filters: 64\n",
      "I1021 13:57:19.987029 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:148] EfficientDet BiFPN num iterations: 3\n",
      "I1021 13:57:19.989258 4547945984 efficientnet_model.py:143] round_filter input=32 output=32\n",
      "I1021 13:57:20.021765 4547945984 efficientnet_model.py:143] round_filter input=32 output=32\n",
      "I1021 13:57:20.021895 4547945984 efficientnet_model.py:143] round_filter input=16 output=16\n",
      "I1021 13:57:20.121143 4547945984 efficientnet_model.py:143] round_filter input=16 output=16\n",
      "I1021 13:57:20.121345 4547945984 efficientnet_model.py:143] round_filter input=24 output=24\n",
      "I1021 13:57:20.342381 4547945984 efficientnet_model.py:143] round_filter input=24 output=24\n",
      "I1021 13:57:20.342511 4547945984 efficientnet_model.py:143] round_filter input=40 output=40\n",
      "I1021 13:57:20.548709 4547945984 efficientnet_model.py:143] round_filter input=40 output=40\n",
      "I1021 13:57:20.548837 4547945984 efficientnet_model.py:143] round_filter input=80 output=80\n",
      "I1021 13:57:20.847000 4547945984 efficientnet_model.py:143] round_filter input=80 output=80\n",
      "I1021 13:57:20.847147 4547945984 efficientnet_model.py:143] round_filter input=112 output=112\n",
      "I1021 13:57:21.192826 4547945984 efficientnet_model.py:143] round_filter input=112 output=112\n",
      "I1021 13:57:21.192970 4547945984 efficientnet_model.py:143] round_filter input=192 output=192\n",
      "I1021 13:57:21.623941 4547945984 efficientnet_model.py:143] round_filter input=192 output=192\n",
      "I1021 13:57:21.624078 4547945984 efficientnet_model.py:143] round_filter input=320 output=320\n",
      "I1021 13:57:21.729537 4547945984 efficientnet_model.py:143] round_filter input=1280 output=1280\n",
      "I1021 13:57:21.796684 4547945984 efficientnet_model.py:453] Building model efficientnet with params ModelConfig(width_coefficient=1.0, depth_coefficient=1.0, resolution=224, dropout_rate=0.2, blocks=(BlockConfig(input_filters=32, output_filters=16, kernel_size=3, num_repeat=1, expand_ratio=1, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=16, output_filters=24, kernel_size=3, num_repeat=2, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=24, output_filters=40, kernel_size=5, num_repeat=2, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=40, output_filters=80, kernel_size=3, num_repeat=3, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=80, output_filters=112, kernel_size=5, num_repeat=3, expand_ratio=6, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=112, output_filters=192, kernel_size=5, num_repeat=4, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=192, output_filters=320, kernel_size=3, num_repeat=1, expand_ratio=6, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise')), stem_base_filters=32, top_base_filters=1280, activation='simple_swish', batch_norm='default', bn_momentum=0.99, bn_epsilon=0.001, weight_decay=5e-06, drop_connect_rate=0.2, depth_divisor=8, min_depth=None, use_se=True, input_channels=3, num_classes=1000, model_name='efficientnet', rescale_input=False, data_format='channels_last', dtype='float32')\n",
      "I1021 13:57:21.872364 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:145] EfficientDet EfficientNet backbone version: efficientnet-b1\n",
      "I1021 13:57:21.872570 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:147] EfficientDet BiFPN num filters: 88\n",
      "I1021 13:57:21.872669 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:148] EfficientDet BiFPN num iterations: 4\n",
      "I1021 13:57:21.874343 4547945984 efficientnet_model.py:143] round_filter input=32 output=32\n",
      "I1021 13:57:21.893571 4547945984 efficientnet_model.py:143] round_filter input=32 output=32\n",
      "I1021 13:57:21.893690 4547945984 efficientnet_model.py:143] round_filter input=16 output=16\n",
      "I1021 13:57:22.072382 4547945984 efficientnet_model.py:143] round_filter input=16 output=16\n",
      "I1021 13:57:22.072507 4547945984 efficientnet_model.py:143] round_filter input=24 output=24\n",
      "I1021 13:57:22.354449 4547945984 efficientnet_model.py:143] round_filter input=24 output=24\n",
      "I1021 13:57:22.354614 4547945984 efficientnet_model.py:143] round_filter input=40 output=40\n",
      "I1021 13:57:22.653064 4547945984 efficientnet_model.py:143] round_filter input=40 output=40\n",
      "I1021 13:57:22.653234 4547945984 efficientnet_model.py:143] round_filter input=80 output=80\n",
      "I1021 13:57:23.025012 4547945984 efficientnet_model.py:143] round_filter input=80 output=80\n",
      "I1021 13:57:23.025144 4547945984 efficientnet_model.py:143] round_filter input=112 output=112\n",
      "I1021 13:57:23.399492 4547945984 efficientnet_model.py:143] round_filter input=112 output=112\n",
      "I1021 13:57:23.399621 4547945984 efficientnet_model.py:143] round_filter input=192 output=192\n",
      "I1021 13:57:23.888134 4547945984 efficientnet_model.py:143] round_filter input=192 output=192\n",
      "I1021 13:57:23.888261 4547945984 efficientnet_model.py:143] round_filter input=320 output=320\n",
      "I1021 13:57:24.105880 4547945984 efficientnet_model.py:143] round_filter input=1280 output=1280\n",
      "I1021 13:57:24.148892 4547945984 efficientnet_model.py:453] Building model efficientnet with params ModelConfig(width_coefficient=1.0, depth_coefficient=1.1, resolution=240, dropout_rate=0.2, blocks=(BlockConfig(input_filters=32, output_filters=16, kernel_size=3, num_repeat=1, expand_ratio=1, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=16, output_filters=24, kernel_size=3, num_repeat=2, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=24, output_filters=40, kernel_size=5, num_repeat=2, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=40, output_filters=80, kernel_size=3, num_repeat=3, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=80, output_filters=112, kernel_size=5, num_repeat=3, expand_ratio=6, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=112, output_filters=192, kernel_size=5, num_repeat=4, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=192, output_filters=320, kernel_size=3, num_repeat=1, expand_ratio=6, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise')), stem_base_filters=32, top_base_filters=1280, activation='simple_swish', batch_norm='default', bn_momentum=0.99, bn_epsilon=0.001, weight_decay=5e-06, drop_connect_rate=0.2, depth_divisor=8, min_depth=None, use_se=True, input_channels=3, num_classes=1000, model_name='efficientnet', rescale_input=False, data_format='channels_last', dtype='float32')\n",
      "I1021 13:57:24.227426 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:145] EfficientDet EfficientNet backbone version: efficientnet-b2\n",
      "I1021 13:57:24.227590 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:147] EfficientDet BiFPN num filters: 112\n",
      "I1021 13:57:24.227678 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:148] EfficientDet BiFPN num iterations: 5\n",
      "I1021 13:57:24.229890 4547945984 efficientnet_model.py:143] round_filter input=32 output=32\n",
      "I1021 13:57:24.259669 4547945984 efficientnet_model.py:143] round_filter input=32 output=32\n",
      "I1021 13:57:24.259819 4547945984 efficientnet_model.py:143] round_filter input=16 output=16\n",
      "I1021 13:57:24.442595 4547945984 efficientnet_model.py:143] round_filter input=16 output=16\n",
      "I1021 13:57:24.442719 4547945984 efficientnet_model.py:143] round_filter input=24 output=24\n",
      "I1021 13:57:24.856902 4547945984 efficientnet_model.py:143] round_filter input=24 output=24\n",
      "I1021 13:57:24.857079 4547945984 efficientnet_model.py:143] round_filter input=40 output=48\n",
      "I1021 13:57:25.434522 4547945984 efficientnet_model.py:143] round_filter input=40 output=48\n",
      "I1021 13:57:25.434703 4547945984 efficientnet_model.py:143] round_filter input=80 output=88\n",
      "I1021 13:57:26.008592 4547945984 efficientnet_model.py:143] round_filter input=80 output=88\n",
      "I1021 13:57:26.008775 4547945984 efficientnet_model.py:143] round_filter input=112 output=120\n",
      "I1021 13:57:26.474539 4547945984 efficientnet_model.py:143] round_filter input=112 output=120\n",
      "I1021 13:57:26.474671 4547945984 efficientnet_model.py:143] round_filter input=192 output=208\n",
      "I1021 13:57:27.183834 4547945984 efficientnet_model.py:143] round_filter input=192 output=208\n",
      "I1021 13:57:27.183968 4547945984 efficientnet_model.py:143] round_filter input=320 output=352\n",
      "I1021 13:57:27.434592 4547945984 efficientnet_model.py:143] round_filter input=1280 output=1408\n",
      "I1021 13:57:27.494632 4547945984 efficientnet_model.py:453] Building model efficientnet with params ModelConfig(width_coefficient=1.1, depth_coefficient=1.2, resolution=260, dropout_rate=0.3, blocks=(BlockConfig(input_filters=32, output_filters=16, kernel_size=3, num_repeat=1, expand_ratio=1, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=16, output_filters=24, kernel_size=3, num_repeat=2, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=24, output_filters=40, kernel_size=5, num_repeat=2, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=40, output_filters=80, kernel_size=3, num_repeat=3, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=80, output_filters=112, kernel_size=5, num_repeat=3, expand_ratio=6, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=112, output_filters=192, kernel_size=5, num_repeat=4, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=192, output_filters=320, kernel_size=3, num_repeat=1, expand_ratio=6, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise')), stem_base_filters=32, top_base_filters=1280, activation='simple_swish', batch_norm='default', bn_momentum=0.99, bn_epsilon=0.001, weight_decay=5e-06, drop_connect_rate=0.2, depth_divisor=8, min_depth=None, use_se=True, input_channels=3, num_classes=1000, model_name='efficientnet', rescale_input=False, data_format='channels_last', dtype='float32')\n",
      "I1021 13:57:27.585879 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:145] EfficientDet EfficientNet backbone version: efficientnet-b3\n",
      "I1021 13:57:27.586087 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:147] EfficientDet BiFPN num filters: 160\n",
      "I1021 13:57:27.586158 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:148] EfficientDet BiFPN num iterations: 6\n",
      "I1021 13:57:27.588146 4547945984 efficientnet_model.py:143] round_filter input=32 output=40\n",
      "I1021 13:57:27.614512 4547945984 efficientnet_model.py:143] round_filter input=32 output=40\n",
      "I1021 13:57:27.614686 4547945984 efficientnet_model.py:143] round_filter input=16 output=24\n",
      "I1021 13:57:27.822858 4547945984 efficientnet_model.py:143] round_filter input=16 output=24\n",
      "I1021 13:57:27.822990 4547945984 efficientnet_model.py:143] round_filter input=24 output=32\n",
      "I1021 13:57:28.131569 4547945984 efficientnet_model.py:143] round_filter input=24 output=32\n",
      "I1021 13:57:28.131701 4547945984 efficientnet_model.py:143] round_filter input=40 output=48\n",
      "I1021 13:57:28.429314 4547945984 efficientnet_model.py:143] round_filter input=40 output=48\n",
      "I1021 13:57:28.429456 4547945984 efficientnet_model.py:143] round_filter input=80 output=96\n",
      "I1021 13:57:28.926888 4547945984 efficientnet_model.py:143] round_filter input=80 output=96\n",
      "I1021 13:57:28.927016 4547945984 efficientnet_model.py:143] round_filter input=112 output=136\n",
      "I1021 13:57:29.430958 4547945984 efficientnet_model.py:143] round_filter input=112 output=136\n",
      "I1021 13:57:29.431087 4547945984 efficientnet_model.py:143] round_filter input=192 output=232\n",
      "I1021 13:57:30.043659 4547945984 efficientnet_model.py:143] round_filter input=192 output=232\n",
      "I1021 13:57:30.043792 4547945984 efficientnet_model.py:143] round_filter input=320 output=384\n",
      "I1021 13:57:30.277719 4547945984 efficientnet_model.py:143] round_filter input=1280 output=1536\n",
      "I1021 13:57:30.331534 4547945984 efficientnet_model.py:453] Building model efficientnet with params ModelConfig(width_coefficient=1.2, depth_coefficient=1.4, resolution=300, dropout_rate=0.3, blocks=(BlockConfig(input_filters=32, output_filters=16, kernel_size=3, num_repeat=1, expand_ratio=1, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=16, output_filters=24, kernel_size=3, num_repeat=2, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=24, output_filters=40, kernel_size=5, num_repeat=2, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=40, output_filters=80, kernel_size=3, num_repeat=3, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=80, output_filters=112, kernel_size=5, num_repeat=3, expand_ratio=6, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=112, output_filters=192, kernel_size=5, num_repeat=4, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=192, output_filters=320, kernel_size=3, num_repeat=1, expand_ratio=6, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise')), stem_base_filters=32, top_base_filters=1280, activation='simple_swish', batch_norm='default', bn_momentum=0.99, bn_epsilon=0.001, weight_decay=5e-06, drop_connect_rate=0.2, depth_divisor=8, min_depth=None, use_se=True, input_channels=3, num_classes=1000, model_name='efficientnet', rescale_input=False, data_format='channels_last', dtype='float32')\n",
      "I1021 13:57:30.410257 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:145] EfficientDet EfficientNet backbone version: efficientnet-b4\n",
      "I1021 13:57:30.410467 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:147] EfficientDet BiFPN num filters: 224\n",
      "I1021 13:57:30.410571 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:148] EfficientDet BiFPN num iterations: 7\n",
      "I1021 13:57:30.413274 4547945984 efficientnet_model.py:143] round_filter input=32 output=48\n",
      "I1021 13:57:30.438714 4547945984 efficientnet_model.py:143] round_filter input=32 output=48\n",
      "I1021 13:57:30.438848 4547945984 efficientnet_model.py:143] round_filter input=16 output=24\n",
      "I1021 13:57:30.626980 4547945984 efficientnet_model.py:143] round_filter input=16 output=24\n",
      "I1021 13:57:30.627103 4547945984 efficientnet_model.py:143] round_filter input=24 output=32\n",
      "I1021 13:57:30.998416 4547945984 efficientnet_model.py:143] round_filter input=24 output=32\n",
      "I1021 13:57:30.998544 4547945984 efficientnet_model.py:143] round_filter input=40 output=56\n",
      "I1021 13:57:31.383427 4547945984 efficientnet_model.py:143] round_filter input=40 output=56\n",
      "I1021 13:57:31.383553 4547945984 efficientnet_model.py:143] round_filter input=80 output=112\n",
      "I1021 13:57:31.956628 4547945984 efficientnet_model.py:143] round_filter input=80 output=112\n",
      "I1021 13:57:31.956792 4547945984 efficientnet_model.py:143] round_filter input=112 output=160\n",
      "I1021 13:57:32.543909 4547945984 efficientnet_model.py:143] round_filter input=112 output=160\n",
      "I1021 13:57:32.544034 4547945984 efficientnet_model.py:143] round_filter input=192 output=272\n",
      "I1021 13:57:33.590694 4547945984 efficientnet_model.py:143] round_filter input=192 output=272\n",
      "I1021 13:57:33.590819 4547945984 efficientnet_model.py:143] round_filter input=320 output=448\n",
      "I1021 13:57:33.827800 4547945984 efficientnet_model.py:143] round_filter input=1280 output=1792\n",
      "I1021 13:57:33.883298 4547945984 efficientnet_model.py:453] Building model efficientnet with params ModelConfig(width_coefficient=1.4, depth_coefficient=1.8, resolution=380, dropout_rate=0.4, blocks=(BlockConfig(input_filters=32, output_filters=16, kernel_size=3, num_repeat=1, expand_ratio=1, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=16, output_filters=24, kernel_size=3, num_repeat=2, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=24, output_filters=40, kernel_size=5, num_repeat=2, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=40, output_filters=80, kernel_size=3, num_repeat=3, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=80, output_filters=112, kernel_size=5, num_repeat=3, expand_ratio=6, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=112, output_filters=192, kernel_size=5, num_repeat=4, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=192, output_filters=320, kernel_size=3, num_repeat=1, expand_ratio=6, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise')), stem_base_filters=32, top_base_filters=1280, activation='simple_swish', batch_norm='default', bn_momentum=0.99, bn_epsilon=0.001, weight_decay=5e-06, drop_connect_rate=0.2, depth_divisor=8, min_depth=None, use_se=True, input_channels=3, num_classes=1000, model_name='efficientnet', rescale_input=False, data_format='channels_last', dtype='float32')\n",
      "I1021 13:57:34.038342 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:145] EfficientDet EfficientNet backbone version: efficientnet-b5\n",
      "I1021 13:57:34.038513 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:147] EfficientDet BiFPN num filters: 288\n",
      "I1021 13:57:34.038593 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:148] EfficientDet BiFPN num iterations: 7\n",
      "I1021 13:57:34.040997 4547945984 efficientnet_model.py:143] round_filter input=32 output=48\n",
      "I1021 13:57:34.067549 4547945984 efficientnet_model.py:143] round_filter input=32 output=48\n",
      "I1021 13:57:34.067710 4547945984 efficientnet_model.py:143] round_filter input=16 output=24\n",
      "I1021 13:57:34.331973 4547945984 efficientnet_model.py:143] round_filter input=16 output=24\n",
      "I1021 13:57:34.332101 4547945984 efficientnet_model.py:143] round_filter input=24 output=40\n",
      "I1021 13:57:34.795357 4547945984 efficientnet_model.py:143] round_filter input=24 output=40\n",
      "I1021 13:57:34.795485 4547945984 efficientnet_model.py:143] round_filter input=40 output=64\n",
      "I1021 13:57:35.485760 4547945984 efficientnet_model.py:143] round_filter input=40 output=64\n",
      "I1021 13:57:35.485925 4547945984 efficientnet_model.py:143] round_filter input=80 output=128\n",
      "I1021 13:57:36.490247 4547945984 efficientnet_model.py:143] round_filter input=80 output=128\n",
      "I1021 13:57:36.490437 4547945984 efficientnet_model.py:143] round_filter input=112 output=176\n",
      "I1021 13:57:37.454149 4547945984 efficientnet_model.py:143] round_filter input=112 output=176\n",
      "I1021 13:57:37.454339 4547945984 efficientnet_model.py:143] round_filter input=192 output=304\n",
      "I1021 13:57:38.529929 4547945984 efficientnet_model.py:143] round_filter input=192 output=304\n",
      "I1021 13:57:38.530115 4547945984 efficientnet_model.py:143] round_filter input=320 output=512\n",
      "I1021 13:57:38.977264 4547945984 efficientnet_model.py:143] round_filter input=1280 output=2048\n",
      "I1021 13:57:39.050194 4547945984 efficientnet_model.py:453] Building model efficientnet with params ModelConfig(width_coefficient=1.6, depth_coefficient=2.2, resolution=456, dropout_rate=0.4, blocks=(BlockConfig(input_filters=32, output_filters=16, kernel_size=3, num_repeat=1, expand_ratio=1, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=16, output_filters=24, kernel_size=3, num_repeat=2, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=24, output_filters=40, kernel_size=5, num_repeat=2, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=40, output_filters=80, kernel_size=3, num_repeat=3, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=80, output_filters=112, kernel_size=5, num_repeat=3, expand_ratio=6, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=112, output_filters=192, kernel_size=5, num_repeat=4, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=192, output_filters=320, kernel_size=3, num_repeat=1, expand_ratio=6, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise')), stem_base_filters=32, top_base_filters=1280, activation='simple_swish', batch_norm='default', bn_momentum=0.99, bn_epsilon=0.001, weight_decay=5e-06, drop_connect_rate=0.2, depth_divisor=8, min_depth=None, use_se=True, input_channels=3, num_classes=1000, model_name='efficientnet', rescale_input=False, data_format='channels_last', dtype='float32')\n",
      "I1021 13:57:39.194155 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:145] EfficientDet EfficientNet backbone version: efficientnet-b6\n",
      "I1021 13:57:39.194298 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:147] EfficientDet BiFPN num filters: 384\n",
      "I1021 13:57:39.194356 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:148] EfficientDet BiFPN num iterations: 8\n",
      "I1021 13:57:39.195960 4547945984 efficientnet_model.py:143] round_filter input=32 output=56\n",
      "I1021 13:57:39.226665 4547945984 efficientnet_model.py:143] round_filter input=32 output=56\n",
      "I1021 13:57:39.226805 4547945984 efficientnet_model.py:143] round_filter input=16 output=32\n",
      "I1021 13:57:39.496409 4547945984 efficientnet_model.py:143] round_filter input=16 output=32\n",
      "I1021 13:57:39.496562 4547945984 efficientnet_model.py:143] round_filter input=24 output=40\n",
      "I1021 13:57:40.092427 4547945984 efficientnet_model.py:143] round_filter input=24 output=40\n",
      "I1021 13:57:40.092561 4547945984 efficientnet_model.py:143] round_filter input=40 output=72\n",
      "I1021 13:57:40.736837 4547945984 efficientnet_model.py:143] round_filter input=40 output=72\n",
      "I1021 13:57:40.736972 4547945984 efficientnet_model.py:143] round_filter input=80 output=144\n",
      "I1021 13:57:41.539566 4547945984 efficientnet_model.py:143] round_filter input=80 output=144\n",
      "I1021 13:57:41.539696 4547945984 efficientnet_model.py:143] round_filter input=112 output=200\n",
      "I1021 13:57:42.626106 4547945984 efficientnet_model.py:143] round_filter input=112 output=200\n",
      "I1021 13:57:42.626240 4547945984 efficientnet_model.py:143] round_filter input=192 output=344\n",
      "I1021 13:57:43.905694 4547945984 efficientnet_model.py:143] round_filter input=192 output=344\n",
      "I1021 13:57:43.905868 4547945984 efficientnet_model.py:143] round_filter input=320 output=576\n",
      "I1021 13:57:44.382702 4547945984 efficientnet_model.py:143] round_filter input=1280 output=2304\n",
      "I1021 13:57:44.445479 4547945984 efficientnet_model.py:453] Building model efficientnet with params ModelConfig(width_coefficient=1.8, depth_coefficient=2.6, resolution=528, dropout_rate=0.5, blocks=(BlockConfig(input_filters=32, output_filters=16, kernel_size=3, num_repeat=1, expand_ratio=1, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=16, output_filters=24, kernel_size=3, num_repeat=2, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=24, output_filters=40, kernel_size=5, num_repeat=2, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=40, output_filters=80, kernel_size=3, num_repeat=3, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=80, output_filters=112, kernel_size=5, num_repeat=3, expand_ratio=6, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=112, output_filters=192, kernel_size=5, num_repeat=4, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=192, output_filters=320, kernel_size=3, num_repeat=1, expand_ratio=6, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise')), stem_base_filters=32, top_base_filters=1280, activation='simple_swish', batch_norm='default', bn_momentum=0.99, bn_epsilon=0.001, weight_decay=5e-06, drop_connect_rate=0.2, depth_divisor=8, min_depth=None, use_se=True, input_channels=3, num_classes=1000, model_name='efficientnet', rescale_input=False, data_format='channels_last', dtype='float32')\n",
      "I1021 13:57:44.600422 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:145] EfficientDet EfficientNet backbone version: efficientnet-b7\n",
      "I1021 13:57:44.600554 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:147] EfficientDet BiFPN num filters: 384\n",
      "I1021 13:57:44.600615 4547945984 ssd_efficientnet_bifpn_feature_extractor.py:148] EfficientDet BiFPN num iterations: 8\n",
      "I1021 13:57:44.602903 4547945984 efficientnet_model.py:143] round_filter input=32 output=64\n",
      "I1021 13:57:44.628327 4547945984 efficientnet_model.py:143] round_filter input=32 output=64\n",
      "I1021 13:57:44.628489 4547945984 efficientnet_model.py:143] round_filter input=16 output=32\n",
      "I1021 13:57:44.967869 4547945984 efficientnet_model.py:143] round_filter input=16 output=32\n",
      "I1021 13:57:44.968012 4547945984 efficientnet_model.py:143] round_filter input=24 output=48\n",
      "I1021 13:57:45.685406 4547945984 efficientnet_model.py:143] round_filter input=24 output=48\n",
      "I1021 13:57:45.685542 4547945984 efficientnet_model.py:143] round_filter input=40 output=80\n",
      "I1021 13:57:46.421605 4547945984 efficientnet_model.py:143] round_filter input=40 output=80\n",
      "I1021 13:57:46.421733 4547945984 efficientnet_model.py:143] round_filter input=80 output=160\n",
      "I1021 13:57:47.868639 4547945984 efficientnet_model.py:143] round_filter input=80 output=160\n",
      "I1021 13:57:47.868826 4547945984 efficientnet_model.py:143] round_filter input=112 output=224\n",
      "I1021 13:57:49.304357 4547945984 efficientnet_model.py:143] round_filter input=112 output=224\n",
      "I1021 13:57:49.304549 4547945984 efficientnet_model.py:143] round_filter input=192 output=384\n",
      "I1021 13:57:51.197710 4547945984 efficientnet_model.py:143] round_filter input=192 output=384\n",
      "I1021 13:57:51.197862 4547945984 efficientnet_model.py:143] round_filter input=320 output=640\n",
      "I1021 13:57:51.838095 4547945984 efficientnet_model.py:143] round_filter input=1280 output=2560\n",
      "I1021 13:57:51.926253 4547945984 efficientnet_model.py:453] Building model efficientnet with params ModelConfig(width_coefficient=2.0, depth_coefficient=3.1, resolution=600, dropout_rate=0.5, blocks=(BlockConfig(input_filters=32, output_filters=16, kernel_size=3, num_repeat=1, expand_ratio=1, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=16, output_filters=24, kernel_size=3, num_repeat=2, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=24, output_filters=40, kernel_size=5, num_repeat=2, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=40, output_filters=80, kernel_size=3, num_repeat=3, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=80, output_filters=112, kernel_size=5, num_repeat=3, expand_ratio=6, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=112, output_filters=192, kernel_size=5, num_repeat=4, expand_ratio=6, strides=(2, 2), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise'), BlockConfig(input_filters=192, output_filters=320, kernel_size=3, num_repeat=1, expand_ratio=6, strides=(1, 1), se_ratio=0.25, id_skip=True, fused_conv=False, conv_type='depthwise')), stem_base_filters=32, top_base_filters=1280, activation='simple_swish', batch_norm='default', bn_momentum=0.99, bn_epsilon=0.001, weight_decay=5e-06, drop_connect_rate=0.2, depth_divisor=8, min_depth=None, use_se=True, input_channels=3, num_classes=1000, model_name='efficientnet', rescale_input=False, data_format='channels_last', dtype='float32')\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_create_ssd_models_from_config): 32.72s\n",
      "I1021 13:57:52.478695 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_create_ssd_models_from_config): 32.72s\n",
      "[       OK ] ModelBuilderTF2Test.test_create_ssd_models_from_config\n",
      "[ RUN      ] ModelBuilderTF2Test.test_invalid_faster_rcnn_batchnorm_update\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_invalid_faster_rcnn_batchnorm_update): 0.0s\n",
      "I1021 13:57:52.525367 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_invalid_faster_rcnn_batchnorm_update): 0.0s\n",
      "[       OK ] ModelBuilderTF2Test.test_invalid_faster_rcnn_batchnorm_update\n",
      "[ RUN      ] ModelBuilderTF2Test.test_invalid_first_stage_nms_iou_threshold\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_invalid_first_stage_nms_iou_threshold): 0.0s\n",
      "I1021 13:57:52.529448 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_invalid_first_stage_nms_iou_threshold): 0.0s\n",
      "[       OK ] ModelBuilderTF2Test.test_invalid_first_stage_nms_iou_threshold\n",
      "[ RUN      ] ModelBuilderTF2Test.test_invalid_model_config_proto\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_invalid_model_config_proto): 0.0s\n",
      "I1021 13:57:52.530189 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_invalid_model_config_proto): 0.0s\n",
      "[       OK ] ModelBuilderTF2Test.test_invalid_model_config_proto\n",
      "[ RUN      ] ModelBuilderTF2Test.test_invalid_second_stage_batch_size\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_invalid_second_stage_batch_size): 0.0s\n",
      "I1021 13:57:52.533784 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_invalid_second_stage_batch_size): 0.0s\n",
      "[       OK ] ModelBuilderTF2Test.test_invalid_second_stage_batch_size\n",
      "[ RUN      ] ModelBuilderTF2Test.test_session\n",
      "[  SKIPPED ] ModelBuilderTF2Test.test_session\n",
      "[ RUN      ] ModelBuilderTF2Test.test_unknown_faster_rcnn_feature_extractor\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_unknown_faster_rcnn_feature_extractor): 0.0s\n",
      "I1021 13:57:52.537247 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_unknown_faster_rcnn_feature_extractor): 0.0s\n",
      "[       OK ] ModelBuilderTF2Test.test_unknown_faster_rcnn_feature_extractor\n",
      "[ RUN      ] ModelBuilderTF2Test.test_unknown_meta_architecture\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_unknown_meta_architecture): 0.0s\n",
      "I1021 13:57:52.537867 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_unknown_meta_architecture): 0.0s\n",
      "[       OK ] ModelBuilderTF2Test.test_unknown_meta_architecture\n",
      "[ RUN      ] ModelBuilderTF2Test.test_unknown_ssd_feature_extractor\n",
      "INFO:tensorflow:time(__main__.ModelBuilderTF2Test.test_unknown_ssd_feature_extractor): 0.0s\n",
      "I1021 13:57:52.540439 4547945984 test_util.py:2460] time(__main__.ModelBuilderTF2Test.test_unknown_ssd_feature_extractor): 0.0s\n",
      "[       OK ] ModelBuilderTF2Test.test_unknown_ssd_feature_extractor\n",
      "----------------------------------------------------------------------\n",
      "Ran 24 tests in 39.470s\n",
      "\n",
      "OK (skipped=1)\n"
     ]
    }
   ],
   "source": [
    "VERIFICATION_SCRIPT = os.path.join(paths['APIMODEL_PATH'], 'research', 'object_detection', 'builders', 'model_builder_tf2_test.py')\n",
    "# Verify Installation\n",
    "!python3 {VERIFICATION_SCRIPT}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-10-21 13:57:55.360436: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n"
     ]
    }
   ],
   "source": [
    "import tensorflow as tf \n",
    "import object_detection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--2022-10-19 19:21:34--  http://download.tensorflow.org/models/object_detection/tf2/20200711/ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8.tar.gz\n",
      "Resolving download.tensorflow.org (download.tensorflow.org)... 172.217.13.176\n",
      "Connecting to download.tensorflow.org (download.tensorflow.org)|172.217.13.176|:80... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 20515344 (20M) [application/x-tar]\n",
      "Saving to: ‘ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8.tar.gz’\n",
      "\n",
      "ssd_mobilenet_v2_fp 100%[===================>]  19.56M  15.5MB/s    in 1.3s    \n",
      "\n",
      "2022-10-19 19:21:36 (15.5 MB/s) - ‘ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8.tar.gz’ saved [20515344/20515344]\n",
      "\n",
      "x ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8/\n",
      "x ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8/checkpoint/\n",
      "x ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8/checkpoint/ckpt-0.data-00000-of-00001\n",
      "x ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8/checkpoint/checkpoint\n",
      "x ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8/checkpoint/ckpt-0.index\n",
      "x ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8/pipeline.config\n",
      "x ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8/saved_model/\n",
      "x ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8/saved_model/saved_model.pb\n",
      "x ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8/saved_model/variables/\n",
      "x ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8/saved_model/variables/variables.data-00000-of-00001\n",
      "x ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8/saved_model/variables/variables.index\n"
     ]
    }
   ],
   "source": [
    "if os.name =='posix':\n",
    "    #!brew install wget\n",
    "    !wget {PRETRAINED_MODEL_URL}\n",
    "    !mv {PRETRAINED_MODEL_NAME+'.tar.gz'} {paths['PRETRAINED_MODEL_PATH']}\n",
    "    !cd {paths['PRETRAINED_MODEL_PATH']} && tar -zxvf {PRETRAINED_MODEL_NAME+'.tar.gz'}\n",
    "if os.name == 'nt':\n",
    "    wget.download(PRETRAINED_MODEL_URL)\n",
    "    !move {PRETRAINED_MODEL_NAME+'.tar.gz'} {paths['PRETRAINED_MODEL_PATH']}\n",
    "    !cd {paths['PRETRAINED_MODEL_PATH']} && tar -zxvf {PRETRAINED_MODEL_NAME+'.tar.gz'}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create Label Map"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# labels = [  {'name':'A', 'id':1}, {'name':'B', 'id':2}, {'name':'C', 'id':3}, {'name':'D', 'id':4}, {'name': 'E', 'id': 5}, {'name': 'F', 'id': 6},\n",
    "#             {'name': 'G', 'id': 7}, {'name': 'H', 'id': 8}, {'name': 'I', 'id': 9}, {'name': 'K', 'id': 10}, {'name': 'L', 'id': 11}, {'name': 'M', 'id': 12}, \n",
    "#             {'name': 'N', 'id': 13}, {'name': 'O', 'id': 14}, {'name': 'P', 'id': 15}, {'name': 'Q', 'id': 16}, {'name': 'R', 'id': 17}, {'name': 'S', 'id': 18}, \n",
    "#             {'name': 'T', 'id': 19}, {'name': 'U', 'id': 20}, {'name': 'V', 'id': 21}, {'name': 'W', 'id': 22}, {'name': 'X', 'id': 23}, {'name': 'Y', 'id': 24} ]\n",
    "\n",
    "labels = [   {'name': 'F', 'id': 1},{'name': 'G', 'id': 2}, {'name': 'H', 'id': 3}, {'name': 'I', 'id': 4}, {'name': 'K', 'id': 5}, {'name': 'L', 'id': 6}, {'name': 'M', 'id': 7}, \n",
    "            {'name': 'N', 'id': 8}, {'name': 'O', 'id': 9}, {'name': 'P', 'id': 10}, {'name': 'Q', 'id': 11}, {'name': 'R', 'id': 12}, {'name': 'S', 'id': 13}, \n",
    "            {'name': 'T', 'id': 14}, {'name': 'U', 'id': 15}, {'name': 'V', 'id': 16}, {'name': 'W', 'id': 17}, {'name': 'X', 'id': 18}, {'name': 'Y', 'id': 19} ]            \n",
    "with open(files['LABELMAP'], 'w') as f:\n",
    "    for label in labels:\n",
    "        f.write('item { \\n')\n",
    "        f.write('\\tname:\\'{}\\'\\n'.format(label['name']))\n",
    "        f.write('\\tid:{}\\n'.format(label['id']))\n",
    "        f.write('}\\n')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create TF record"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "ARCHIVE_FILES = os.path.join(paths['IMAGE_PATH'], 'archive.tar.gz')\n",
    "if os.path.exists(ARCHIVE_FILES):\n",
    "  !tar -zxvf {ARCHIVE_FILES}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "if not os.path.exists(files['TF_RECORD_SCRIPT']):\n",
    "    !git clone https://github.com/nicknochnack/GenerateTFRecord {paths['SCRIPTS_PATH']}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Successfully created the TFRecord file: Tensorflow/workspace/annotations/train.record\n",
      "Successfully created the TFRecord file: Tensorflow/workspace/annotations/test.record\n"
     ]
    }
   ],
   "source": [
    "!python {files['TF_RECORD_SCRIPT']} -x {os.path.join(paths['IMAGE_PATH'], 'train')} -l {files['LABELMAP']} -o {os.path.join(paths['ANNOTATION_PATH'], 'train.record')} \n",
    "!python {files['TF_RECORD_SCRIPT']} -x {os.path.join(paths['IMAGE_PATH'], 'test')} -l {files['LABELMAP']} -o {os.path.join(paths['ANNOTATION_PATH'], 'test.record')} "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "if os.name =='posix':\n",
    "    !cp {os.path.join(paths['PRETRAINED_MODEL_PATH'], PRETRAINED_MODEL_NAME, 'pipeline.config')} {os.path.join(paths['CHECKPOINT_PATH'])}\n",
    "if os.name == 'nt':\n",
    "    !copy {os.path.join(paths['PRETRAINED_MODEL_PATH'], PRETRAINED_MODEL_NAME, 'pipeline.config')} {os.path.join(paths['CHECKPOINT_PATH'])}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Update Config for Transfer Learning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from object_detection.utils import config_util\n",
    "from object_detection.protos import pipeline_pb2\n",
    "from google.protobuf import text_format"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "config = config_util.get_configs_from_pipeline_file(files['PIPELINE_CONFIG'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "pipeline_config = pipeline_pb2.TrainEvalPipelineConfig()\n",
    "with tf.io.gfile.GFile(files['PIPELINE_CONFIG'], \"r\") as f:                                                                                                                                                                                                                     \n",
    "    proto_str = f.read()                                                                                                                                                                                                                                          \n",
    "    text_format.Merge(proto_str, pipeline_config) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "pipeline_config.model.ssd.num_classes = len(labels)\n",
    "pipeline_config.train_config.batch_size = 4\n",
    "pipeline_config.train_config.fine_tune_checkpoint = os.path.join(paths['PRETRAINED_MODEL_PATH'], PRETRAINED_MODEL_NAME, 'checkpoint', 'ckpt-0')\n",
    "pipeline_config.train_config.fine_tune_checkpoint_type = \"detection\"\n",
    "pipeline_config.train_input_reader.label_map_path= files['LABELMAP']\n",
    "pipeline_config.train_input_reader.tf_record_input_reader.input_path[:] = [os.path.join(paths['ANNOTATION_PATH'], 'train.record')]\n",
    "pipeline_config.eval_input_reader[0].label_map_path = files['LABELMAP']\n",
    "pipeline_config.eval_input_reader[0].tf_record_input_reader.input_path[:] = [os.path.join(paths['ANNOTATION_PATH'], 'test.record')]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "config_text = text_format.MessageToString(pipeline_config)                                                                                                                                                                                                        \n",
    "with tf.io.gfile.GFile(files['PIPELINE_CONFIG'], \"wb\") as f:                                                                                                                                                                                                                     \n",
    "    f.write(config_text)   "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Train the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "TRAINING_SCRIPT = os.path.join(paths['APIMODEL_PATH'], 'research', 'object_detection', 'model_main_tf2.py')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "command = \"python {} --model_dir={} --pipeline_config_path={} --num_train_steps=60000\".format(TRAINING_SCRIPT, paths['CHECKPOINT_PATH'],files['PIPELINE_CONFIG'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2022-10-19 19:23:17.849422: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "2022-10-19 19:23:24.317233: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "WARNING:tensorflow:There are non-GPU devices in `tf.distribute.Strategy`, not using nccl allreduce.\n",
      "W1019 19:23:24.323853 4682925568 cross_device_ops.py:1387] There are non-GPU devices in `tf.distribute.Strategy`, not using nccl allreduce.\n",
      "INFO:tensorflow:Using MirroredStrategy with devices ('/job:localhost/replica:0/task:0/device:CPU:0',)\n",
      "I1019 19:23:24.389832 4682925568 mirrored_strategy.py:374] Using MirroredStrategy with devices ('/job:localhost/replica:0/task:0/device:CPU:0',)\n",
      "INFO:tensorflow:Maybe overwriting train_steps: 60000\n",
      "I1019 19:23:24.395215 4682925568 config_util.py:552] Maybe overwriting train_steps: 60000\n",
      "INFO:tensorflow:Maybe overwriting use_bfloat16: False\n",
      "I1019 19:23:24.395364 4682925568 config_util.py:552] Maybe overwriting use_bfloat16: False\n",
      "WARNING:tensorflow:From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/object_detection/model_lib_v2.py:563: StrategyBase.experimental_distribute_datasets_from_function (from tensorflow.python.distribute.distribute_lib) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "rename to distribute_datasets_from_function\n",
      "W1019 19:23:24.440004 4682925568 deprecation.py:350] From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/object_detection/model_lib_v2.py:563: StrategyBase.experimental_distribute_datasets_from_function (from tensorflow.python.distribute.distribute_lib) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "rename to distribute_datasets_from_function\n",
      "INFO:tensorflow:Reading unweighted datasets: ['Tensorflow/workspace/annotations/train.record']\n",
      "I1019 19:23:24.456514 4682925568 dataset_builder.py:162] Reading unweighted datasets: ['Tensorflow/workspace/annotations/train.record']\n",
      "INFO:tensorflow:Reading record datasets for input file: ['Tensorflow/workspace/annotations/train.record']\n",
      "I1019 19:23:24.459177 4682925568 dataset_builder.py:79] Reading record datasets for input file: ['Tensorflow/workspace/annotations/train.record']\n",
      "INFO:tensorflow:Number of filenames to read: 1\n",
      "I1019 19:23:24.459316 4682925568 dataset_builder.py:80] Number of filenames to read: 1\n",
      "WARNING:tensorflow:num_readers has been reduced to 1 to match input file shards.\n",
      "W1019 19:23:24.459375 4682925568 dataset_builder.py:86] num_readers has been reduced to 1 to match input file shards.\n",
      "WARNING:tensorflow:From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/object_detection/builders/dataset_builder.py:100: parallel_interleave (from tensorflow.python.data.experimental.ops.interleave_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use `tf.data.Dataset.interleave(map_func, cycle_length, block_length, num_parallel_calls=tf.data.AUTOTUNE)` instead. If sloppy execution is desired, use `tf.data.Options.deterministic`.\n",
      "W1019 19:23:24.474282 4682925568 deprecation.py:350] From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/object_detection/builders/dataset_builder.py:100: parallel_interleave (from tensorflow.python.data.experimental.ops.interleave_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use `tf.data.Dataset.interleave(map_func, cycle_length, block_length, num_parallel_calls=tf.data.AUTOTUNE)` instead. If sloppy execution is desired, use `tf.data.Options.deterministic`.\n",
      "WARNING:tensorflow:From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/object_detection/builders/dataset_builder.py:235: DatasetV1.map_with_legacy_function (from tensorflow.python.data.ops.dataset_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use `tf.data.Dataset.map()\n",
      "W1019 19:23:24.505779 4682925568 deprecation.py:350] From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/object_detection/builders/dataset_builder.py:235: DatasetV1.map_with_legacy_function (from tensorflow.python.data.ops.dataset_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use `tf.data.Dataset.map()\n",
      "WARNING:tensorflow:From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/tensorflow/python/util/dispatch.py:1176: sparse_to_dense (from tensorflow.python.ops.sparse_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Create a `tf.sparse.SparseTensor` and use `tf.sparse.to_dense` instead.\n",
      "W1019 19:23:30.702205 4682925568 deprecation.py:350] From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/tensorflow/python/util/dispatch.py:1176: sparse_to_dense (from tensorflow.python.ops.sparse_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Create a `tf.sparse.SparseTensor` and use `tf.sparse.to_dense` instead.\n",
      "WARNING:tensorflow:From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/tensorflow/python/util/dispatch.py:1176: sample_distorted_bounding_box (from tensorflow.python.ops.image_ops_impl) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "`seed2` arg is deprecated.Use sample_distorted_bounding_box_v2 instead.\n",
      "W1019 19:23:33.557003 4682925568 deprecation.py:350] From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/tensorflow/python/util/dispatch.py:1176: sample_distorted_bounding_box (from tensorflow.python.ops.image_ops_impl) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "`seed2` arg is deprecated.Use sample_distorted_bounding_box_v2 instead.\n",
      "WARNING:tensorflow:From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/tensorflow/python/util/dispatch.py:1176: to_float (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use `tf.cast` instead.\n",
      "W1019 19:23:35.261534 4682925568 deprecation.py:350] From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/tensorflow/python/util/dispatch.py:1176: to_float (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use `tf.cast` instead.\n",
      "2022-10-19 19:23:37.947154: W tensorflow/core/framework/dataset.cc:769] Input of GeneratorDatasetOp::Dataset will not be optimized because the dataset does not implement the AsGraphDefInternal() method needed to apply optimizations.\n",
      "/Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/keras/backend.py:451: UserWarning: `tf.keras.backend.set_learning_phase` is deprecated and will be removed after 2020-10-11. To update it, simply pass a True/False value to the `training` argument of the `__call__` method of your layer or model.\n",
      "  warnings.warn(\n",
      "WARNING:tensorflow:From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/tensorflow/python/util/deprecation.py:629: calling map_fn_v2 (from tensorflow.python.ops.map_fn) with dtype is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use fn_output_signature instead\n",
      "W1019 19:23:57.935636 123145486327808 deprecation.py:554] From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/tensorflow/python/util/deprecation.py:629: calling map_fn_v2 (from tensorflow.python.ops.map_fn) with dtype is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use fn_output_signature instead\n",
      "INFO:tensorflow:Step 100 per-step time 1.221s\n",
      "I1019 19:25:59.768016 4682925568 model_lib_v2.py:705] Step 100 per-step time 1.221s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.6717789,\n",
      " 'Loss/localization_loss': 0.36233348,\n",
      " 'Loss/regularization_loss': 0.1538505,\n",
      " 'Loss/total_loss': 1.1879629,\n",
      " 'learning_rate': 0.0319994}\n",
      "I1019 19:25:59.772602 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.6717789,\n",
      " 'Loss/localization_loss': 0.36233348,\n",
      " 'Loss/regularization_loss': 0.1538505,\n",
      " 'Loss/total_loss': 1.1879629,\n",
      " 'learning_rate': 0.0319994}\n",
      "INFO:tensorflow:Step 200 per-step time 0.968s\n",
      "I1019 19:27:36.483079 4682925568 model_lib_v2.py:705] Step 200 per-step time 0.968s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.6471089,\n",
      " 'Loss/localization_loss': 0.2361843,\n",
      " 'Loss/regularization_loss': 0.15394482,\n",
      " 'Loss/total_loss': 1.037238,\n",
      " 'learning_rate': 0.0373328}\n",
      "I1019 19:27:36.483329 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.6471089,\n",
      " 'Loss/localization_loss': 0.2361843,\n",
      " 'Loss/regularization_loss': 0.15394482,\n",
      " 'Loss/total_loss': 1.037238,\n",
      " 'learning_rate': 0.0373328}\n",
      "INFO:tensorflow:Step 300 per-step time 0.966s\n",
      "I1019 19:29:13.082653 4682925568 model_lib_v2.py:705] Step 300 per-step time 0.966s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.3910179,\n",
      " 'Loss/localization_loss': 0.15700759,\n",
      " 'Loss/regularization_loss': 0.15383494,\n",
      " 'Loss/total_loss': 0.7018604,\n",
      " 'learning_rate': 0.0426662}\n",
      "I1019 19:29:13.085186 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.3910179,\n",
      " 'Loss/localization_loss': 0.15700759,\n",
      " 'Loss/regularization_loss': 0.15383494,\n",
      " 'Loss/total_loss': 0.7018604,\n",
      " 'learning_rate': 0.0426662}\n",
      "INFO:tensorflow:Step 400 per-step time 0.907s\n",
      "I1019 19:30:43.806084 4682925568 model_lib_v2.py:705] Step 400 per-step time 0.907s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.5424468,\n",
      " 'Loss/localization_loss': 0.25799495,\n",
      " 'Loss/regularization_loss': 0.1538267,\n",
      " 'Loss/total_loss': 0.95426846,\n",
      " 'learning_rate': 0.047999598}\n",
      "I1019 19:30:43.807132 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.5424468,\n",
      " 'Loss/localization_loss': 0.25799495,\n",
      " 'Loss/regularization_loss': 0.1538267,\n",
      " 'Loss/total_loss': 0.95426846,\n",
      " 'learning_rate': 0.047999598}\n",
      "INFO:tensorflow:Step 500 per-step time 0.887s\n",
      "I1019 19:32:12.509327 4682925568 model_lib_v2.py:705] Step 500 per-step time 0.887s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.36096057,\n",
      " 'Loss/localization_loss': 0.17054626,\n",
      " 'Loss/regularization_loss': 0.15374273,\n",
      " 'Loss/total_loss': 0.68524957,\n",
      " 'learning_rate': 0.053333}\n",
      "I1019 19:32:12.511291 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.36096057,\n",
      " 'Loss/localization_loss': 0.17054626,\n",
      " 'Loss/regularization_loss': 0.15374273,\n",
      " 'Loss/total_loss': 0.68524957,\n",
      " 'learning_rate': 0.053333}\n",
      "INFO:tensorflow:Step 600 per-step time 0.884s\n",
      "I1019 19:33:40.883647 4682925568 model_lib_v2.py:705] Step 600 per-step time 0.884s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.38677168,\n",
      " 'Loss/localization_loss': 0.1872246,\n",
      " 'Loss/regularization_loss': 0.15362868,\n",
      " 'Loss/total_loss': 0.72762495,\n",
      " 'learning_rate': 0.0586664}\n",
      "I1019 19:33:40.887058 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.38677168,\n",
      " 'Loss/localization_loss': 0.1872246,\n",
      " 'Loss/regularization_loss': 0.15362868,\n",
      " 'Loss/total_loss': 0.72762495,\n",
      " 'learning_rate': 0.0586664}\n",
      "INFO:tensorflow:Step 700 per-step time 1.024s\n",
      "I1019 19:35:23.266504 4682925568 model_lib_v2.py:705] Step 700 per-step time 1.024s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.5498964,\n",
      " 'Loss/localization_loss': 0.17062202,\n",
      " 'Loss/regularization_loss': 0.15359882,\n",
      " 'Loss/total_loss': 0.87411726,\n",
      " 'learning_rate': 0.0639998}\n",
      "I1019 19:35:23.270404 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.5498964,\n",
      " 'Loss/localization_loss': 0.17062202,\n",
      " 'Loss/regularization_loss': 0.15359882,\n",
      " 'Loss/total_loss': 0.87411726,\n",
      " 'learning_rate': 0.0639998}\n",
      "INFO:tensorflow:Step 800 per-step time 0.894s\n",
      "I1019 19:36:52.644091 4682925568 model_lib_v2.py:705] Step 800 per-step time 0.894s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.4380314,\n",
      " 'Loss/localization_loss': 0.19312446,\n",
      " 'Loss/regularization_loss': 0.15348122,\n",
      " 'Loss/total_loss': 0.7846371,\n",
      " 'learning_rate': 0.069333196}\n",
      "I1019 19:36:52.645018 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.4380314,\n",
      " 'Loss/localization_loss': 0.19312446,\n",
      " 'Loss/regularization_loss': 0.15348122,\n",
      " 'Loss/total_loss': 0.7846371,\n",
      " 'learning_rate': 0.069333196}\n",
      "INFO:tensorflow:Step 900 per-step time 0.947s\n",
      "I1019 19:38:27.353359 4682925568 model_lib_v2.py:705] Step 900 per-step time 0.947s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.48653755,\n",
      " 'Loss/localization_loss': 0.1660217,\n",
      " 'Loss/regularization_loss': 0.15336493,\n",
      " 'Loss/total_loss': 0.8059242,\n",
      " 'learning_rate': 0.074666604}\n",
      "I1019 19:38:27.356780 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.48653755,\n",
      " 'Loss/localization_loss': 0.1660217,\n",
      " 'Loss/regularization_loss': 0.15336493,\n",
      " 'Loss/total_loss': 0.8059242,\n",
      " 'learning_rate': 0.074666604}\n",
      "INFO:tensorflow:Step 1000 per-step time 1.119s\n",
      "I1019 19:40:19.226861 4682925568 model_lib_v2.py:705] Step 1000 per-step time 1.119s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.5466489,\n",
      " 'Loss/localization_loss': 0.17889982,\n",
      " 'Loss/regularization_loss': 0.15362968,\n",
      " 'Loss/total_loss': 0.8791784,\n",
      " 'learning_rate': 0.08}\n",
      "I1019 19:40:19.230072 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.5466489,\n",
      " 'Loss/localization_loss': 0.17889982,\n",
      " 'Loss/regularization_loss': 0.15362968,\n",
      " 'Loss/total_loss': 0.8791784,\n",
      " 'learning_rate': 0.08}\n",
      "INFO:tensorflow:Step 1100 per-step time 1.079s\n",
      "I1019 19:42:07.131659 4682925568 model_lib_v2.py:705] Step 1100 per-step time 1.079s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.5163011,\n",
      " 'Loss/localization_loss': 0.15461703,\n",
      " 'Loss/regularization_loss': 0.15366563,\n",
      " 'Loss/total_loss': 0.82458377,\n",
      " 'learning_rate': 0.07999918}\n",
      "I1019 19:42:07.139845 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.5163011,\n",
      " 'Loss/localization_loss': 0.15461703,\n",
      " 'Loss/regularization_loss': 0.15366563,\n",
      " 'Loss/total_loss': 0.82458377,\n",
      " 'learning_rate': 0.07999918}\n",
      "INFO:tensorflow:Step 1200 per-step time 1.129s\n",
      "I1019 19:44:00.058115 4682925568 model_lib_v2.py:705] Step 1200 per-step time 1.129s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.27469075,\n",
      " 'Loss/localization_loss': 0.09727819,\n",
      " 'Loss/regularization_loss': 0.15332372,\n",
      " 'Loss/total_loss': 0.52529263,\n",
      " 'learning_rate': 0.079996705}\n",
      "I1019 19:44:00.063067 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.27469075,\n",
      " 'Loss/localization_loss': 0.09727819,\n",
      " 'Loss/regularization_loss': 0.15332372,\n",
      " 'Loss/total_loss': 0.52529263,\n",
      " 'learning_rate': 0.079996705}\n",
      "INFO:tensorflow:Step 1300 per-step time 1.352s\n",
      "I1019 19:46:15.358400 4682925568 model_lib_v2.py:705] Step 1300 per-step time 1.352s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.29841736,\n",
      " 'Loss/localization_loss': 0.06584638,\n",
      " 'Loss/regularization_loss': 0.15308225,\n",
      " 'Loss/total_loss': 0.517346,\n",
      " 'learning_rate': 0.0799926}\n",
      "I1019 19:46:15.364879 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.29841736,\n",
      " 'Loss/localization_loss': 0.06584638,\n",
      " 'Loss/regularization_loss': 0.15308225,\n",
      " 'Loss/total_loss': 0.517346,\n",
      " 'learning_rate': 0.0799926}\n",
      "INFO:tensorflow:Step 1400 per-step time 1.141s\n",
      "I1019 19:48:09.405333 4682925568 model_lib_v2.py:705] Step 1400 per-step time 1.141s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.2605697,\n",
      " 'Loss/localization_loss': 0.10381213,\n",
      " 'Loss/regularization_loss': 0.15263481,\n",
      " 'Loss/total_loss': 0.51701665,\n",
      " 'learning_rate': 0.07998685}\n",
      "I1019 19:48:09.412041 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.2605697,\n",
      " 'Loss/localization_loss': 0.10381213,\n",
      " 'Loss/regularization_loss': 0.15263481,\n",
      " 'Loss/total_loss': 0.51701665,\n",
      " 'learning_rate': 0.07998685}\n",
      "INFO:tensorflow:Step 1500 per-step time 1.218s\n",
      "I1019 19:50:11.201285 4682925568 model_lib_v2.py:705] Step 1500 per-step time 1.218s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.27466798,\n",
      " 'Loss/localization_loss': 0.103864655,\n",
      " 'Loss/regularization_loss': 0.15219298,\n",
      " 'Loss/total_loss': 0.5307256,\n",
      " 'learning_rate': 0.07997945}\n",
      "I1019 19:50:11.204751 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.27466798,\n",
      " 'Loss/localization_loss': 0.103864655,\n",
      " 'Loss/regularization_loss': 0.15219298,\n",
      " 'Loss/total_loss': 0.5307256,\n",
      " 'learning_rate': 0.07997945}\n",
      "INFO:tensorflow:Step 1600 per-step time 1.032s\n",
      "I1019 19:51:54.415261 4682925568 model_lib_v2.py:705] Step 1600 per-step time 1.032s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.4352565,\n",
      " 'Loss/localization_loss': 0.08384915,\n",
      " 'Loss/regularization_loss': 0.15162857,\n",
      " 'Loss/total_loss': 0.6707342,\n",
      " 'learning_rate': 0.079970405}\n",
      "I1019 19:51:54.420011 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.4352565,\n",
      " 'Loss/localization_loss': 0.08384915,\n",
      " 'Loss/regularization_loss': 0.15162857,\n",
      " 'Loss/total_loss': 0.6707342,\n",
      " 'learning_rate': 0.079970405}\n",
      "INFO:tensorflow:Step 1700 per-step time 1.040s\n",
      "I1019 19:53:38.441170 4682925568 model_lib_v2.py:705] Step 1700 per-step time 1.040s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.22746687,\n",
      " 'Loss/localization_loss': 0.086169966,\n",
      " 'Loss/regularization_loss': 0.15121342,\n",
      " 'Loss/total_loss': 0.46485025,\n",
      " 'learning_rate': 0.07995972}\n",
      "I1019 19:53:38.444452 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.22746687,\n",
      " 'Loss/localization_loss': 0.086169966,\n",
      " 'Loss/regularization_loss': 0.15121342,\n",
      " 'Loss/total_loss': 0.46485025,\n",
      " 'learning_rate': 0.07995972}\n",
      "INFO:tensorflow:Step 1800 per-step time 1.126s\n",
      "I1019 19:55:31.038224 4682925568 model_lib_v2.py:705] Step 1800 per-step time 1.126s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.2851164,\n",
      " 'Loss/localization_loss': 0.060819827,\n",
      " 'Loss/regularization_loss': 0.1506607,\n",
      " 'Loss/total_loss': 0.49659693,\n",
      " 'learning_rate': 0.0799474}\n",
      "I1019 19:55:31.041308 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.2851164,\n",
      " 'Loss/localization_loss': 0.060819827,\n",
      " 'Loss/regularization_loss': 0.1506607,\n",
      " 'Loss/total_loss': 0.49659693,\n",
      " 'learning_rate': 0.0799474}\n",
      "INFO:tensorflow:Step 1900 per-step time 1.069s\n",
      "I1019 19:57:17.871187 4682925568 model_lib_v2.py:705] Step 1900 per-step time 1.069s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.30304304,\n",
      " 'Loss/localization_loss': 0.13496825,\n",
      " 'Loss/regularization_loss': 0.14999399,\n",
      " 'Loss/total_loss': 0.5880053,\n",
      " 'learning_rate': 0.07993342}\n",
      "I1019 19:57:17.875082 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.30304304,\n",
      " 'Loss/localization_loss': 0.13496825,\n",
      " 'Loss/regularization_loss': 0.14999399,\n",
      " 'Loss/total_loss': 0.5880053,\n",
      " 'learning_rate': 0.07993342}\n",
      "INFO:tensorflow:Step 2000 per-step time 1.024s\n",
      "I1019 19:59:00.216851 4682925568 model_lib_v2.py:705] Step 2000 per-step time 1.024s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.21525776,\n",
      " 'Loss/localization_loss': 0.13291654,\n",
      " 'Loss/regularization_loss': 0.14948687,\n",
      " 'Loss/total_loss': 0.49766117,\n",
      " 'learning_rate': 0.07991781}\n",
      "I1019 19:59:00.220407 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.21525776,\n",
      " 'Loss/localization_loss': 0.13291654,\n",
      " 'Loss/regularization_loss': 0.14948687,\n",
      " 'Loss/total_loss': 0.49766117,\n",
      " 'learning_rate': 0.07991781}\n",
      "INFO:tensorflow:Step 2100 per-step time 1.140s\n",
      "I1019 20:00:54.251803 4682925568 model_lib_v2.py:705] Step 2100 per-step time 1.140s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.23326522,\n",
      " 'Loss/localization_loss': 0.09548673,\n",
      " 'Loss/regularization_loss': 0.14899072,\n",
      " 'Loss/total_loss': 0.47774267,\n",
      " 'learning_rate': 0.07990056}\n",
      "I1019 20:00:54.256633 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.23326522,\n",
      " 'Loss/localization_loss': 0.09548673,\n",
      " 'Loss/regularization_loss': 0.14899072,\n",
      " 'Loss/total_loss': 0.47774267,\n",
      " 'learning_rate': 0.07990056}\n",
      "INFO:tensorflow:Step 2200 per-step time 0.976s\n",
      "I1019 20:02:31.808667 4682925568 model_lib_v2.py:705] Step 2200 per-step time 0.976s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.22492109,\n",
      " 'Loss/localization_loss': 0.051160056,\n",
      " 'Loss/regularization_loss': 0.14843419,\n",
      " 'Loss/total_loss': 0.42451534,\n",
      " 'learning_rate': 0.07988167}\n",
      "I1019 20:02:31.812176 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.22492109,\n",
      " 'Loss/localization_loss': 0.051160056,\n",
      " 'Loss/regularization_loss': 0.14843419,\n",
      " 'Loss/total_loss': 0.42451534,\n",
      " 'learning_rate': 0.07988167}\n",
      "INFO:tensorflow:Step 2300 per-step time 1.024s\n",
      "I1019 20:04:14.220490 4682925568 model_lib_v2.py:705] Step 2300 per-step time 1.024s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.20070621,\n",
      " 'Loss/localization_loss': 0.1099877,\n",
      " 'Loss/regularization_loss': 0.1479054,\n",
      " 'Loss/total_loss': 0.4585993,\n",
      " 'learning_rate': 0.07986114}\n",
      "I1019 20:04:14.223674 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.20070621,\n",
      " 'Loss/localization_loss': 0.1099877,\n",
      " 'Loss/regularization_loss': 0.1479054,\n",
      " 'Loss/total_loss': 0.4585993,\n",
      " 'learning_rate': 0.07986114}\n",
      "INFO:tensorflow:Step 2400 per-step time 1.001s\n",
      "I1019 20:05:54.354506 4682925568 model_lib_v2.py:705] Step 2400 per-step time 1.001s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.113252684,\n",
      " 'Loss/localization_loss': 0.049933463,\n",
      " 'Loss/regularization_loss': 0.14740255,\n",
      " 'Loss/total_loss': 0.31058872,\n",
      " 'learning_rate': 0.07983897}\n",
      "I1019 20:05:54.357069 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.113252684,\n",
      " 'Loss/localization_loss': 0.049933463,\n",
      " 'Loss/regularization_loss': 0.14740255,\n",
      " 'Loss/total_loss': 0.31058872,\n",
      " 'learning_rate': 0.07983897}\n",
      "INFO:tensorflow:Step 2500 per-step time 0.928s\n",
      "I1019 20:07:27.170131 4682925568 model_lib_v2.py:705] Step 2500 per-step time 0.928s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.24548447,\n",
      " 'Loss/localization_loss': 0.14645657,\n",
      " 'Loss/regularization_loss': 0.1468182,\n",
      " 'Loss/total_loss': 0.53875923,\n",
      " 'learning_rate': 0.079815164}\n",
      "I1019 20:07:27.173741 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.24548447,\n",
      " 'Loss/localization_loss': 0.14645657,\n",
      " 'Loss/regularization_loss': 0.1468182,\n",
      " 'Loss/total_loss': 0.53875923,\n",
      " 'learning_rate': 0.079815164}\n",
      "INFO:tensorflow:Step 2600 per-step time 0.864s\n",
      "I1019 20:08:53.551236 4682925568 model_lib_v2.py:705] Step 2600 per-step time 0.864s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.12388015,\n",
      " 'Loss/localization_loss': 0.05189615,\n",
      " 'Loss/regularization_loss': 0.14632139,\n",
      " 'Loss/total_loss': 0.3220977,\n",
      " 'learning_rate': 0.07978972}\n",
      "I1019 20:08:53.553181 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.12388015,\n",
      " 'Loss/localization_loss': 0.05189615,\n",
      " 'Loss/regularization_loss': 0.14632139,\n",
      " 'Loss/total_loss': 0.3220977,\n",
      " 'learning_rate': 0.07978972}\n",
      "INFO:tensorflow:Step 2700 per-step time 0.860s\n",
      "I1019 20:10:19.519003 4682925568 model_lib_v2.py:705] Step 2700 per-step time 0.860s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.122495405,\n",
      " 'Loss/localization_loss': 0.10360649,\n",
      " 'Loss/regularization_loss': 0.14582421,\n",
      " 'Loss/total_loss': 0.3719261,\n",
      " 'learning_rate': 0.07976264}\n",
      "I1019 20:10:19.520099 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.122495405,\n",
      " 'Loss/localization_loss': 0.10360649,\n",
      " 'Loss/regularization_loss': 0.14582421,\n",
      " 'Loss/total_loss': 0.3719261,\n",
      " 'learning_rate': 0.07976264}\n",
      "INFO:tensorflow:Step 2800 per-step time 0.957s\n",
      "I1019 20:11:55.240591 4682925568 model_lib_v2.py:705] Step 2800 per-step time 0.957s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.15624002,\n",
      " 'Loss/localization_loss': 0.044450916,\n",
      " 'Loss/regularization_loss': 0.1453129,\n",
      " 'Loss/total_loss': 0.34600383,\n",
      " 'learning_rate': 0.07973392}\n",
      "I1019 20:11:55.244194 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.15624002,\n",
      " 'Loss/localization_loss': 0.044450916,\n",
      " 'Loss/regularization_loss': 0.1453129,\n",
      " 'Loss/total_loss': 0.34600383,\n",
      " 'learning_rate': 0.07973392}\n",
      "INFO:tensorflow:Step 2900 per-step time 0.853s\n",
      "I1019 20:13:20.538158 4682925568 model_lib_v2.py:705] Step 2900 per-step time 0.853s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.19458003,\n",
      " 'Loss/localization_loss': 0.044549692,\n",
      " 'Loss/regularization_loss': 0.14491613,\n",
      " 'Loss/total_loss': 0.38404584,\n",
      " 'learning_rate': 0.07970358}\n",
      "I1019 20:13:20.539013 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.19458003,\n",
      " 'Loss/localization_loss': 0.044549692,\n",
      " 'Loss/regularization_loss': 0.14491613,\n",
      " 'Loss/total_loss': 0.38404584,\n",
      " 'learning_rate': 0.07970358}\n",
      "INFO:tensorflow:Step 3000 per-step time 0.846s\n",
      "I1019 20:14:45.190699 4682925568 model_lib_v2.py:705] Step 3000 per-step time 0.846s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.24399264,\n",
      " 'Loss/localization_loss': 0.109773956,\n",
      " 'Loss/regularization_loss': 0.14437623,\n",
      " 'Loss/total_loss': 0.49814284,\n",
      " 'learning_rate': 0.0796716}\n",
      "I1019 20:14:45.191738 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.24399264,\n",
      " 'Loss/localization_loss': 0.109773956,\n",
      " 'Loss/regularization_loss': 0.14437623,\n",
      " 'Loss/total_loss': 0.49814284,\n",
      " 'learning_rate': 0.0796716}\n",
      "INFO:tensorflow:Step 3100 per-step time 0.885s\n",
      "I1019 20:16:13.661840 4682925568 model_lib_v2.py:705] Step 3100 per-step time 0.885s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.15010774,\n",
      " 'Loss/localization_loss': 0.08448484,\n",
      " 'Loss/regularization_loss': 0.14396264,\n",
      " 'Loss/total_loss': 0.3785552,\n",
      " 'learning_rate': 0.07963799}\n",
      "I1019 20:16:13.666364 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.15010774,\n",
      " 'Loss/localization_loss': 0.08448484,\n",
      " 'Loss/regularization_loss': 0.14396264,\n",
      " 'Loss/total_loss': 0.3785552,\n",
      " 'learning_rate': 0.07963799}\n",
      "INFO:tensorflow:Step 3200 per-step time 1.600s\n",
      "I1019 20:18:53.723447 4682925568 model_lib_v2.py:705] Step 3200 per-step time 1.600s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.2938043,\n",
      " 'Loss/localization_loss': 0.07412177,\n",
      " 'Loss/regularization_loss': 0.14343245,\n",
      " 'Loss/total_loss': 0.5113585,\n",
      " 'learning_rate': 0.07960275}\n",
      "I1019 20:18:53.731696 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.2938043,\n",
      " 'Loss/localization_loss': 0.07412177,\n",
      " 'Loss/regularization_loss': 0.14343245,\n",
      " 'Loss/total_loss': 0.5113585,\n",
      " 'learning_rate': 0.07960275}\n",
      "INFO:tensorflow:Step 3300 per-step time 1.437s\n",
      "I1019 20:21:17.411845 4682925568 model_lib_v2.py:705] Step 3300 per-step time 1.437s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.22022517,\n",
      " 'Loss/localization_loss': 0.10669107,\n",
      " 'Loss/regularization_loss': 0.1430412,\n",
      " 'Loss/total_loss': 0.4699574,\n",
      " 'learning_rate': 0.07956588}\n",
      "I1019 20:21:17.417851 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.22022517,\n",
      " 'Loss/localization_loss': 0.10669107,\n",
      " 'Loss/regularization_loss': 0.1430412,\n",
      " 'Loss/total_loss': 0.4699574,\n",
      " 'learning_rate': 0.07956588}\n",
      "INFO:tensorflow:Step 3400 per-step time 1.129s\n",
      "I1019 20:23:10.235754 4682925568 model_lib_v2.py:705] Step 3400 per-step time 1.129s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.14843713,\n",
      " 'Loss/localization_loss': 0.05656062,\n",
      " 'Loss/regularization_loss': 0.14256874,\n",
      " 'Loss/total_loss': 0.3475665,\n",
      " 'learning_rate': 0.079527386}\n",
      "I1019 20:23:10.240451 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.14843713,\n",
      " 'Loss/localization_loss': 0.05656062,\n",
      " 'Loss/regularization_loss': 0.14256874,\n",
      " 'Loss/total_loss': 0.3475665,\n",
      " 'learning_rate': 0.079527386}\n",
      "INFO:tensorflow:Step 3500 per-step time 0.953s\n",
      "I1019 20:24:45.512037 4682925568 model_lib_v2.py:705] Step 3500 per-step time 0.953s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.21324433,\n",
      " 'Loss/localization_loss': 0.11051793,\n",
      " 'Loss/regularization_loss': 0.14211604,\n",
      " 'Loss/total_loss': 0.4658783,\n",
      " 'learning_rate': 0.07948727}\n",
      "I1019 20:24:45.515285 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.21324433,\n",
      " 'Loss/localization_loss': 0.11051793,\n",
      " 'Loss/regularization_loss': 0.14211604,\n",
      " 'Loss/total_loss': 0.4658783,\n",
      " 'learning_rate': 0.07948727}\n",
      "INFO:tensorflow:Step 3600 per-step time 0.959s\n",
      "I1019 20:26:21.365091 4682925568 model_lib_v2.py:705] Step 3600 per-step time 0.959s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.19324815,\n",
      " 'Loss/localization_loss': 0.07824193,\n",
      " 'Loss/regularization_loss': 0.14164168,\n",
      " 'Loss/total_loss': 0.41313177,\n",
      " 'learning_rate': 0.079445526}\n",
      "I1019 20:26:21.368478 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.19324815,\n",
      " 'Loss/localization_loss': 0.07824193,\n",
      " 'Loss/regularization_loss': 0.14164168,\n",
      " 'Loss/total_loss': 0.41313177,\n",
      " 'learning_rate': 0.079445526}\n",
      "INFO:tensorflow:Step 3700 per-step time 0.967s\n",
      "I1019 20:27:58.048573 4682925568 model_lib_v2.py:705] Step 3700 per-step time 0.967s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.22170068,\n",
      " 'Loss/localization_loss': 0.083349094,\n",
      " 'Loss/regularization_loss': 0.14115953,\n",
      " 'Loss/total_loss': 0.4462093,\n",
      " 'learning_rate': 0.07940216}\n",
      "I1019 20:27:58.052069 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.22170068,\n",
      " 'Loss/localization_loss': 0.083349094,\n",
      " 'Loss/regularization_loss': 0.14115953,\n",
      " 'Loss/total_loss': 0.4462093,\n",
      " 'learning_rate': 0.07940216}\n",
      "INFO:tensorflow:Step 3800 per-step time 0.885s\n",
      "I1019 20:29:26.564476 4682925568 model_lib_v2.py:705] Step 3800 per-step time 0.885s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.16174696,\n",
      " 'Loss/localization_loss': 0.08833302,\n",
      " 'Loss/regularization_loss': 0.14063548,\n",
      " 'Loss/total_loss': 0.39071548,\n",
      " 'learning_rate': 0.079357184}\n",
      "I1019 20:29:26.567439 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.16174696,\n",
      " 'Loss/localization_loss': 0.08833302,\n",
      " 'Loss/regularization_loss': 0.14063548,\n",
      " 'Loss/total_loss': 0.39071548,\n",
      " 'learning_rate': 0.079357184}\n",
      "INFO:tensorflow:Step 3900 per-step time 0.875s\n",
      "I1019 20:30:54.103821 4682925568 model_lib_v2.py:705] Step 3900 per-step time 0.875s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.15074165,\n",
      " 'Loss/localization_loss': 0.072436444,\n",
      " 'Loss/regularization_loss': 0.14011264,\n",
      " 'Loss/total_loss': 0.36329073,\n",
      " 'learning_rate': 0.07931058}\n",
      "I1019 20:30:54.104902 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.15074165,\n",
      " 'Loss/localization_loss': 0.072436444,\n",
      " 'Loss/regularization_loss': 0.14011264,\n",
      " 'Loss/total_loss': 0.36329073,\n",
      " 'learning_rate': 0.07931058}\n",
      "INFO:tensorflow:Step 4000 per-step time 1.055s\n",
      "I1019 20:32:39.584896 4682925568 model_lib_v2.py:705] Step 4000 per-step time 1.055s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.21724838,\n",
      " 'Loss/localization_loss': 0.044746842,\n",
      " 'Loss/regularization_loss': 0.13975719,\n",
      " 'Loss/total_loss': 0.4017524,\n",
      " 'learning_rate': 0.07926236}\n",
      "I1019 20:32:39.588858 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.21724838,\n",
      " 'Loss/localization_loss': 0.044746842,\n",
      " 'Loss/regularization_loss': 0.13975719,\n",
      " 'Loss/total_loss': 0.4017524,\n",
      " 'learning_rate': 0.07926236}\n",
      "INFO:tensorflow:Step 4100 per-step time 1.254s\n",
      "I1019 20:34:45.078558 4682925568 model_lib_v2.py:705] Step 4100 per-step time 1.254s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.16072959,\n",
      " 'Loss/localization_loss': 0.05948294,\n",
      " 'Loss/regularization_loss': 0.13929793,\n",
      " 'Loss/total_loss': 0.35951045,\n",
      " 'learning_rate': 0.07921253}\n",
      "I1019 20:34:45.090117 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.16072959,\n",
      " 'Loss/localization_loss': 0.05948294,\n",
      " 'Loss/regularization_loss': 0.13929793,\n",
      " 'Loss/total_loss': 0.35951045,\n",
      " 'learning_rate': 0.07921253}\n",
      "INFO:tensorflow:Step 4200 per-step time 1.007s\n",
      "I1019 20:36:25.697667 4682925568 model_lib_v2.py:705] Step 4200 per-step time 1.007s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.16414742,\n",
      " 'Loss/localization_loss': 0.056434568,\n",
      " 'Loss/regularization_loss': 0.13891032,\n",
      " 'Loss/total_loss': 0.3594923,\n",
      " 'learning_rate': 0.07916109}\n",
      "I1019 20:36:25.701662 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.16414742,\n",
      " 'Loss/localization_loss': 0.056434568,\n",
      " 'Loss/regularization_loss': 0.13891032,\n",
      " 'Loss/total_loss': 0.3594923,\n",
      " 'learning_rate': 0.07916109}\n",
      "INFO:tensorflow:Step 4300 per-step time 0.938s\n",
      "I1019 20:37:59.514608 4682925568 model_lib_v2.py:705] Step 4300 per-step time 0.938s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.10703184,\n",
      " 'Loss/localization_loss': 0.06575014,\n",
      " 'Loss/regularization_loss': 0.13851346,\n",
      " 'Loss/total_loss': 0.31129545,\n",
      " 'learning_rate': 0.07910804}\n",
      "I1019 20:37:59.515930 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.10703184,\n",
      " 'Loss/localization_loss': 0.06575014,\n",
      " 'Loss/regularization_loss': 0.13851346,\n",
      " 'Loss/total_loss': 0.31129545,\n",
      " 'learning_rate': 0.07910804}\n",
      "INFO:tensorflow:Step 4400 per-step time 0.896s\n",
      "I1019 20:39:29.123557 4682925568 model_lib_v2.py:705] Step 4400 per-step time 0.896s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.20552835,\n",
      " 'Loss/localization_loss': 0.046843544,\n",
      " 'Loss/regularization_loss': 0.13800898,\n",
      " 'Loss/total_loss': 0.39038086,\n",
      " 'learning_rate': 0.07905338}\n",
      "I1019 20:39:29.126784 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.20552835,\n",
      " 'Loss/localization_loss': 0.046843544,\n",
      " 'Loss/regularization_loss': 0.13800898,\n",
      " 'Loss/total_loss': 0.39038086,\n",
      " 'learning_rate': 0.07905338}\n",
      "INFO:tensorflow:Step 4500 per-step time 0.875s\n",
      "I1019 20:40:56.600598 4682925568 model_lib_v2.py:705] Step 4500 per-step time 0.875s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.13487686,\n",
      " 'Loss/localization_loss': 0.04877624,\n",
      " 'Loss/regularization_loss': 0.13746299,\n",
      " 'Loss/total_loss': 0.3211161,\n",
      " 'learning_rate': 0.07899711}\n",
      "I1019 20:40:56.601509 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.13487686,\n",
      " 'Loss/localization_loss': 0.04877624,\n",
      " 'Loss/regularization_loss': 0.13746299,\n",
      " 'Loss/total_loss': 0.3211161,\n",
      " 'learning_rate': 0.07899711}\n",
      "INFO:tensorflow:Step 4600 per-step time 0.885s\n",
      "I1019 20:42:25.082171 4682925568 model_lib_v2.py:705] Step 4600 per-step time 0.885s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.17138152,\n",
      " 'Loss/localization_loss': 0.06141231,\n",
      " 'Loss/regularization_loss': 0.13691759,\n",
      " 'Loss/total_loss': 0.3697114,\n",
      " 'learning_rate': 0.078939244}\n",
      "I1019 20:42:25.085063 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.17138152,\n",
      " 'Loss/localization_loss': 0.06141231,\n",
      " 'Loss/regularization_loss': 0.13691759,\n",
      " 'Loss/total_loss': 0.3697114,\n",
      " 'learning_rate': 0.078939244}\n",
      "INFO:tensorflow:Step 4700 per-step time 0.983s\n",
      "I1019 20:44:03.411488 4682925568 model_lib_v2.py:705] Step 4700 per-step time 0.983s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.118568964,\n",
      " 'Loss/localization_loss': 0.040040337,\n",
      " 'Loss/regularization_loss': 0.13644506,\n",
      " 'Loss/total_loss': 0.29505435,\n",
      " 'learning_rate': 0.07887978}\n",
      "I1019 20:44:03.414644 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.118568964,\n",
      " 'Loss/localization_loss': 0.040040337,\n",
      " 'Loss/regularization_loss': 0.13644506,\n",
      " 'Loss/total_loss': 0.29505435,\n",
      " 'learning_rate': 0.07887978}\n",
      "INFO:tensorflow:Step 4800 per-step time 0.909s\n",
      "I1019 20:45:34.267462 4682925568 model_lib_v2.py:705] Step 4800 per-step time 0.909s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.10909485,\n",
      " 'Loss/localization_loss': 0.04060741,\n",
      " 'Loss/regularization_loss': 0.13600796,\n",
      " 'Loss/total_loss': 0.28571022,\n",
      " 'learning_rate': 0.07881871}\n",
      "I1019 20:45:34.270349 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.10909485,\n",
      " 'Loss/localization_loss': 0.04060741,\n",
      " 'Loss/regularization_loss': 0.13600796,\n",
      " 'Loss/total_loss': 0.28571022,\n",
      " 'learning_rate': 0.07881871}\n",
      "INFO:tensorflow:Step 4900 per-step time 0.901s\n",
      "I1019 20:47:04.348676 4682925568 model_lib_v2.py:705] Step 4900 per-step time 0.901s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.13555254,\n",
      " 'Loss/localization_loss': 0.04508223,\n",
      " 'Loss/regularization_loss': 0.13552071,\n",
      " 'Loss/total_loss': 0.3161555,\n",
      " 'learning_rate': 0.07875605}\n",
      "I1019 20:47:04.352340 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.13555254,\n",
      " 'Loss/localization_loss': 0.04508223,\n",
      " 'Loss/regularization_loss': 0.13552071,\n",
      " 'Loss/total_loss': 0.3161555,\n",
      " 'learning_rate': 0.07875605}\n",
      "INFO:tensorflow:Step 5000 per-step time 0.954s\n",
      "I1019 20:48:39.764314 4682925568 model_lib_v2.py:705] Step 5000 per-step time 0.954s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.10359704,\n",
      " 'Loss/localization_loss': 0.06786304,\n",
      " 'Loss/regularization_loss': 0.13521734,\n",
      " 'Loss/total_loss': 0.30667743,\n",
      " 'learning_rate': 0.078691795}\n",
      "I1019 20:48:39.767807 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.10359704,\n",
      " 'Loss/localization_loss': 0.06786304,\n",
      " 'Loss/regularization_loss': 0.13521734,\n",
      " 'Loss/total_loss': 0.30667743,\n",
      " 'learning_rate': 0.078691795}\n",
      "INFO:tensorflow:Step 5100 per-step time 0.960s\n",
      "I1019 20:50:15.742381 4682925568 model_lib_v2.py:705] Step 5100 per-step time 0.960s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.14319696,\n",
      " 'Loss/localization_loss': 0.04259515,\n",
      " 'Loss/regularization_loss': 0.13468632,\n",
      " 'Loss/total_loss': 0.32047844,\n",
      " 'learning_rate': 0.07862595}\n",
      "I1019 20:50:15.746634 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.14319696,\n",
      " 'Loss/localization_loss': 0.04259515,\n",
      " 'Loss/regularization_loss': 0.13468632,\n",
      " 'Loss/total_loss': 0.32047844,\n",
      " 'learning_rate': 0.07862595}\n",
      "INFO:tensorflow:Step 5200 per-step time 0.873s\n",
      "I1019 20:51:43.008029 4682925568 model_lib_v2.py:705] Step 5200 per-step time 0.873s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.15931235,\n",
      " 'Loss/localization_loss': 0.13418674,\n",
      " 'Loss/regularization_loss': 0.13419332,\n",
      " 'Loss/total_loss': 0.4276924,\n",
      " 'learning_rate': 0.07855851}\n",
      "I1019 20:51:43.011394 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.15931235,\n",
      " 'Loss/localization_loss': 0.13418674,\n",
      " 'Loss/regularization_loss': 0.13419332,\n",
      " 'Loss/total_loss': 0.4276924,\n",
      " 'learning_rate': 0.07855851}\n",
      "INFO:tensorflow:Step 5300 per-step time 0.861s\n",
      "I1019 20:53:09.071553 4682925568 model_lib_v2.py:705] Step 5300 per-step time 0.861s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.14433551,\n",
      " 'Loss/localization_loss': 0.046311166,\n",
      " 'Loss/regularization_loss': 0.13376711,\n",
      " 'Loss/total_loss': 0.32441378,\n",
      " 'learning_rate': 0.07848949}\n",
      "I1019 20:53:09.071805 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.14433551,\n",
      " 'Loss/localization_loss': 0.046311166,\n",
      " 'Loss/regularization_loss': 0.13376711,\n",
      " 'Loss/total_loss': 0.32441378,\n",
      " 'learning_rate': 0.07848949}\n",
      "INFO:tensorflow:Step 5400 per-step time 0.861s\n",
      "I1019 20:54:35.212765 4682925568 model_lib_v2.py:705] Step 5400 per-step time 0.861s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.24899803,\n",
      " 'Loss/localization_loss': 0.07386053,\n",
      " 'Loss/regularization_loss': 0.13323635,\n",
      " 'Loss/total_loss': 0.45609492,\n",
      " 'learning_rate': 0.078418896}\n",
      "I1019 20:54:35.213036 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.24899803,\n",
      " 'Loss/localization_loss': 0.07386053,\n",
      " 'Loss/regularization_loss': 0.13323635,\n",
      " 'Loss/total_loss': 0.45609492,\n",
      " 'learning_rate': 0.078418896}\n",
      "INFO:tensorflow:Step 5500 per-step time 0.890s\n",
      "I1019 20:56:04.231299 4682925568 model_lib_v2.py:705] Step 5500 per-step time 0.890s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.093443416,\n",
      " 'Loss/localization_loss': 0.055699598,\n",
      " 'Loss/regularization_loss': 0.13276353,\n",
      " 'Loss/total_loss': 0.28190655,\n",
      " 'learning_rate': 0.078346714}\n",
      "I1019 20:56:04.234311 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.093443416,\n",
      " 'Loss/localization_loss': 0.055699598,\n",
      " 'Loss/regularization_loss': 0.13276353,\n",
      " 'Loss/total_loss': 0.28190655,\n",
      " 'learning_rate': 0.078346714}\n",
      "INFO:tensorflow:Step 5600 per-step time 0.863s\n",
      "I1019 20:57:30.561721 4682925568 model_lib_v2.py:705] Step 5600 per-step time 0.863s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.14867647,\n",
      " 'Loss/localization_loss': 0.07586909,\n",
      " 'Loss/regularization_loss': 0.13225006,\n",
      " 'Loss/total_loss': 0.3567956,\n",
      " 'learning_rate': 0.07827295}\n",
      "I1019 20:57:30.565169 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.14867647,\n",
      " 'Loss/localization_loss': 0.07586909,\n",
      " 'Loss/regularization_loss': 0.13225006,\n",
      " 'Loss/total_loss': 0.3567956,\n",
      " 'learning_rate': 0.07827295}\n",
      "INFO:tensorflow:Step 5700 per-step time 0.898s\n",
      "I1019 20:59:00.379115 4682925568 model_lib_v2.py:705] Step 5700 per-step time 0.898s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.099303804,\n",
      " 'Loss/localization_loss': 0.03380699,\n",
      " 'Loss/regularization_loss': 0.13173878,\n",
      " 'Loss/total_loss': 0.26484957,\n",
      " 'learning_rate': 0.07819763}\n",
      "I1019 20:59:00.382153 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.099303804,\n",
      " 'Loss/localization_loss': 0.03380699,\n",
      " 'Loss/regularization_loss': 0.13173878,\n",
      " 'Loss/total_loss': 0.26484957,\n",
      " 'learning_rate': 0.07819763}\n",
      "INFO:tensorflow:Step 5800 per-step time 0.956s\n",
      "I1019 21:00:35.984624 4682925568 model_lib_v2.py:705] Step 5800 per-step time 0.956s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.13475326,\n",
      " 'Loss/localization_loss': 0.035947755,\n",
      " 'Loss/regularization_loss': 0.13140088,\n",
      " 'Loss/total_loss': 0.3021019,\n",
      " 'learning_rate': 0.07812072}\n",
      "I1019 21:00:35.989740 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.13475326,\n",
      " 'Loss/localization_loss': 0.035947755,\n",
      " 'Loss/regularization_loss': 0.13140088,\n",
      " 'Loss/total_loss': 0.3021019,\n",
      " 'learning_rate': 0.07812072}\n",
      "INFO:tensorflow:Step 5900 per-step time 0.943s\n",
      "I1019 21:02:10.266040 4682925568 model_lib_v2.py:705] Step 5900 per-step time 0.943s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.10849415,\n",
      " 'Loss/localization_loss': 0.06084062,\n",
      " 'Loss/regularization_loss': 0.13108407,\n",
      " 'Loss/total_loss': 0.30041885,\n",
      " 'learning_rate': 0.078042254}\n",
      "I1019 21:02:10.269235 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.10849415,\n",
      " 'Loss/localization_loss': 0.06084062,\n",
      " 'Loss/regularization_loss': 0.13108407,\n",
      " 'Loss/total_loss': 0.30041885,\n",
      " 'learning_rate': 0.078042254}\n",
      "INFO:tensorflow:Step 6000 per-step time 0.861s\n",
      "I1019 21:03:36.357648 4682925568 model_lib_v2.py:705] Step 6000 per-step time 0.861s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.14881168,\n",
      " 'Loss/localization_loss': 0.08391894,\n",
      " 'Loss/regularization_loss': 0.1305587,\n",
      " 'Loss/total_loss': 0.3632893,\n",
      " 'learning_rate': 0.07796223}\n",
      "I1019 21:03:36.358532 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.14881168,\n",
      " 'Loss/localization_loss': 0.08391894,\n",
      " 'Loss/regularization_loss': 0.1305587,\n",
      " 'Loss/total_loss': 0.3632893,\n",
      " 'learning_rate': 0.07796223}\n",
      "INFO:tensorflow:Step 6100 per-step time 0.946s\n",
      "I1019 21:05:10.999709 4682925568 model_lib_v2.py:705] Step 6100 per-step time 0.946s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.18667808,\n",
      " 'Loss/localization_loss': 0.034993332,\n",
      " 'Loss/regularization_loss': 0.13009942,\n",
      " 'Loss/total_loss': 0.35177082,\n",
      " 'learning_rate': 0.077880636}\n",
      "I1019 21:05:11.004127 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.18667808,\n",
      " 'Loss/localization_loss': 0.034993332,\n",
      " 'Loss/regularization_loss': 0.13009942,\n",
      " 'Loss/total_loss': 0.35177082,\n",
      " 'learning_rate': 0.077880636}\n",
      "INFO:tensorflow:Step 6200 per-step time 0.846s\n",
      "I1019 21:06:35.600044 4682925568 model_lib_v2.py:705] Step 6200 per-step time 0.846s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.12335269,\n",
      " 'Loss/localization_loss': 0.088849165,\n",
      " 'Loss/regularization_loss': 0.12951648,\n",
      " 'Loss/total_loss': 0.34171832,\n",
      " 'learning_rate': 0.07779749}\n",
      "I1019 21:06:35.600294 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.12335269,\n",
      " 'Loss/localization_loss': 0.088849165,\n",
      " 'Loss/regularization_loss': 0.12951648,\n",
      " 'Loss/total_loss': 0.34171832,\n",
      " 'learning_rate': 0.07779749}\n",
      "INFO:tensorflow:Step 6300 per-step time 0.829s\n",
      "I1019 21:07:58.454746 4682925568 model_lib_v2.py:705] Step 6300 per-step time 0.829s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.1365002,\n",
      " 'Loss/localization_loss': 0.049836025,\n",
      " 'Loss/regularization_loss': 0.12904082,\n",
      " 'Loss/total_loss': 0.31537706,\n",
      " 'learning_rate': 0.07771279}\n",
      "I1019 21:07:58.455823 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.1365002,\n",
      " 'Loss/localization_loss': 0.049836025,\n",
      " 'Loss/regularization_loss': 0.12904082,\n",
      " 'Loss/total_loss': 0.31537706,\n",
      " 'learning_rate': 0.07771279}\n",
      "INFO:tensorflow:Step 6400 per-step time 0.846s\n",
      "I1019 21:09:23.089315 4682925568 model_lib_v2.py:705] Step 6400 per-step time 0.846s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.095243365,\n",
      " 'Loss/localization_loss': 0.0421379,\n",
      " 'Loss/regularization_loss': 0.12856528,\n",
      " 'Loss/total_loss': 0.26594654,\n",
      " 'learning_rate': 0.077626534}\n",
      "I1019 21:09:23.091539 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.095243365,\n",
      " 'Loss/localization_loss': 0.0421379,\n",
      " 'Loss/regularization_loss': 0.12856528,\n",
      " 'Loss/total_loss': 0.26594654,\n",
      " 'learning_rate': 0.077626534}\n",
      "INFO:tensorflow:Step 6500 per-step time 0.835s\n",
      "I1019 21:10:46.573381 4682925568 model_lib_v2.py:705] Step 6500 per-step time 0.835s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.103441305,\n",
      " 'Loss/localization_loss': 0.036715664,\n",
      " 'Loss/regularization_loss': 0.12810305,\n",
      " 'Loss/total_loss': 0.26826,\n",
      " 'learning_rate': 0.077538736}\n",
      "I1019 21:10:46.574462 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.103441305,\n",
      " 'Loss/localization_loss': 0.036715664,\n",
      " 'Loss/regularization_loss': 0.12810305,\n",
      " 'Loss/total_loss': 0.26826,\n",
      " 'learning_rate': 0.077538736}\n",
      "INFO:tensorflow:Step 6600 per-step time 0.830s\n",
      "I1019 21:12:09.569257 4682925568 model_lib_v2.py:705] Step 6600 per-step time 0.830s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.20569907,\n",
      " 'Loss/localization_loss': 0.20237815,\n",
      " 'Loss/regularization_loss': 0.12768811,\n",
      " 'Loss/total_loss': 0.53576535,\n",
      " 'learning_rate': 0.07744939}\n",
      "I1019 21:12:09.569562 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.20569907,\n",
      " 'Loss/localization_loss': 0.20237815,\n",
      " 'Loss/regularization_loss': 0.12768811,\n",
      " 'Loss/total_loss': 0.53576535,\n",
      " 'learning_rate': 0.07744939}\n",
      "INFO:tensorflow:Step 6700 per-step time 0.832s\n",
      "I1019 21:13:32.733552 4682925568 model_lib_v2.py:705] Step 6700 per-step time 0.832s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.14684011,\n",
      " 'Loss/localization_loss': 0.071483344,\n",
      " 'Loss/regularization_loss': 0.12721106,\n",
      " 'Loss/total_loss': 0.3455345,\n",
      " 'learning_rate': 0.077358514}\n",
      "I1019 21:13:32.733788 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.14684011,\n",
      " 'Loss/localization_loss': 0.071483344,\n",
      " 'Loss/regularization_loss': 0.12721106,\n",
      " 'Loss/total_loss': 0.3455345,\n",
      " 'learning_rate': 0.077358514}\n",
      "INFO:tensorflow:Step 6800 per-step time 0.820s\n",
      "I1019 21:14:54.742162 4682925568 model_lib_v2.py:705] Step 6800 per-step time 0.820s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.09970892,\n",
      " 'Loss/localization_loss': 0.06504965,\n",
      " 'Loss/regularization_loss': 0.12677406,\n",
      " 'Loss/total_loss': 0.29153264,\n",
      " 'learning_rate': 0.0772661}\n",
      "I1019 21:14:54.742395 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.09970892,\n",
      " 'Loss/localization_loss': 0.06504965,\n",
      " 'Loss/regularization_loss': 0.12677406,\n",
      " 'Loss/total_loss': 0.29153264,\n",
      " 'learning_rate': 0.0772661}\n",
      "INFO:tensorflow:Step 6900 per-step time 0.838s\n",
      "I1019 21:16:18.500448 4682925568 model_lib_v2.py:705] Step 6900 per-step time 0.838s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.12411801,\n",
      " 'Loss/localization_loss': 0.051206212,\n",
      " 'Loss/regularization_loss': 0.12625039,\n",
      " 'Loss/total_loss': 0.30157462,\n",
      " 'learning_rate': 0.077172145}\n",
      "I1019 21:16:18.503557 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.12411801,\n",
      " 'Loss/localization_loss': 0.051206212,\n",
      " 'Loss/regularization_loss': 0.12625039,\n",
      " 'Loss/total_loss': 0.30157462,\n",
      " 'learning_rate': 0.077172145}\n",
      "INFO:tensorflow:Step 7000 per-step time 0.878s\n",
      "I1019 21:17:46.275090 4682925568 model_lib_v2.py:705] Step 7000 per-step time 0.878s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.1433274,\n",
      " 'Loss/localization_loss': 0.06277574,\n",
      " 'Loss/regularization_loss': 0.12573662,\n",
      " 'Loss/total_loss': 0.33183974,\n",
      " 'learning_rate': 0.07707667}\n",
      "I1019 21:17:46.277817 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.1433274,\n",
      " 'Loss/localization_loss': 0.06277574,\n",
      " 'Loss/regularization_loss': 0.12573662,\n",
      " 'Loss/total_loss': 0.33183974,\n",
      " 'learning_rate': 0.07707667}\n",
      "INFO:tensorflow:Step 7100 per-step time 0.930s\n",
      "I1019 21:19:19.271461 4682925568 model_lib_v2.py:705] Step 7100 per-step time 0.930s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.13897517,\n",
      " 'Loss/localization_loss': 0.04735563,\n",
      " 'Loss/regularization_loss': 0.12540756,\n",
      " 'Loss/total_loss': 0.31173837,\n",
      " 'learning_rate': 0.07697967}\n",
      "I1019 21:19:19.275322 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.13897517,\n",
      " 'Loss/localization_loss': 0.04735563,\n",
      " 'Loss/regularization_loss': 0.12540756,\n",
      " 'Loss/total_loss': 0.31173837,\n",
      " 'learning_rate': 0.07697967}\n",
      "INFO:tensorflow:Step 7200 per-step time 0.816s\n",
      "I1019 21:20:40.843574 4682925568 model_lib_v2.py:705] Step 7200 per-step time 0.816s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08134881,\n",
      " 'Loss/localization_loss': 0.057111118,\n",
      " 'Loss/regularization_loss': 0.12489884,\n",
      " 'Loss/total_loss': 0.26335877,\n",
      " 'learning_rate': 0.07688115}\n",
      "I1019 21:20:40.844660 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08134881,\n",
      " 'Loss/localization_loss': 0.057111118,\n",
      " 'Loss/regularization_loss': 0.12489884,\n",
      " 'Loss/total_loss': 0.26335877,\n",
      " 'learning_rate': 0.07688115}\n",
      "INFO:tensorflow:Step 7300 per-step time 0.815s\n",
      "I1019 21:22:02.357347 4682925568 model_lib_v2.py:705] Step 7300 per-step time 0.815s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.13761689,\n",
      " 'Loss/localization_loss': 0.08108266,\n",
      " 'Loss/regularization_loss': 0.124409795,\n",
      " 'Loss/total_loss': 0.34310934,\n",
      " 'learning_rate': 0.07678111}\n",
      "I1019 21:22:02.357586 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.13761689,\n",
      " 'Loss/localization_loss': 0.08108266,\n",
      " 'Loss/regularization_loss': 0.124409795,\n",
      " 'Loss/total_loss': 0.34310934,\n",
      " 'learning_rate': 0.07678111}\n",
      "INFO:tensorflow:Step 7400 per-step time 0.833s\n",
      "I1019 21:23:25.701858 4682925568 model_lib_v2.py:705] Step 7400 per-step time 0.833s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.09266747,\n",
      " 'Loss/localization_loss': 0.031329293,\n",
      " 'Loss/regularization_loss': 0.123982325,\n",
      " 'Loss/total_loss': 0.24797909,\n",
      " 'learning_rate': 0.076679565}\n",
      "I1019 21:23:25.702192 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.09266747,\n",
      " 'Loss/localization_loss': 0.031329293,\n",
      " 'Loss/regularization_loss': 0.123982325,\n",
      " 'Loss/total_loss': 0.24797909,\n",
      " 'learning_rate': 0.076679565}\n",
      "INFO:tensorflow:Step 7500 per-step time 0.896s\n",
      "I1019 21:24:55.332235 4682925568 model_lib_v2.py:705] Step 7500 per-step time 0.896s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06012171,\n",
      " 'Loss/localization_loss': 0.029265331,\n",
      " 'Loss/regularization_loss': 0.123578265,\n",
      " 'Loss/total_loss': 0.21296531,\n",
      " 'learning_rate': 0.0765765}\n",
      "I1019 21:24:55.336655 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06012171,\n",
      " 'Loss/localization_loss': 0.029265331,\n",
      " 'Loss/regularization_loss': 0.123578265,\n",
      " 'Loss/total_loss': 0.21296531,\n",
      " 'learning_rate': 0.0765765}\n",
      "INFO:tensorflow:Step 7600 per-step time 0.833s\n",
      "I1019 21:26:18.645797 4682925568 model_lib_v2.py:705] Step 7600 per-step time 0.833s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.12770373,\n",
      " 'Loss/localization_loss': 0.044668622,\n",
      " 'Loss/regularization_loss': 0.12306009,\n",
      " 'Loss/total_loss': 0.29543245,\n",
      " 'learning_rate': 0.07647194}\n",
      "I1019 21:26:18.646708 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.12770373,\n",
      " 'Loss/localization_loss': 0.044668622,\n",
      " 'Loss/regularization_loss': 0.12306009,\n",
      " 'Loss/total_loss': 0.29543245,\n",
      " 'learning_rate': 0.07647194}\n",
      "INFO:tensorflow:Step 7700 per-step time 0.820s\n",
      "I1019 21:27:40.622437 4682925568 model_lib_v2.py:705] Step 7700 per-step time 0.820s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.100149736,\n",
      " 'Loss/localization_loss': 0.021937042,\n",
      " 'Loss/regularization_loss': 0.12257561,\n",
      " 'Loss/total_loss': 0.24466239,\n",
      " 'learning_rate': 0.07636588}\n",
      "I1019 21:27:40.623502 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.100149736,\n",
      " 'Loss/localization_loss': 0.021937042,\n",
      " 'Loss/regularization_loss': 0.12257561,\n",
      " 'Loss/total_loss': 0.24466239,\n",
      " 'learning_rate': 0.07636588}\n",
      "INFO:tensorflow:Step 7800 per-step time 0.857s\n",
      "I1019 21:29:06.355031 4682925568 model_lib_v2.py:705] Step 7800 per-step time 0.857s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.1092891,\n",
      " 'Loss/localization_loss': 0.026261522,\n",
      " 'Loss/regularization_loss': 0.122068375,\n",
      " 'Loss/total_loss': 0.257619,\n",
      " 'learning_rate': 0.07625833}\n",
      "I1019 21:29:06.357342 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.1092891,\n",
      " 'Loss/localization_loss': 0.026261522,\n",
      " 'Loss/regularization_loss': 0.122068375,\n",
      " 'Loss/total_loss': 0.257619,\n",
      " 'learning_rate': 0.07625833}\n",
      "INFO:tensorflow:Step 7900 per-step time 0.831s\n",
      "I1019 21:30:29.476879 4682925568 model_lib_v2.py:705] Step 7900 per-step time 0.831s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.088429645,\n",
      " 'Loss/localization_loss': 0.020741545,\n",
      " 'Loss/regularization_loss': 0.12174362,\n",
      " 'Loss/total_loss': 0.23091482,\n",
      " 'learning_rate': 0.07614928}\n",
      "I1019 21:30:29.477815 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.088429645,\n",
      " 'Loss/localization_loss': 0.020741545,\n",
      " 'Loss/regularization_loss': 0.12174362,\n",
      " 'Loss/total_loss': 0.23091482,\n",
      " 'learning_rate': 0.07614928}\n",
      "INFO:tensorflow:Step 8000 per-step time 0.830s\n",
      "I1019 21:31:52.499110 4682925568 model_lib_v2.py:705] Step 8000 per-step time 0.830s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.17087342,\n",
      " 'Loss/localization_loss': 0.04334192,\n",
      " 'Loss/regularization_loss': 0.121680796,\n",
      " 'Loss/total_loss': 0.33589613,\n",
      " 'learning_rate': 0.07603875}\n",
      "I1019 21:31:52.499930 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.17087342,\n",
      " 'Loss/localization_loss': 0.04334192,\n",
      " 'Loss/regularization_loss': 0.121680796,\n",
      " 'Loss/total_loss': 0.33589613,\n",
      " 'learning_rate': 0.07603875}\n",
      "INFO:tensorflow:Step 8100 per-step time 0.911s\n",
      "I1019 21:33:23.591331 4682925568 model_lib_v2.py:705] Step 8100 per-step time 0.911s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0969913,\n",
      " 'Loss/localization_loss': 0.032781195,\n",
      " 'Loss/regularization_loss': 0.121454805,\n",
      " 'Loss/total_loss': 0.25122732,\n",
      " 'learning_rate': 0.07592674}\n",
      "I1019 21:33:23.595245 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0969913,\n",
      " 'Loss/localization_loss': 0.032781195,\n",
      " 'Loss/regularization_loss': 0.121454805,\n",
      " 'Loss/total_loss': 0.25122732,\n",
      " 'learning_rate': 0.07592674}\n",
      "INFO:tensorflow:Step 8200 per-step time 0.853s\n",
      "I1019 21:34:48.888145 4682925568 model_lib_v2.py:705] Step 8200 per-step time 0.853s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.081730284,\n",
      " 'Loss/localization_loss': 0.05012145,\n",
      " 'Loss/regularization_loss': 0.12093449,\n",
      " 'Loss/total_loss': 0.25278622,\n",
      " 'learning_rate': 0.075813256}\n",
      "I1019 21:34:48.890398 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.081730284,\n",
      " 'Loss/localization_loss': 0.05012145,\n",
      " 'Loss/regularization_loss': 0.12093449,\n",
      " 'Loss/total_loss': 0.25278622,\n",
      " 'learning_rate': 0.075813256}\n",
      "INFO:tensorflow:Step 8300 per-step time 0.846s\n",
      "I1019 21:36:13.472980 4682925568 model_lib_v2.py:705] Step 8300 per-step time 0.846s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.121902555,\n",
      " 'Loss/localization_loss': 0.021611702,\n",
      " 'Loss/regularization_loss': 0.12053364,\n",
      " 'Loss/total_loss': 0.2640479,\n",
      " 'learning_rate': 0.07569829}\n",
      "I1019 21:36:13.473237 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.121902555,\n",
      " 'Loss/localization_loss': 0.021611702,\n",
      " 'Loss/regularization_loss': 0.12053364,\n",
      " 'Loss/total_loss': 0.2640479,\n",
      " 'learning_rate': 0.07569829}\n",
      "INFO:tensorflow:Step 8400 per-step time 0.896s\n",
      "I1019 21:37:43.069555 4682925568 model_lib_v2.py:705] Step 8400 per-step time 0.896s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.12927672,\n",
      " 'Loss/localization_loss': 0.0702494,\n",
      " 'Loss/regularization_loss': 0.120214894,\n",
      " 'Loss/total_loss': 0.319741,\n",
      " 'learning_rate': 0.07558186}\n",
      "I1019 21:37:43.071962 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.12927672,\n",
      " 'Loss/localization_loss': 0.0702494,\n",
      " 'Loss/regularization_loss': 0.120214894,\n",
      " 'Loss/total_loss': 0.319741,\n",
      " 'learning_rate': 0.07558186}\n",
      "INFO:tensorflow:Step 8500 per-step time 0.838s\n",
      "I1019 21:39:06.808688 4682925568 model_lib_v2.py:705] Step 8500 per-step time 0.838s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.11195433,\n",
      " 'Loss/localization_loss': 0.046466492,\n",
      " 'Loss/regularization_loss': 0.11972945,\n",
      " 'Loss/total_loss': 0.27815026,\n",
      " 'learning_rate': 0.07546397}\n",
      "I1019 21:39:06.809778 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.11195433,\n",
      " 'Loss/localization_loss': 0.046466492,\n",
      " 'Loss/regularization_loss': 0.11972945,\n",
      " 'Loss/total_loss': 0.27815026,\n",
      " 'learning_rate': 0.07546397}\n",
      "INFO:tensorflow:Step 8600 per-step time 0.835s\n",
      "I1019 21:40:30.318659 4682925568 model_lib_v2.py:705] Step 8600 per-step time 0.835s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07180731,\n",
      " 'Loss/localization_loss': 0.027239706,\n",
      " 'Loss/regularization_loss': 0.119264066,\n",
      " 'Loss/total_loss': 0.21831109,\n",
      " 'learning_rate': 0.075344615}\n",
      "I1019 21:40:30.320888 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07180731,\n",
      " 'Loss/localization_loss': 0.027239706,\n",
      " 'Loss/regularization_loss': 0.119264066,\n",
      " 'Loss/total_loss': 0.21831109,\n",
      " 'learning_rate': 0.075344615}\n",
      "INFO:tensorflow:Step 8700 per-step time 0.876s\n",
      "I1019 21:41:57.879525 4682925568 model_lib_v2.py:705] Step 8700 per-step time 0.876s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08355874,\n",
      " 'Loss/localization_loss': 0.031746447,\n",
      " 'Loss/regularization_loss': 0.11873453,\n",
      " 'Loss/total_loss': 0.23403972,\n",
      " 'learning_rate': 0.07522382}\n",
      "I1019 21:41:57.879785 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08355874,\n",
      " 'Loss/localization_loss': 0.031746447,\n",
      " 'Loss/regularization_loss': 0.11873453,\n",
      " 'Loss/total_loss': 0.23403972,\n",
      " 'learning_rate': 0.07522382}\n",
      "INFO:tensorflow:Step 8800 per-step time 0.872s\n",
      "I1019 21:43:25.093518 4682925568 model_lib_v2.py:705] Step 8800 per-step time 0.872s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.068518944,\n",
      " 'Loss/localization_loss': 0.042790037,\n",
      " 'Loss/regularization_loss': 0.11829723,\n",
      " 'Loss/total_loss': 0.22960621,\n",
      " 'learning_rate': 0.07510157}\n",
      "I1019 21:43:25.093760 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.068518944,\n",
      " 'Loss/localization_loss': 0.042790037,\n",
      " 'Loss/regularization_loss': 0.11829723,\n",
      " 'Loss/total_loss': 0.22960621,\n",
      " 'learning_rate': 0.07510157}\n",
      "INFO:tensorflow:Step 8900 per-step time 0.894s\n",
      "I1019 21:44:54.545082 4682925568 model_lib_v2.py:705] Step 8900 per-step time 0.894s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.10828966,\n",
      " 'Loss/localization_loss': 0.086433135,\n",
      " 'Loss/regularization_loss': 0.11783343,\n",
      " 'Loss/total_loss': 0.3125562,\n",
      " 'learning_rate': 0.074977875}\n",
      "I1019 21:44:54.547518 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.10828966,\n",
      " 'Loss/localization_loss': 0.086433135,\n",
      " 'Loss/regularization_loss': 0.11783343,\n",
      " 'Loss/total_loss': 0.3125562,\n",
      " 'learning_rate': 0.074977875}\n",
      "INFO:tensorflow:Step 9000 per-step time 0.932s\n",
      "I1019 21:46:27.741443 4682925568 model_lib_v2.py:705] Step 9000 per-step time 0.932s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.10680505,\n",
      " 'Loss/localization_loss': 0.029891592,\n",
      " 'Loss/regularization_loss': 0.11742584,\n",
      " 'Loss/total_loss': 0.25412247,\n",
      " 'learning_rate': 0.07485275}\n",
      "I1019 21:46:27.743700 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.10680505,\n",
      " 'Loss/localization_loss': 0.029891592,\n",
      " 'Loss/regularization_loss': 0.11742584,\n",
      " 'Loss/total_loss': 0.25412247,\n",
      " 'learning_rate': 0.07485275}\n",
      "INFO:tensorflow:Step 9100 per-step time 0.950s\n",
      "I1019 21:48:02.744796 4682925568 model_lib_v2.py:705] Step 9100 per-step time 0.950s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.16006148,\n",
      " 'Loss/localization_loss': 0.027875926,\n",
      " 'Loss/regularization_loss': 0.11711388,\n",
      " 'Loss/total_loss': 0.30505127,\n",
      " 'learning_rate': 0.07472619}\n",
      "I1019 21:48:02.747984 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.16006148,\n",
      " 'Loss/localization_loss': 0.027875926,\n",
      " 'Loss/regularization_loss': 0.11711388,\n",
      " 'Loss/total_loss': 0.30505127,\n",
      " 'learning_rate': 0.07472619}\n",
      "INFO:tensorflow:Step 9200 per-step time 0.878s\n",
      "I1019 21:49:30.535877 4682925568 model_lib_v2.py:705] Step 9200 per-step time 0.878s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.13760386,\n",
      " 'Loss/localization_loss': 0.06167852,\n",
      " 'Loss/regularization_loss': 0.1168074,\n",
      " 'Loss/total_loss': 0.3160898,\n",
      " 'learning_rate': 0.07459819}\n",
      "I1019 21:49:30.536112 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.13760386,\n",
      " 'Loss/localization_loss': 0.06167852,\n",
      " 'Loss/regularization_loss': 0.1168074,\n",
      " 'Loss/total_loss': 0.3160898,\n",
      " 'learning_rate': 0.07459819}\n",
      "INFO:tensorflow:Step 9300 per-step time 0.827s\n",
      "I1019 21:50:53.201370 4682925568 model_lib_v2.py:705] Step 9300 per-step time 0.827s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.10384738,\n",
      " 'Loss/localization_loss': 0.026990116,\n",
      " 'Loss/regularization_loss': 0.116339035,\n",
      " 'Loss/total_loss': 0.24717653,\n",
      " 'learning_rate': 0.074468784}\n",
      "I1019 21:50:53.203639 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.10384738,\n",
      " 'Loss/localization_loss': 0.026990116,\n",
      " 'Loss/regularization_loss': 0.116339035,\n",
      " 'Loss/total_loss': 0.24717653,\n",
      " 'learning_rate': 0.074468784}\n",
      "INFO:tensorflow:Step 9400 per-step time 0.832s\n",
      "I1019 21:52:16.337869 4682925568 model_lib_v2.py:705] Step 9400 per-step time 0.832s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.16471614,\n",
      " 'Loss/localization_loss': 0.043611716,\n",
      " 'Loss/regularization_loss': 0.11613705,\n",
      " 'Loss/total_loss': 0.32446492,\n",
      " 'learning_rate': 0.074337944}\n",
      "I1019 21:52:16.338140 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.16471614,\n",
      " 'Loss/localization_loss': 0.043611716,\n",
      " 'Loss/regularization_loss': 0.11613705,\n",
      " 'Loss/total_loss': 0.32446492,\n",
      " 'learning_rate': 0.074337944}\n",
      "INFO:tensorflow:Step 9500 per-step time 0.861s\n",
      "I1019 21:53:42.422389 4682925568 model_lib_v2.py:705] Step 9500 per-step time 0.861s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.12554817,\n",
      " 'Loss/localization_loss': 0.12386387,\n",
      " 'Loss/regularization_loss': 0.11570708,\n",
      " 'Loss/total_loss': 0.3651191,\n",
      " 'learning_rate': 0.074205704}\n",
      "I1019 21:53:42.424650 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.12554817,\n",
      " 'Loss/localization_loss': 0.12386387,\n",
      " 'Loss/regularization_loss': 0.11570708,\n",
      " 'Loss/total_loss': 0.3651191,\n",
      " 'learning_rate': 0.074205704}\n",
      "INFO:tensorflow:Step 9600 per-step time 0.840s\n",
      "I1019 21:55:06.448041 4682925568 model_lib_v2.py:705] Step 9600 per-step time 0.840s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.10016337,\n",
      " 'Loss/localization_loss': 0.08816961,\n",
      " 'Loss/regularization_loss': 0.11520242,\n",
      " 'Loss/total_loss': 0.3035354,\n",
      " 'learning_rate': 0.07407206}\n",
      "I1019 21:55:06.450279 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.10016337,\n",
      " 'Loss/localization_loss': 0.08816961,\n",
      " 'Loss/regularization_loss': 0.11520242,\n",
      " 'Loss/total_loss': 0.3035354,\n",
      " 'learning_rate': 0.07407206}\n",
      "INFO:tensorflow:Step 9700 per-step time 0.847s\n",
      "I1019 21:56:31.147566 4682925568 model_lib_v2.py:705] Step 9700 per-step time 0.847s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.16951409,\n",
      " 'Loss/localization_loss': 0.091153435,\n",
      " 'Loss/regularization_loss': 0.114887,\n",
      " 'Loss/total_loss': 0.37555453,\n",
      " 'learning_rate': 0.073937014}\n",
      "I1019 21:56:31.150047 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.16951409,\n",
      " 'Loss/localization_loss': 0.091153435,\n",
      " 'Loss/regularization_loss': 0.114887,\n",
      " 'Loss/total_loss': 0.37555453,\n",
      " 'learning_rate': 0.073937014}\n",
      "INFO:tensorflow:Step 9800 per-step time 0.828s\n",
      "I1019 21:57:53.956970 4682925568 model_lib_v2.py:705] Step 9800 per-step time 0.828s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.16173647,\n",
      " 'Loss/localization_loss': 0.051054373,\n",
      " 'Loss/regularization_loss': 0.11451941,\n",
      " 'Loss/total_loss': 0.32731026,\n",
      " 'learning_rate': 0.07380057}\n",
      "I1019 21:57:53.957216 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.16173647,\n",
      " 'Loss/localization_loss': 0.051054373,\n",
      " 'Loss/regularization_loss': 0.11451941,\n",
      " 'Loss/total_loss': 0.32731026,\n",
      " 'learning_rate': 0.07380057}\n",
      "INFO:tensorflow:Step 9900 per-step time 0.839s\n",
      "I1019 21:59:17.868901 4682925568 model_lib_v2.py:705] Step 9900 per-step time 0.839s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.077453025,\n",
      " 'Loss/localization_loss': 0.03739703,\n",
      " 'Loss/regularization_loss': 0.114223965,\n",
      " 'Loss/total_loss': 0.22907403,\n",
      " 'learning_rate': 0.073662736}\n",
      "I1019 21:59:17.871378 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.077453025,\n",
      " 'Loss/localization_loss': 0.03739703,\n",
      " 'Loss/regularization_loss': 0.114223965,\n",
      " 'Loss/total_loss': 0.22907403,\n",
      " 'learning_rate': 0.073662736}\n",
      "INFO:tensorflow:Step 10000 per-step time 0.856s\n",
      "I1019 22:00:43.470952 4682925568 model_lib_v2.py:705] Step 10000 per-step time 0.856s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.1372298,\n",
      " 'Loss/localization_loss': 0.08199541,\n",
      " 'Loss/regularization_loss': 0.11389138,\n",
      " 'Loss/total_loss': 0.3331166,\n",
      " 'learning_rate': 0.07352352}\n",
      "I1019 22:00:43.471897 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.1372298,\n",
      " 'Loss/localization_loss': 0.08199541,\n",
      " 'Loss/regularization_loss': 0.11389138,\n",
      " 'Loss/total_loss': 0.3331166,\n",
      " 'learning_rate': 0.07352352}\n",
      "INFO:tensorflow:Step 10100 per-step time 0.918s\n",
      "I1019 22:02:15.319356 4682925568 model_lib_v2.py:705] Step 10100 per-step time 0.918s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.064717114,\n",
      " 'Loss/localization_loss': 0.047907304,\n",
      " 'Loss/regularization_loss': 0.11347833,\n",
      " 'Loss/total_loss': 0.22610275,\n",
      " 'learning_rate': 0.07338293}\n",
      "I1019 22:02:15.321846 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.064717114,\n",
      " 'Loss/localization_loss': 0.047907304,\n",
      " 'Loss/regularization_loss': 0.11347833,\n",
      " 'Loss/total_loss': 0.22610275,\n",
      " 'learning_rate': 0.07338293}\n",
      "INFO:tensorflow:Step 10200 per-step time 0.856s\n",
      "I1019 22:03:40.944741 4682925568 model_lib_v2.py:705] Step 10200 per-step time 0.856s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.110561974,\n",
      " 'Loss/localization_loss': 0.022193529,\n",
      " 'Loss/regularization_loss': 0.112994164,\n",
      " 'Loss/total_loss': 0.24574967,\n",
      " 'learning_rate': 0.073240966}\n",
      "I1019 22:03:40.948026 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.110561974,\n",
      " 'Loss/localization_loss': 0.022193529,\n",
      " 'Loss/regularization_loss': 0.112994164,\n",
      " 'Loss/total_loss': 0.24574967,\n",
      " 'learning_rate': 0.073240966}\n",
      "INFO:tensorflow:Step 10300 per-step time 0.837s\n",
      "I1019 22:05:04.650300 4682925568 model_lib_v2.py:705] Step 10300 per-step time 0.837s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.068547204,\n",
      " 'Loss/localization_loss': 0.033539247,\n",
      " 'Loss/regularization_loss': 0.112528846,\n",
      " 'Loss/total_loss': 0.2146153,\n",
      " 'learning_rate': 0.07309763}\n",
      "I1019 22:05:04.650544 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.068547204,\n",
      " 'Loss/localization_loss': 0.033539247,\n",
      " 'Loss/regularization_loss': 0.112528846,\n",
      " 'Loss/total_loss': 0.2146153,\n",
      " 'learning_rate': 0.07309763}\n",
      "INFO:tensorflow:Step 10400 per-step time 0.826s\n",
      "I1019 22:06:27.254173 4682925568 model_lib_v2.py:705] Step 10400 per-step time 0.826s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05882235,\n",
      " 'Loss/localization_loss': 0.03960629,\n",
      " 'Loss/regularization_loss': 0.1120386,\n",
      " 'Loss/total_loss': 0.21046723,\n",
      " 'learning_rate': 0.07295293}\n",
      "I1019 22:06:27.254430 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05882235,\n",
      " 'Loss/localization_loss': 0.03960629,\n",
      " 'Loss/regularization_loss': 0.1120386,\n",
      " 'Loss/total_loss': 0.21046723,\n",
      " 'learning_rate': 0.07295293}\n",
      "INFO:tensorflow:Step 10500 per-step time 0.829s\n",
      "I1019 22:07:50.117386 4682925568 model_lib_v2.py:705] Step 10500 per-step time 0.829s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.059104342,\n",
      " 'Loss/localization_loss': 0.029854959,\n",
      " 'Loss/regularization_loss': 0.11157289,\n",
      " 'Loss/total_loss': 0.2005322,\n",
      " 'learning_rate': 0.07280689}\n",
      "I1019 22:07:50.117621 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.059104342,\n",
      " 'Loss/localization_loss': 0.029854959,\n",
      " 'Loss/regularization_loss': 0.11157289,\n",
      " 'Loss/total_loss': 0.2005322,\n",
      " 'learning_rate': 0.07280689}\n",
      "INFO:tensorflow:Step 10600 per-step time 0.843s\n",
      "I1019 22:09:14.378139 4682925568 model_lib_v2.py:705] Step 10600 per-step time 0.843s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.09668045,\n",
      " 'Loss/localization_loss': 0.019242037,\n",
      " 'Loss/regularization_loss': 0.11110813,\n",
      " 'Loss/total_loss': 0.2270306,\n",
      " 'learning_rate': 0.07265949}\n",
      "I1019 22:09:14.380418 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.09668045,\n",
      " 'Loss/localization_loss': 0.019242037,\n",
      " 'Loss/regularization_loss': 0.11110813,\n",
      " 'Loss/total_loss': 0.2270306,\n",
      " 'learning_rate': 0.07265949}\n",
      "INFO:tensorflow:Step 10700 per-step time 0.889s\n",
      "I1019 22:10:43.258615 4682925568 model_lib_v2.py:705] Step 10700 per-step time 0.889s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.09714031,\n",
      " 'Loss/localization_loss': 0.047008578,\n",
      " 'Loss/regularization_loss': 0.110765845,\n",
      " 'Loss/total_loss': 0.25491473,\n",
      " 'learning_rate': 0.07251076}\n",
      "I1019 22:10:43.258975 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.09714031,\n",
      " 'Loss/localization_loss': 0.047008578,\n",
      " 'Loss/regularization_loss': 0.110765845,\n",
      " 'Loss/total_loss': 0.25491473,\n",
      " 'learning_rate': 0.07251076}\n",
      "INFO:tensorflow:Step 10800 per-step time 0.857s\n",
      "I1019 22:12:08.990087 4682925568 model_lib_v2.py:705] Step 10800 per-step time 0.857s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.2693747,\n",
      " 'Loss/localization_loss': 0.047194228,\n",
      " 'Loss/regularization_loss': 0.11092593,\n",
      " 'Loss/total_loss': 0.42749482,\n",
      " 'learning_rate': 0.07236068}\n",
      "I1019 22:12:08.990426 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.2693747,\n",
      " 'Loss/localization_loss': 0.047194228,\n",
      " 'Loss/regularization_loss': 0.11092593,\n",
      " 'Loss/total_loss': 0.42749482,\n",
      " 'learning_rate': 0.07236068}\n",
      "INFO:tensorflow:Step 10900 per-step time 0.813s\n",
      "I1019 22:13:30.236195 4682925568 model_lib_v2.py:705] Step 10900 per-step time 0.813s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.1399087,\n",
      " 'Loss/localization_loss': 0.048225775,\n",
      " 'Loss/regularization_loss': 0.11056034,\n",
      " 'Loss/total_loss': 0.29869482,\n",
      " 'learning_rate': 0.07220927}\n",
      "I1019 22:13:30.236431 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.1399087,\n",
      " 'Loss/localization_loss': 0.048225775,\n",
      " 'Loss/regularization_loss': 0.11056034,\n",
      " 'Loss/total_loss': 0.29869482,\n",
      " 'learning_rate': 0.07220927}\n",
      "INFO:tensorflow:Step 11000 per-step time 0.817s\n",
      "I1019 22:14:51.928580 4682925568 model_lib_v2.py:705] Step 11000 per-step time 0.817s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.10658219,\n",
      " 'Loss/localization_loss': 0.031162906,\n",
      " 'Loss/regularization_loss': 0.11010165,\n",
      " 'Loss/total_loss': 0.24784675,\n",
      " 'learning_rate': 0.07205654}\n",
      "I1019 22:14:51.928946 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.10658219,\n",
      " 'Loss/localization_loss': 0.031162906,\n",
      " 'Loss/regularization_loss': 0.11010165,\n",
      " 'Loss/total_loss': 0.24784675,\n",
      " 'learning_rate': 0.07205654}\n",
      "INFO:tensorflow:Step 11100 per-step time 0.853s\n",
      "I1019 22:16:17.268154 4682925568 model_lib_v2.py:705] Step 11100 per-step time 0.853s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.14869983,\n",
      " 'Loss/localization_loss': 0.039040454,\n",
      " 'Loss/regularization_loss': 0.1096761,\n",
      " 'Loss/total_loss': 0.2974164,\n",
      " 'learning_rate': 0.07190249}\n",
      "I1019 22:16:17.268393 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.14869983,\n",
      " 'Loss/localization_loss': 0.039040454,\n",
      " 'Loss/regularization_loss': 0.1096761,\n",
      " 'Loss/total_loss': 0.2974164,\n",
      " 'learning_rate': 0.07190249}\n",
      "INFO:tensorflow:Step 11200 per-step time 0.847s\n",
      "I1019 22:17:42.006600 4682925568 model_lib_v2.py:705] Step 11200 per-step time 0.847s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.092829436,\n",
      " 'Loss/localization_loss': 0.022748554,\n",
      " 'Loss/regularization_loss': 0.10920985,\n",
      " 'Loss/total_loss': 0.22478783,\n",
      " 'learning_rate': 0.07174714}\n",
      "I1019 22:17:42.008981 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.092829436,\n",
      " 'Loss/localization_loss': 0.022748554,\n",
      " 'Loss/regularization_loss': 0.10920985,\n",
      " 'Loss/total_loss': 0.22478783,\n",
      " 'learning_rate': 0.07174714}\n",
      "INFO:tensorflow:Step 11300 per-step time 0.842s\n",
      "I1019 22:19:06.192101 4682925568 model_lib_v2.py:705] Step 11300 per-step time 0.842s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0913843,\n",
      " 'Loss/localization_loss': 0.03335189,\n",
      " 'Loss/regularization_loss': 0.10882433,\n",
      " 'Loss/total_loss': 0.23356052,\n",
      " 'learning_rate': 0.071590476}\n",
      "I1019 22:19:06.193998 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0913843,\n",
      " 'Loss/localization_loss': 0.03335189,\n",
      " 'Loss/regularization_loss': 0.10882433,\n",
      " 'Loss/total_loss': 0.23356052,\n",
      " 'learning_rate': 0.071590476}\n",
      "INFO:tensorflow:Step 11400 per-step time 0.839s\n",
      "I1019 22:20:30.063467 4682925568 model_lib_v2.py:705] Step 11400 per-step time 0.839s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.070803024,\n",
      " 'Loss/localization_loss': 0.041584123,\n",
      " 'Loss/regularization_loss': 0.10834281,\n",
      " 'Loss/total_loss': 0.22072996,\n",
      " 'learning_rate': 0.071432516}\n",
      "I1019 22:20:30.065313 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.070803024,\n",
      " 'Loss/localization_loss': 0.041584123,\n",
      " 'Loss/regularization_loss': 0.10834281,\n",
      " 'Loss/total_loss': 0.22072996,\n",
      " 'learning_rate': 0.071432516}\n",
      "INFO:tensorflow:Step 11500 per-step time 0.826s\n",
      "I1019 22:21:52.643935 4682925568 model_lib_v2.py:705] Step 11500 per-step time 0.826s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.087306075,\n",
      " 'Loss/localization_loss': 0.043698754,\n",
      " 'Loss/regularization_loss': 0.107906766,\n",
      " 'Loss/total_loss': 0.2389116,\n",
      " 'learning_rate': 0.07127326}\n",
      "I1019 22:21:52.647217 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.087306075,\n",
      " 'Loss/localization_loss': 0.043698754,\n",
      " 'Loss/regularization_loss': 0.107906766,\n",
      " 'Loss/total_loss': 0.2389116,\n",
      " 'learning_rate': 0.07127326}\n",
      "INFO:tensorflow:Step 11600 per-step time 0.815s\n",
      "I1019 22:23:14.113210 4682925568 model_lib_v2.py:705] Step 11600 per-step time 0.815s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.13342756,\n",
      " 'Loss/localization_loss': 0.033493657,\n",
      " 'Loss/regularization_loss': 0.10751127,\n",
      " 'Loss/total_loss': 0.27443248,\n",
      " 'learning_rate': 0.071112715}\n",
      "I1019 22:23:14.115226 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.13342756,\n",
      " 'Loss/localization_loss': 0.033493657,\n",
      " 'Loss/regularization_loss': 0.10751127,\n",
      " 'Loss/total_loss': 0.27443248,\n",
      " 'learning_rate': 0.071112715}\n",
      "INFO:tensorflow:Step 11700 per-step time 0.828s\n",
      "I1019 22:24:36.934688 4682925568 model_lib_v2.py:705] Step 11700 per-step time 0.828s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.122701354,\n",
      " 'Loss/localization_loss': 0.083435096,\n",
      " 'Loss/regularization_loss': 0.107138865,\n",
      " 'Loss/total_loss': 0.3132753,\n",
      " 'learning_rate': 0.070950896}\n",
      "I1019 22:24:36.935066 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.122701354,\n",
      " 'Loss/localization_loss': 0.083435096,\n",
      " 'Loss/regularization_loss': 0.107138865,\n",
      " 'Loss/total_loss': 0.3132753,\n",
      " 'learning_rate': 0.070950896}\n",
      "INFO:tensorflow:Step 11800 per-step time 0.833s\n",
      "I1019 22:26:00.268575 4682925568 model_lib_v2.py:705] Step 11800 per-step time 0.833s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.13395497,\n",
      " 'Loss/localization_loss': 0.073588245,\n",
      " 'Loss/regularization_loss': 0.10676275,\n",
      " 'Loss/total_loss': 0.31430596,\n",
      " 'learning_rate': 0.07078781}\n",
      "I1019 22:26:00.268907 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.13395497,\n",
      " 'Loss/localization_loss': 0.073588245,\n",
      " 'Loss/regularization_loss': 0.10676275,\n",
      " 'Loss/total_loss': 0.31430596,\n",
      " 'learning_rate': 0.07078781}\n",
      "INFO:tensorflow:Step 11900 per-step time 0.841s\n",
      "I1019 22:27:24.425840 4682925568 model_lib_v2.py:705] Step 11900 per-step time 0.841s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.09353165,\n",
      " 'Loss/localization_loss': 0.04269873,\n",
      " 'Loss/regularization_loss': 0.10640224,\n",
      " 'Loss/total_loss': 0.24263263,\n",
      " 'learning_rate': 0.07062345}\n",
      "I1019 22:27:24.428118 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.09353165,\n",
      " 'Loss/localization_loss': 0.04269873,\n",
      " 'Loss/regularization_loss': 0.10640224,\n",
      " 'Loss/total_loss': 0.24263263,\n",
      " 'learning_rate': 0.07062345}\n",
      "INFO:tensorflow:Step 12000 per-step time 0.818s\n",
      "I1019 22:28:46.186700 4682925568 model_lib_v2.py:705] Step 12000 per-step time 0.818s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07733403,\n",
      " 'Loss/localization_loss': 0.044265233,\n",
      " 'Loss/regularization_loss': 0.10611957,\n",
      " 'Loss/total_loss': 0.22771883,\n",
      " 'learning_rate': 0.07045784}\n",
      "I1019 22:28:46.186993 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07733403,\n",
      " 'Loss/localization_loss': 0.044265233,\n",
      " 'Loss/regularization_loss': 0.10611957,\n",
      " 'Loss/total_loss': 0.22771883,\n",
      " 'learning_rate': 0.07045784}\n",
      "INFO:tensorflow:Step 12100 per-step time 0.852s\n",
      "I1019 22:30:11.433570 4682925568 model_lib_v2.py:705] Step 12100 per-step time 0.852s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.047036286,\n",
      " 'Loss/localization_loss': 0.030025687,\n",
      " 'Loss/regularization_loss': 0.10576107,\n",
      " 'Loss/total_loss': 0.18282305,\n",
      " 'learning_rate': 0.07029097}\n",
      "I1019 22:30:11.437519 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.047036286,\n",
      " 'Loss/localization_loss': 0.030025687,\n",
      " 'Loss/regularization_loss': 0.10576107,\n",
      " 'Loss/total_loss': 0.18282305,\n",
      " 'learning_rate': 0.07029097}\n",
      "INFO:tensorflow:Step 12200 per-step time 0.826s\n",
      "I1019 22:31:33.983080 4682925568 model_lib_v2.py:705] Step 12200 per-step time 0.826s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.055096924,\n",
      " 'Loss/localization_loss': 0.016082982,\n",
      " 'Loss/regularization_loss': 0.10554251,\n",
      " 'Loss/total_loss': 0.1767224,\n",
      " 'learning_rate': 0.07012285}\n",
      "I1019 22:31:33.983319 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.055096924,\n",
      " 'Loss/localization_loss': 0.016082982,\n",
      " 'Loss/regularization_loss': 0.10554251,\n",
      " 'Loss/total_loss': 0.1767224,\n",
      " 'learning_rate': 0.07012285}\n",
      "INFO:tensorflow:Step 12300 per-step time 0.827s\n",
      "I1019 22:32:56.699817 4682925568 model_lib_v2.py:705] Step 12300 per-step time 0.827s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.14688803,\n",
      " 'Loss/localization_loss': 0.086512394,\n",
      " 'Loss/regularization_loss': 0.10527962,\n",
      " 'Loss/total_loss': 0.33868006,\n",
      " 'learning_rate': 0.06995351}\n",
      "I1019 22:32:56.700191 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.14688803,\n",
      " 'Loss/localization_loss': 0.086512394,\n",
      " 'Loss/regularization_loss': 0.10527962,\n",
      " 'Loss/total_loss': 0.33868006,\n",
      " 'learning_rate': 0.06995351}\n",
      "INFO:tensorflow:Step 12400 per-step time 0.814s\n",
      "I1019 22:34:18.109369 4682925568 model_lib_v2.py:705] Step 12400 per-step time 0.814s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08186553,\n",
      " 'Loss/localization_loss': 0.03242824,\n",
      " 'Loss/regularization_loss': 0.10504149,\n",
      " 'Loss/total_loss': 0.21933526,\n",
      " 'learning_rate': 0.06978292}\n",
      "I1019 22:34:18.109603 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08186553,\n",
      " 'Loss/localization_loss': 0.03242824,\n",
      " 'Loss/regularization_loss': 0.10504149,\n",
      " 'Loss/total_loss': 0.21933526,\n",
      " 'learning_rate': 0.06978292}\n",
      "INFO:tensorflow:Step 12500 per-step time 0.828s\n",
      "I1019 22:35:40.970584 4682925568 model_lib_v2.py:705] Step 12500 per-step time 0.828s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.09759872,\n",
      " 'Loss/localization_loss': 0.041725665,\n",
      " 'Loss/regularization_loss': 0.104648605,\n",
      " 'Loss/total_loss': 0.24397299,\n",
      " 'learning_rate': 0.06961112}\n",
      "I1019 22:35:40.972879 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.09759872,\n",
      " 'Loss/localization_loss': 0.041725665,\n",
      " 'Loss/regularization_loss': 0.104648605,\n",
      " 'Loss/total_loss': 0.24397299,\n",
      " 'learning_rate': 0.06961112}\n",
      "INFO:tensorflow:Step 12600 per-step time 0.816s\n",
      "I1019 22:37:02.560432 4682925568 model_lib_v2.py:705] Step 12600 per-step time 0.816s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07541756,\n",
      " 'Loss/localization_loss': 0.052273586,\n",
      " 'Loss/regularization_loss': 0.104373194,\n",
      " 'Loss/total_loss': 0.23206434,\n",
      " 'learning_rate': 0.06943809}\n",
      "I1019 22:37:02.560676 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07541756,\n",
      " 'Loss/localization_loss': 0.052273586,\n",
      " 'Loss/regularization_loss': 0.104373194,\n",
      " 'Loss/total_loss': 0.23206434,\n",
      " 'learning_rate': 0.06943809}\n",
      "INFO:tensorflow:Step 12700 per-step time 0.816s\n",
      "I1019 22:38:24.122191 4682925568 model_lib_v2.py:705] Step 12700 per-step time 0.816s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.11665913,\n",
      " 'Loss/localization_loss': 0.033586226,\n",
      " 'Loss/regularization_loss': 0.10406257,\n",
      " 'Loss/total_loss': 0.25430793,\n",
      " 'learning_rate': 0.06926386}\n",
      "I1019 22:38:24.122458 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.11665913,\n",
      " 'Loss/localization_loss': 0.033586226,\n",
      " 'Loss/regularization_loss': 0.10406257,\n",
      " 'Loss/total_loss': 0.25430793,\n",
      " 'learning_rate': 0.06926386}\n",
      "INFO:tensorflow:Step 12800 per-step time 0.823s\n",
      "I1019 22:39:46.466309 4682925568 model_lib_v2.py:705] Step 12800 per-step time 0.823s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08517285,\n",
      " 'Loss/localization_loss': 0.025847608,\n",
      " 'Loss/regularization_loss': 0.10361796,\n",
      " 'Loss/total_loss': 0.21463841,\n",
      " 'learning_rate': 0.06908843}\n",
      "I1019 22:39:46.466620 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08517285,\n",
      " 'Loss/localization_loss': 0.025847608,\n",
      " 'Loss/regularization_loss': 0.10361796,\n",
      " 'Loss/total_loss': 0.21463841,\n",
      " 'learning_rate': 0.06908843}\n",
      "INFO:tensorflow:Step 12900 per-step time 0.860s\n",
      "I1019 22:41:12.522854 4682925568 model_lib_v2.py:705] Step 12900 per-step time 0.860s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.081979565,\n",
      " 'Loss/localization_loss': 0.035645552,\n",
      " 'Loss/regularization_loss': 0.103151456,\n",
      " 'Loss/total_loss': 0.22077657,\n",
      " 'learning_rate': 0.06891179}\n",
      "I1019 22:41:12.526138 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.081979565,\n",
      " 'Loss/localization_loss': 0.035645552,\n",
      " 'Loss/regularization_loss': 0.103151456,\n",
      " 'Loss/total_loss': 0.22077657,\n",
      " 'learning_rate': 0.06891179}\n",
      "INFO:tensorflow:Step 13000 per-step time 0.870s\n",
      "I1019 22:42:39.514816 4682925568 model_lib_v2.py:705] Step 13000 per-step time 0.870s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06676011,\n",
      " 'Loss/localization_loss': 0.040029086,\n",
      " 'Loss/regularization_loss': 0.10272761,\n",
      " 'Loss/total_loss': 0.2095168,\n",
      " 'learning_rate': 0.068733975}\n",
      "I1019 22:42:39.517978 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06676011,\n",
      " 'Loss/localization_loss': 0.040029086,\n",
      " 'Loss/regularization_loss': 0.10272761,\n",
      " 'Loss/total_loss': 0.2095168,\n",
      " 'learning_rate': 0.068733975}\n",
      "INFO:tensorflow:Step 13100 per-step time 0.811s\n",
      "I1019 22:44:00.581176 4682925568 model_lib_v2.py:705] Step 13100 per-step time 0.811s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.2205864,\n",
      " 'Loss/localization_loss': 0.06454786,\n",
      " 'Loss/regularization_loss': 0.10227362,\n",
      " 'Loss/total_loss': 0.38740787,\n",
      " 'learning_rate': 0.068554975}\n",
      "I1019 22:44:00.581459 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.2205864,\n",
      " 'Loss/localization_loss': 0.06454786,\n",
      " 'Loss/regularization_loss': 0.10227362,\n",
      " 'Loss/total_loss': 0.38740787,\n",
      " 'learning_rate': 0.068554975}\n",
      "INFO:tensorflow:Step 13200 per-step time 0.852s\n",
      "I1019 22:45:25.748474 4682925568 model_lib_v2.py:705] Step 13200 per-step time 0.852s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.11647082,\n",
      " 'Loss/localization_loss': 0.06225187,\n",
      " 'Loss/regularization_loss': 0.1018776,\n",
      " 'Loss/total_loss': 0.2806003,\n",
      " 'learning_rate': 0.0683748}\n",
      "I1019 22:45:25.750669 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.11647082,\n",
      " 'Loss/localization_loss': 0.06225187,\n",
      " 'Loss/regularization_loss': 0.1018776,\n",
      " 'Loss/total_loss': 0.2806003,\n",
      " 'learning_rate': 0.0683748}\n",
      "INFO:tensorflow:Step 13300 per-step time 0.879s\n",
      "I1019 22:46:53.614808 4682925568 model_lib_v2.py:705] Step 13300 per-step time 0.879s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.10169787,\n",
      " 'Loss/localization_loss': 0.095495835,\n",
      " 'Loss/regularization_loss': 0.10167197,\n",
      " 'Loss/total_loss': 0.29886568,\n",
      " 'learning_rate': 0.06819345}\n",
      "I1019 22:46:53.617845 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.10169787,\n",
      " 'Loss/localization_loss': 0.095495835,\n",
      " 'Loss/regularization_loss': 0.10167197,\n",
      " 'Loss/total_loss': 0.29886568,\n",
      " 'learning_rate': 0.06819345}\n",
      "INFO:tensorflow:Step 13400 per-step time 0.822s\n",
      "I1019 22:48:15.846612 4682925568 model_lib_v2.py:705] Step 13400 per-step time 0.822s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.16375127,\n",
      " 'Loss/localization_loss': 0.082810916,\n",
      " 'Loss/regularization_loss': 0.10141932,\n",
      " 'Loss/total_loss': 0.3479815,\n",
      " 'learning_rate': 0.06801095}\n",
      "I1019 22:48:15.846857 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.16375127,\n",
      " 'Loss/localization_loss': 0.082810916,\n",
      " 'Loss/regularization_loss': 0.10141932,\n",
      " 'Loss/total_loss': 0.3479815,\n",
      " 'learning_rate': 0.06801095}\n",
      "INFO:tensorflow:Step 13500 per-step time 0.821s\n",
      "I1019 22:49:37.917608 4682925568 model_lib_v2.py:705] Step 13500 per-step time 0.821s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.101886205,\n",
      " 'Loss/localization_loss': 0.089428015,\n",
      " 'Loss/regularization_loss': 0.1010445,\n",
      " 'Loss/total_loss': 0.29235873,\n",
      " 'learning_rate': 0.0678273}\n",
      "I1019 22:49:37.917845 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.101886205,\n",
      " 'Loss/localization_loss': 0.089428015,\n",
      " 'Loss/regularization_loss': 0.1010445,\n",
      " 'Loss/total_loss': 0.29235873,\n",
      " 'learning_rate': 0.0678273}\n",
      "INFO:tensorflow:Step 13600 per-step time 1.047s\n",
      "I1019 22:51:22.705981 4682925568 model_lib_v2.py:705] Step 13600 per-step time 1.047s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07833924,\n",
      " 'Loss/localization_loss': 0.01728768,\n",
      " 'Loss/regularization_loss': 0.10069309,\n",
      " 'Loss/total_loss': 0.19632001,\n",
      " 'learning_rate': 0.0676425}\n",
      "I1019 22:51:22.712089 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07833924,\n",
      " 'Loss/localization_loss': 0.01728768,\n",
      " 'Loss/regularization_loss': 0.10069309,\n",
      " 'Loss/total_loss': 0.19632001,\n",
      " 'learning_rate': 0.0676425}\n",
      "INFO:tensorflow:Step 13700 per-step time 1.325s\n",
      "I1019 22:53:35.261603 4682925568 model_lib_v2.py:705] Step 13700 per-step time 1.325s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.3218349,\n",
      " 'Loss/localization_loss': 0.14240333,\n",
      " 'Loss/regularization_loss': 0.10033403,\n",
      " 'Loss/total_loss': 0.5645723,\n",
      " 'learning_rate': 0.06745657}\n",
      "I1019 22:53:35.268028 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.3218349,\n",
      " 'Loss/localization_loss': 0.14240333,\n",
      " 'Loss/regularization_loss': 0.10033403,\n",
      " 'Loss/total_loss': 0.5645723,\n",
      " 'learning_rate': 0.06745657}\n",
      "INFO:tensorflow:Step 13800 per-step time 0.898s\n",
      "I1019 22:55:05.032881 4682925568 model_lib_v2.py:705] Step 13800 per-step time 0.898s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.18340318,\n",
      " 'Loss/localization_loss': 0.1648939,\n",
      " 'Loss/regularization_loss': 0.099999264,\n",
      " 'Loss/total_loss': 0.4482963,\n",
      " 'learning_rate': 0.06726951}\n",
      "I1019 22:55:05.035310 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.18340318,\n",
      " 'Loss/localization_loss': 0.1648939,\n",
      " 'Loss/regularization_loss': 0.099999264,\n",
      " 'Loss/total_loss': 0.4482963,\n",
      " 'learning_rate': 0.06726951}\n",
      "INFO:tensorflow:Step 13900 per-step time 0.958s\n",
      "I1019 22:56:40.797477 4682925568 model_lib_v2.py:705] Step 13900 per-step time 0.958s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.10301314,\n",
      " 'Loss/localization_loss': 0.038311053,\n",
      " 'Loss/regularization_loss': 0.0999468,\n",
      " 'Loss/total_loss': 0.24127099,\n",
      " 'learning_rate': 0.067081325}\n",
      "I1019 22:56:40.801526 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.10301314,\n",
      " 'Loss/localization_loss': 0.038311053,\n",
      " 'Loss/regularization_loss': 0.0999468,\n",
      " 'Loss/total_loss': 0.24127099,\n",
      " 'learning_rate': 0.067081325}\n",
      "INFO:tensorflow:Step 14000 per-step time 0.877s\n",
      "I1019 22:58:08.489258 4682925568 model_lib_v2.py:705] Step 14000 per-step time 0.877s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08873044,\n",
      " 'Loss/localization_loss': 0.04850568,\n",
      " 'Loss/regularization_loss': 0.09993149,\n",
      " 'Loss/total_loss': 0.23716761,\n",
      " 'learning_rate': 0.06689203}\n",
      "I1019 22:58:08.491559 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08873044,\n",
      " 'Loss/localization_loss': 0.04850568,\n",
      " 'Loss/regularization_loss': 0.09993149,\n",
      " 'Loss/total_loss': 0.23716761,\n",
      " 'learning_rate': 0.06689203}\n",
      "INFO:tensorflow:Step 14100 per-step time 1.030s\n",
      "I1019 22:59:51.546553 4682925568 model_lib_v2.py:705] Step 14100 per-step time 1.030s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08291115,\n",
      " 'Loss/localization_loss': 0.017235715,\n",
      " 'Loss/regularization_loss': 0.09980138,\n",
      " 'Loss/total_loss': 0.19994824,\n",
      " 'learning_rate': 0.06670163}\n",
      "I1019 22:59:51.551206 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08291115,\n",
      " 'Loss/localization_loss': 0.017235715,\n",
      " 'Loss/regularization_loss': 0.09980138,\n",
      " 'Loss/total_loss': 0.19994824,\n",
      " 'learning_rate': 0.06670163}\n",
      "INFO:tensorflow:Step 14200 per-step time 0.856s\n",
      "I1019 23:01:17.125876 4682925568 model_lib_v2.py:705] Step 14200 per-step time 0.856s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.069108374,\n",
      " 'Loss/localization_loss': 0.015832158,\n",
      " 'Loss/regularization_loss': 0.099465944,\n",
      " 'Loss/total_loss': 0.18440647,\n",
      " 'learning_rate': 0.06651013}\n",
      "I1019 23:01:17.126129 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.069108374,\n",
      " 'Loss/localization_loss': 0.015832158,\n",
      " 'Loss/regularization_loss': 0.099465944,\n",
      " 'Loss/total_loss': 0.18440647,\n",
      " 'learning_rate': 0.06651013}\n",
      "INFO:tensorflow:Step 14300 per-step time 0.836s\n",
      "I1019 23:02:40.738291 4682925568 model_lib_v2.py:705] Step 14300 per-step time 0.836s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.20802268,\n",
      " 'Loss/localization_loss': 0.06049127,\n",
      " 'Loss/regularization_loss': 0.09904715,\n",
      " 'Loss/total_loss': 0.3675611,\n",
      " 'learning_rate': 0.06631755}\n",
      "I1019 23:02:40.738543 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.20802268,\n",
      " 'Loss/localization_loss': 0.06049127,\n",
      " 'Loss/regularization_loss': 0.09904715,\n",
      " 'Loss/total_loss': 0.3675611,\n",
      " 'learning_rate': 0.06631755}\n",
      "INFO:tensorflow:Step 14400 per-step time 0.827s\n",
      "I1019 23:04:03.466808 4682925568 model_lib_v2.py:705] Step 14400 per-step time 0.827s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.071026094,\n",
      " 'Loss/localization_loss': 0.021047683,\n",
      " 'Loss/regularization_loss': 0.09864926,\n",
      " 'Loss/total_loss': 0.19072303,\n",
      " 'learning_rate': 0.06612387}\n",
      "I1019 23:04:03.467053 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.071026094,\n",
      " 'Loss/localization_loss': 0.021047683,\n",
      " 'Loss/regularization_loss': 0.09864926,\n",
      " 'Loss/total_loss': 0.19072303,\n",
      " 'learning_rate': 0.06612387}\n",
      "INFO:tensorflow:Step 14500 per-step time 0.820s\n",
      "I1019 23:05:25.498999 4682925568 model_lib_v2.py:705] Step 14500 per-step time 0.820s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06351812,\n",
      " 'Loss/localization_loss': 0.030823635,\n",
      " 'Loss/regularization_loss': 0.09823588,\n",
      " 'Loss/total_loss': 0.19257763,\n",
      " 'learning_rate': 0.06592914}\n",
      "I1019 23:05:25.499310 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06351812,\n",
      " 'Loss/localization_loss': 0.030823635,\n",
      " 'Loss/regularization_loss': 0.09823588,\n",
      " 'Loss/total_loss': 0.19257763,\n",
      " 'learning_rate': 0.06592914}\n",
      "INFO:tensorflow:Step 14600 per-step time 0.823s\n",
      "I1019 23:06:47.798492 4682925568 model_lib_v2.py:705] Step 14600 per-step time 0.823s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.1041027,\n",
      " 'Loss/localization_loss': 0.06794126,\n",
      " 'Loss/regularization_loss': 0.09787792,\n",
      " 'Loss/total_loss': 0.2699219,\n",
      " 'learning_rate': 0.06573333}\n",
      "I1019 23:06:47.798722 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.1041027,\n",
      " 'Loss/localization_loss': 0.06794126,\n",
      " 'Loss/regularization_loss': 0.09787792,\n",
      " 'Loss/total_loss': 0.2699219,\n",
      " 'learning_rate': 0.06573333}\n",
      "INFO:tensorflow:Step 14700 per-step time 0.839s\n",
      "I1019 23:08:11.708949 4682925568 model_lib_v2.py:705] Step 14700 per-step time 0.839s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06886989,\n",
      " 'Loss/localization_loss': 0.02489774,\n",
      " 'Loss/regularization_loss': 0.09750829,\n",
      " 'Loss/total_loss': 0.19127592,\n",
      " 'learning_rate': 0.06553646}\n",
      "I1019 23:08:11.710504 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06886989,\n",
      " 'Loss/localization_loss': 0.02489774,\n",
      " 'Loss/regularization_loss': 0.09750829,\n",
      " 'Loss/total_loss': 0.19127592,\n",
      " 'learning_rate': 0.06553646}\n",
      "INFO:tensorflow:Step 14800 per-step time 0.830s\n",
      "I1019 23:09:34.758804 4682925568 model_lib_v2.py:705] Step 14800 per-step time 0.830s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.093225464,\n",
      " 'Loss/localization_loss': 0.027037038,\n",
      " 'Loss/regularization_loss': 0.097122885,\n",
      " 'Loss/total_loss': 0.21738538,\n",
      " 'learning_rate': 0.065338545}\n",
      "I1019 23:09:34.761168 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.093225464,\n",
      " 'Loss/localization_loss': 0.027037038,\n",
      " 'Loss/regularization_loss': 0.097122885,\n",
      " 'Loss/total_loss': 0.21738538,\n",
      " 'learning_rate': 0.065338545}\n",
      "INFO:tensorflow:Step 14900 per-step time 0.827s\n",
      "I1019 23:10:57.473413 4682925568 model_lib_v2.py:705] Step 14900 per-step time 0.827s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08497665,\n",
      " 'Loss/localization_loss': 0.023153681,\n",
      " 'Loss/regularization_loss': 0.09675134,\n",
      " 'Loss/total_loss': 0.20488167,\n",
      " 'learning_rate': 0.06513958}\n",
      "I1019 23:10:57.473677 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08497665,\n",
      " 'Loss/localization_loss': 0.023153681,\n",
      " 'Loss/regularization_loss': 0.09675134,\n",
      " 'Loss/total_loss': 0.20488167,\n",
      " 'learning_rate': 0.06513958}\n",
      "INFO:tensorflow:Step 15000 per-step time 0.813s\n",
      "I1019 23:12:18.734598 4682925568 model_lib_v2.py:705] Step 15000 per-step time 0.813s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.087879635,\n",
      " 'Loss/localization_loss': 0.038775064,\n",
      " 'Loss/regularization_loss': 0.09635906,\n",
      " 'Loss/total_loss': 0.22301376,\n",
      " 'learning_rate': 0.064939596}\n",
      "I1019 23:12:18.734927 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.087879635,\n",
      " 'Loss/localization_loss': 0.038775064,\n",
      " 'Loss/regularization_loss': 0.09635906,\n",
      " 'Loss/total_loss': 0.22301376,\n",
      " 'learning_rate': 0.064939596}\n",
      "INFO:tensorflow:Step 15100 per-step time 0.909s\n",
      "I1019 23:13:49.604948 4682925568 model_lib_v2.py:705] Step 15100 per-step time 0.909s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.15787357,\n",
      " 'Loss/localization_loss': 0.055311922,\n",
      " 'Loss/regularization_loss': 0.09598048,\n",
      " 'Loss/total_loss': 0.30916595,\n",
      " 'learning_rate': 0.06473857}\n",
      "I1019 23:13:49.607211 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.15787357,\n",
      " 'Loss/localization_loss': 0.055311922,\n",
      " 'Loss/regularization_loss': 0.09598048,\n",
      " 'Loss/total_loss': 0.30916595,\n",
      " 'learning_rate': 0.06473857}\n",
      "INFO:tensorflow:Step 15200 per-step time 0.851s\n",
      "I1019 23:15:14.734213 4682925568 model_lib_v2.py:705] Step 15200 per-step time 0.851s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.13919213,\n",
      " 'Loss/localization_loss': 0.041229524,\n",
      " 'Loss/regularization_loss': 0.0956203,\n",
      " 'Loss/total_loss': 0.27604195,\n",
      " 'learning_rate': 0.06453654}\n",
      "I1019 23:15:14.734459 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.13919213,\n",
      " 'Loss/localization_loss': 0.041229524,\n",
      " 'Loss/regularization_loss': 0.0956203,\n",
      " 'Loss/total_loss': 0.27604195,\n",
      " 'learning_rate': 0.06453654}\n",
      "INFO:tensorflow:Step 15300 per-step time 0.804s\n",
      "I1019 23:16:35.129163 4682925568 model_lib_v2.py:705] Step 15300 per-step time 0.804s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.10434659,\n",
      " 'Loss/localization_loss': 0.01939763,\n",
      " 'Loss/regularization_loss': 0.09526539,\n",
      " 'Loss/total_loss': 0.21900961,\n",
      " 'learning_rate': 0.06433349}\n",
      "I1019 23:16:35.129402 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.10434659,\n",
      " 'Loss/localization_loss': 0.01939763,\n",
      " 'Loss/regularization_loss': 0.09526539,\n",
      " 'Loss/total_loss': 0.21900961,\n",
      " 'learning_rate': 0.06433349}\n",
      "INFO:tensorflow:Step 15400 per-step time 0.811s\n",
      "I1019 23:17:56.253250 4682925568 model_lib_v2.py:705] Step 15400 per-step time 0.811s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06329469,\n",
      " 'Loss/localization_loss': 0.021856945,\n",
      " 'Loss/regularization_loss': 0.094890505,\n",
      " 'Loss/total_loss': 0.18004213,\n",
      " 'learning_rate': 0.06412946}\n",
      "I1019 23:17:56.253739 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06329469,\n",
      " 'Loss/localization_loss': 0.021856945,\n",
      " 'Loss/regularization_loss': 0.094890505,\n",
      " 'Loss/total_loss': 0.18004213,\n",
      " 'learning_rate': 0.06412946}\n",
      "INFO:tensorflow:Step 15500 per-step time 0.824s\n",
      "I1019 23:19:18.665025 4682925568 model_lib_v2.py:705] Step 15500 per-step time 0.824s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07915945,\n",
      " 'Loss/localization_loss': 0.0194083,\n",
      " 'Loss/regularization_loss': 0.09456885,\n",
      " 'Loss/total_loss': 0.1931366,\n",
      " 'learning_rate': 0.063924424}\n",
      "I1019 23:19:18.665301 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07915945,\n",
      " 'Loss/localization_loss': 0.0194083,\n",
      " 'Loss/regularization_loss': 0.09456885,\n",
      " 'Loss/total_loss': 0.1931366,\n",
      " 'learning_rate': 0.063924424}\n",
      "INFO:tensorflow:Step 15600 per-step time 0.797s\n",
      "I1019 23:20:38.363157 4682925568 model_lib_v2.py:705] Step 15600 per-step time 0.797s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.048345648,\n",
      " 'Loss/localization_loss': 0.0414449,\n",
      " 'Loss/regularization_loss': 0.09422469,\n",
      " 'Loss/total_loss': 0.18401524,\n",
      " 'learning_rate': 0.0637184}\n",
      "I1019 23:20:38.363397 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.048345648,\n",
      " 'Loss/localization_loss': 0.0414449,\n",
      " 'Loss/regularization_loss': 0.09422469,\n",
      " 'Loss/total_loss': 0.18401524,\n",
      " 'learning_rate': 0.0637184}\n",
      "INFO:tensorflow:Step 15700 per-step time 0.802s\n",
      "I1019 23:21:58.517101 4682925568 model_lib_v2.py:705] Step 15700 per-step time 0.802s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.079305984,\n",
      " 'Loss/localization_loss': 0.012523058,\n",
      " 'Loss/regularization_loss': 0.09390043,\n",
      " 'Loss/total_loss': 0.18572947,\n",
      " 'learning_rate': 0.06351141}\n",
      "I1019 23:21:58.517410 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.079305984,\n",
      " 'Loss/localization_loss': 0.012523058,\n",
      " 'Loss/regularization_loss': 0.09390043,\n",
      " 'Loss/total_loss': 0.18572947,\n",
      " 'learning_rate': 0.06351141}\n",
      "INFO:tensorflow:Step 15800 per-step time 0.798s\n",
      "I1019 23:23:18.282286 4682925568 model_lib_v2.py:705] Step 15800 per-step time 0.798s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07915263,\n",
      " 'Loss/localization_loss': 0.018617395,\n",
      " 'Loss/regularization_loss': 0.093579315,\n",
      " 'Loss/total_loss': 0.19134933,\n",
      " 'learning_rate': 0.06330345}\n",
      "I1019 23:23:18.282523 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07915263,\n",
      " 'Loss/localization_loss': 0.018617395,\n",
      " 'Loss/regularization_loss': 0.093579315,\n",
      " 'Loss/total_loss': 0.19134933,\n",
      " 'learning_rate': 0.06330345}\n",
      "INFO:tensorflow:Step 15900 per-step time 1.070s\n",
      "I1019 23:25:05.328455 4682925568 model_lib_v2.py:705] Step 15900 per-step time 1.070s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07185056,\n",
      " 'Loss/localization_loss': 0.043516874,\n",
      " 'Loss/regularization_loss': 0.09326482,\n",
      " 'Loss/total_loss': 0.20863226,\n",
      " 'learning_rate': 0.06309453}\n",
      "I1019 23:25:05.337585 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07185056,\n",
      " 'Loss/localization_loss': 0.043516874,\n",
      " 'Loss/regularization_loss': 0.09326482,\n",
      " 'Loss/total_loss': 0.20863226,\n",
      " 'learning_rate': 0.06309453}\n",
      "INFO:tensorflow:Step 16000 per-step time 1.042s\n",
      "I1019 23:26:49.506404 4682925568 model_lib_v2.py:705] Step 16000 per-step time 1.042s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05510328,\n",
      " 'Loss/localization_loss': 0.047678575,\n",
      " 'Loss/regularization_loss': 0.09287453,\n",
      " 'Loss/total_loss': 0.19565639,\n",
      " 'learning_rate': 0.062884666}\n",
      "I1019 23:26:49.509043 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05510328,\n",
      " 'Loss/localization_loss': 0.047678575,\n",
      " 'Loss/regularization_loss': 0.09287453,\n",
      " 'Loss/total_loss': 0.19565639,\n",
      " 'learning_rate': 0.062884666}\n",
      "INFO:tensorflow:Step 16100 per-step time 1.118s\n",
      "I1019 23:28:41.359919 4682925568 model_lib_v2.py:705] Step 16100 per-step time 1.118s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.045802224,\n",
      " 'Loss/localization_loss': 0.022662664,\n",
      " 'Loss/regularization_loss': 0.09253622,\n",
      " 'Loss/total_loss': 0.16100112,\n",
      " 'learning_rate': 0.06267385}\n",
      "I1019 23:28:41.365139 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.045802224,\n",
      " 'Loss/localization_loss': 0.022662664,\n",
      " 'Loss/regularization_loss': 0.09253622,\n",
      " 'Loss/total_loss': 0.16100112,\n",
      " 'learning_rate': 0.06267385}\n",
      "INFO:tensorflow:Step 16200 per-step time 0.999s\n",
      "I1019 23:30:21.238739 4682925568 model_lib_v2.py:705] Step 16200 per-step time 0.999s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06156271,\n",
      " 'Loss/localization_loss': 0.025599966,\n",
      " 'Loss/regularization_loss': 0.09229492,\n",
      " 'Loss/total_loss': 0.17945759,\n",
      " 'learning_rate': 0.062462118}\n",
      "I1019 23:30:21.241594 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06156271,\n",
      " 'Loss/localization_loss': 0.025599966,\n",
      " 'Loss/regularization_loss': 0.09229492,\n",
      " 'Loss/total_loss': 0.17945759,\n",
      " 'learning_rate': 0.062462118}\n",
      "INFO:tensorflow:Step 16300 per-step time 0.940s\n",
      "I1019 23:31:55.274329 4682925568 model_lib_v2.py:705] Step 16300 per-step time 0.940s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.067330934,\n",
      " 'Loss/localization_loss': 0.030917581,\n",
      " 'Loss/regularization_loss': 0.09196007,\n",
      " 'Loss/total_loss': 0.19020858,\n",
      " 'learning_rate': 0.06224946}\n",
      "I1019 23:31:55.277969 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.067330934,\n",
      " 'Loss/localization_loss': 0.030917581,\n",
      " 'Loss/regularization_loss': 0.09196007,\n",
      " 'Loss/total_loss': 0.19020858,\n",
      " 'learning_rate': 0.06224946}\n",
      "INFO:tensorflow:Step 16400 per-step time 0.966s\n",
      "I1019 23:33:31.892732 4682925568 model_lib_v2.py:705] Step 16400 per-step time 0.966s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.090270415,\n",
      " 'Loss/localization_loss': 0.06881549,\n",
      " 'Loss/regularization_loss': 0.09161972,\n",
      " 'Loss/total_loss': 0.25070563,\n",
      " 'learning_rate': 0.062035877}\n",
      "I1019 23:33:31.897665 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.090270415,\n",
      " 'Loss/localization_loss': 0.06881549,\n",
      " 'Loss/regularization_loss': 0.09161972,\n",
      " 'Loss/total_loss': 0.25070563,\n",
      " 'learning_rate': 0.062035877}\n",
      "INFO:tensorflow:Step 16500 per-step time 0.916s\n",
      "I1019 23:35:03.434397 4682925568 model_lib_v2.py:705] Step 16500 per-step time 0.916s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.12681684,\n",
      " 'Loss/localization_loss': 0.07599521,\n",
      " 'Loss/regularization_loss': 0.09126414,\n",
      " 'Loss/total_loss': 0.2940762,\n",
      " 'learning_rate': 0.061821394}\n",
      "I1019 23:35:03.437453 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.12681684,\n",
      " 'Loss/localization_loss': 0.07599521,\n",
      " 'Loss/regularization_loss': 0.09126414,\n",
      " 'Loss/total_loss': 0.2940762,\n",
      " 'learning_rate': 0.061821394}\n",
      "INFO:tensorflow:Step 16600 per-step time 0.878s\n",
      "I1019 23:36:31.279549 4682925568 model_lib_v2.py:705] Step 16600 per-step time 0.878s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07179982,\n",
      " 'Loss/localization_loss': 0.014379565,\n",
      " 'Loss/regularization_loss': 0.09092552,\n",
      " 'Loss/total_loss': 0.17710492,\n",
      " 'learning_rate': 0.061606016}\n",
      "I1019 23:36:31.283643 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07179982,\n",
      " 'Loss/localization_loss': 0.014379565,\n",
      " 'Loss/regularization_loss': 0.09092552,\n",
      " 'Loss/total_loss': 0.17710492,\n",
      " 'learning_rate': 0.061606016}\n",
      "INFO:tensorflow:Step 16700 per-step time 1.026s\n",
      "I1019 23:38:13.923722 4682925568 model_lib_v2.py:705] Step 16700 per-step time 1.026s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05028619,\n",
      " 'Loss/localization_loss': 0.039677974,\n",
      " 'Loss/regularization_loss': 0.09067264,\n",
      " 'Loss/total_loss': 0.18063681,\n",
      " 'learning_rate': 0.06138974}\n",
      "I1019 23:38:13.930737 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05028619,\n",
      " 'Loss/localization_loss': 0.039677974,\n",
      " 'Loss/regularization_loss': 0.09067264,\n",
      " 'Loss/total_loss': 0.18063681,\n",
      " 'learning_rate': 0.06138974}\n",
      "INFO:tensorflow:Step 16800 per-step time 0.993s\n",
      "I1019 23:39:53.246945 4682925568 model_lib_v2.py:705] Step 16800 per-step time 0.993s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.09035277,\n",
      " 'Loss/localization_loss': 0.025766505,\n",
      " 'Loss/regularization_loss': 0.090359785,\n",
      " 'Loss/total_loss': 0.20647907,\n",
      " 'learning_rate': 0.061172597}\n",
      "I1019 23:39:53.249963 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.09035277,\n",
      " 'Loss/localization_loss': 0.025766505,\n",
      " 'Loss/regularization_loss': 0.090359785,\n",
      " 'Loss/total_loss': 0.20647907,\n",
      " 'learning_rate': 0.061172597}\n",
      "INFO:tensorflow:Step 16900 per-step time 0.915s\n",
      "I1019 23:41:24.728787 4682925568 model_lib_v2.py:705] Step 16900 per-step time 0.915s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.13277745,\n",
      " 'Loss/localization_loss': 0.03090181,\n",
      " 'Loss/regularization_loss': 0.09003711,\n",
      " 'Loss/total_loss': 0.25371635,\n",
      " 'learning_rate': 0.06095458}\n",
      "I1019 23:41:24.731300 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.13277745,\n",
      " 'Loss/localization_loss': 0.03090181,\n",
      " 'Loss/regularization_loss': 0.09003711,\n",
      " 'Loss/total_loss': 0.25371635,\n",
      " 'learning_rate': 0.06095458}\n",
      "INFO:tensorflow:Step 17000 per-step time 0.854s\n",
      "I1019 23:42:50.073107 4682925568 model_lib_v2.py:705] Step 17000 per-step time 0.854s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.13649225,\n",
      " 'Loss/localization_loss': 0.03402796,\n",
      " 'Loss/regularization_loss': 0.089713246,\n",
      " 'Loss/total_loss': 0.26023346,\n",
      " 'learning_rate': 0.060735703}\n",
      "I1019 23:42:50.073364 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.13649225,\n",
      " 'Loss/localization_loss': 0.03402796,\n",
      " 'Loss/regularization_loss': 0.089713246,\n",
      " 'Loss/total_loss': 0.26023346,\n",
      " 'learning_rate': 0.060735703}\n",
      "INFO:tensorflow:Step 17100 per-step time 0.897s\n",
      "I1019 23:44:19.776622 4682925568 model_lib_v2.py:705] Step 17100 per-step time 0.897s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.15726712,\n",
      " 'Loss/localization_loss': 0.07578254,\n",
      " 'Loss/regularization_loss': 0.08967486,\n",
      " 'Loss/total_loss': 0.32272452,\n",
      " 'learning_rate': 0.060515966}\n",
      "I1019 23:44:19.779785 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.15726712,\n",
      " 'Loss/localization_loss': 0.07578254,\n",
      " 'Loss/regularization_loss': 0.08967486,\n",
      " 'Loss/total_loss': 0.32272452,\n",
      " 'learning_rate': 0.060515966}\n",
      "INFO:tensorflow:Step 17200 per-step time 0.921s\n",
      "I1019 23:45:51.850348 4682925568 model_lib_v2.py:705] Step 17200 per-step time 0.921s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.09742766,\n",
      " 'Loss/localization_loss': 0.026745718,\n",
      " 'Loss/regularization_loss': 0.08959478,\n",
      " 'Loss/total_loss': 0.21376815,\n",
      " 'learning_rate': 0.0602954}\n",
      "I1019 23:45:51.852754 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.09742766,\n",
      " 'Loss/localization_loss': 0.026745718,\n",
      " 'Loss/regularization_loss': 0.08959478,\n",
      " 'Loss/total_loss': 0.21376815,\n",
      " 'learning_rate': 0.0602954}\n",
      "INFO:tensorflow:Step 17300 per-step time 0.849s\n",
      "I1019 23:47:16.705433 4682925568 model_lib_v2.py:705] Step 17300 per-step time 0.849s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.068566605,\n",
      " 'Loss/localization_loss': 0.008939867,\n",
      " 'Loss/regularization_loss': 0.08926878,\n",
      " 'Loss/total_loss': 0.16677526,\n",
      " 'learning_rate': 0.060073983}\n",
      "I1019 23:47:16.705672 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.068566605,\n",
      " 'Loss/localization_loss': 0.008939867,\n",
      " 'Loss/regularization_loss': 0.08926878,\n",
      " 'Loss/total_loss': 0.16677526,\n",
      " 'learning_rate': 0.060073983}\n",
      "INFO:tensorflow:Step 17400 per-step time 0.814s\n",
      "I1019 23:48:38.144343 4682925568 model_lib_v2.py:705] Step 17400 per-step time 0.814s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08931001,\n",
      " 'Loss/localization_loss': 0.019816667,\n",
      " 'Loss/regularization_loss': 0.089008205,\n",
      " 'Loss/total_loss': 0.19813488,\n",
      " 'learning_rate': 0.05985175}\n",
      "I1019 23:48:38.146642 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08931001,\n",
      " 'Loss/localization_loss': 0.019816667,\n",
      " 'Loss/regularization_loss': 0.089008205,\n",
      " 'Loss/total_loss': 0.19813488,\n",
      " 'learning_rate': 0.05985175}\n",
      "INFO:tensorflow:Step 17500 per-step time 0.820s\n",
      "I1019 23:50:00.148247 4682925568 model_lib_v2.py:705] Step 17500 per-step time 0.820s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.043597043,\n",
      " 'Loss/localization_loss': 0.016790157,\n",
      " 'Loss/regularization_loss': 0.08870222,\n",
      " 'Loss/total_loss': 0.14908943,\n",
      " 'learning_rate': 0.0596287}\n",
      "I1019 23:50:00.148479 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.043597043,\n",
      " 'Loss/localization_loss': 0.016790157,\n",
      " 'Loss/regularization_loss': 0.08870222,\n",
      " 'Loss/total_loss': 0.14908943,\n",
      " 'learning_rate': 0.0596287}\n",
      "INFO:tensorflow:Step 17600 per-step time 0.847s\n",
      "I1019 23:51:24.895370 4682925568 model_lib_v2.py:705] Step 17600 per-step time 0.847s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08771014,\n",
      " 'Loss/localization_loss': 0.022361372,\n",
      " 'Loss/regularization_loss': 0.088723905,\n",
      " 'Loss/total_loss': 0.19879541,\n",
      " 'learning_rate': 0.05940484}\n",
      "I1019 23:51:24.897526 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08771014,\n",
      " 'Loss/localization_loss': 0.022361372,\n",
      " 'Loss/regularization_loss': 0.088723905,\n",
      " 'Loss/total_loss': 0.19879541,\n",
      " 'learning_rate': 0.05940484}\n",
      "INFO:tensorflow:Step 17700 per-step time 0.843s\n",
      "I1019 23:52:49.166969 4682925568 model_lib_v2.py:705] Step 17700 per-step time 0.843s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07122704,\n",
      " 'Loss/localization_loss': 0.014817817,\n",
      " 'Loss/regularization_loss': 0.08842569,\n",
      " 'Loss/total_loss': 0.17447054,\n",
      " 'learning_rate': 0.059180185}\n",
      "I1019 23:52:49.169255 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07122704,\n",
      " 'Loss/localization_loss': 0.014817817,\n",
      " 'Loss/regularization_loss': 0.08842569,\n",
      " 'Loss/total_loss': 0.17447054,\n",
      " 'learning_rate': 0.059180185}\n",
      "INFO:tensorflow:Step 17800 per-step time 0.810s\n",
      "I1019 23:54:10.126157 4682925568 model_lib_v2.py:705] Step 17800 per-step time 0.810s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06304894,\n",
      " 'Loss/localization_loss': 0.009791288,\n",
      " 'Loss/regularization_loss': 0.08813141,\n",
      " 'Loss/total_loss': 0.16097164,\n",
      " 'learning_rate': 0.05895474}\n",
      "I1019 23:54:10.126448 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06304894,\n",
      " 'Loss/localization_loss': 0.009791288,\n",
      " 'Loss/regularization_loss': 0.08813141,\n",
      " 'Loss/total_loss': 0.16097164,\n",
      " 'learning_rate': 0.05895474}\n",
      "INFO:tensorflow:Step 17900 per-step time 0.801s\n",
      "I1019 23:55:30.188889 4682925568 model_lib_v2.py:705] Step 17900 per-step time 0.801s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.082972884,\n",
      " 'Loss/localization_loss': 0.045704413,\n",
      " 'Loss/regularization_loss': 0.087874606,\n",
      " 'Loss/total_loss': 0.2165519,\n",
      " 'learning_rate': 0.058728524}\n",
      "I1019 23:55:30.189136 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.082972884,\n",
      " 'Loss/localization_loss': 0.045704413,\n",
      " 'Loss/regularization_loss': 0.087874606,\n",
      " 'Loss/total_loss': 0.2165519,\n",
      " 'learning_rate': 0.058728524}\n",
      "INFO:tensorflow:Step 18000 per-step time 0.797s\n",
      "I1019 23:56:49.906119 4682925568 model_lib_v2.py:705] Step 18000 per-step time 0.797s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.103991136,\n",
      " 'Loss/localization_loss': 0.050460014,\n",
      " 'Loss/regularization_loss': 0.08761779,\n",
      " 'Loss/total_loss': 0.24206895,\n",
      " 'learning_rate': 0.058501527}\n",
      "I1019 23:56:49.906352 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.103991136,\n",
      " 'Loss/localization_loss': 0.050460014,\n",
      " 'Loss/regularization_loss': 0.08761779,\n",
      " 'Loss/total_loss': 0.24206895,\n",
      " 'learning_rate': 0.058501527}\n",
      "INFO:tensorflow:Step 18100 per-step time 0.922s\n",
      "I1019 23:58:22.083811 4682925568 model_lib_v2.py:705] Step 18100 per-step time 0.922s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07028492,\n",
      " 'Loss/localization_loss': 0.02173203,\n",
      " 'Loss/regularization_loss': 0.08736418,\n",
      " 'Loss/total_loss': 0.17938113,\n",
      " 'learning_rate': 0.058273777}\n",
      "I1019 23:58:22.085993 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07028492,\n",
      " 'Loss/localization_loss': 0.02173203,\n",
      " 'Loss/regularization_loss': 0.08736418,\n",
      " 'Loss/total_loss': 0.17938113,\n",
      " 'learning_rate': 0.058273777}\n",
      "INFO:tensorflow:Step 18200 per-step time 0.819s\n",
      "I1019 23:59:43.942230 4682925568 model_lib_v2.py:705] Step 18200 per-step time 0.819s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.1308987,\n",
      " 'Loss/localization_loss': 0.0924744,\n",
      " 'Loss/regularization_loss': 0.08707097,\n",
      " 'Loss/total_loss': 0.31044406,\n",
      " 'learning_rate': 0.05804527}\n",
      "I1019 23:59:43.942501 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.1308987,\n",
      " 'Loss/localization_loss': 0.0924744,\n",
      " 'Loss/regularization_loss': 0.08707097,\n",
      " 'Loss/total_loss': 0.31044406,\n",
      " 'learning_rate': 0.05804527}\n",
      "INFO:tensorflow:Step 18300 per-step time 0.961s\n",
      "I1020 00:01:20.007942 4682925568 model_lib_v2.py:705] Step 18300 per-step time 0.961s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.09614253,\n",
      " 'Loss/localization_loss': 0.0146615,\n",
      " 'Loss/regularization_loss': 0.08676975,\n",
      " 'Loss/total_loss': 0.19757378,\n",
      " 'learning_rate': 0.05781603}\n",
      "I1020 00:01:20.011868 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.09614253,\n",
      " 'Loss/localization_loss': 0.0146615,\n",
      " 'Loss/regularization_loss': 0.08676975,\n",
      " 'Loss/total_loss': 0.19757378,\n",
      " 'learning_rate': 0.05781603}\n",
      "INFO:tensorflow:Step 18400 per-step time 0.892s\n",
      "I1020 00:02:49.184490 4682925568 model_lib_v2.py:705] Step 18400 per-step time 0.892s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05915336,\n",
      " 'Loss/localization_loss': 0.027066002,\n",
      " 'Loss/regularization_loss': 0.086456425,\n",
      " 'Loss/total_loss': 0.17267579,\n",
      " 'learning_rate': 0.057586048}\n",
      "I1020 00:02:49.186738 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05915336,\n",
      " 'Loss/localization_loss': 0.027066002,\n",
      " 'Loss/regularization_loss': 0.086456425,\n",
      " 'Loss/total_loss': 0.17267579,\n",
      " 'learning_rate': 0.057586048}\n",
      "INFO:tensorflow:Step 18500 per-step time 0.871s\n",
      "I1020 00:04:16.246140 4682925568 model_lib_v2.py:705] Step 18500 per-step time 0.871s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0691518,\n",
      " 'Loss/localization_loss': 0.056411054,\n",
      " 'Loss/regularization_loss': 0.08619212,\n",
      " 'Loss/total_loss': 0.21175498,\n",
      " 'learning_rate': 0.05735535}\n",
      "I1020 00:04:16.248373 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0691518,\n",
      " 'Loss/localization_loss': 0.056411054,\n",
      " 'Loss/regularization_loss': 0.08619212,\n",
      " 'Loss/total_loss': 0.21175498,\n",
      " 'learning_rate': 0.05735535}\n",
      "INFO:tensorflow:Step 18600 per-step time 0.821s\n",
      "I1020 00:05:38.333530 4682925568 model_lib_v2.py:705] Step 18600 per-step time 0.821s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06739557,\n",
      " 'Loss/localization_loss': 0.020285217,\n",
      " 'Loss/regularization_loss': 0.085971974,\n",
      " 'Loss/total_loss': 0.17365277,\n",
      " 'learning_rate': 0.057123937}\n",
      "I1020 00:05:38.333770 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06739557,\n",
      " 'Loss/localization_loss': 0.020285217,\n",
      " 'Loss/regularization_loss': 0.085971974,\n",
      " 'Loss/total_loss': 0.17365277,\n",
      " 'learning_rate': 0.057123937}\n",
      "INFO:tensorflow:Step 18700 per-step time 0.839s\n",
      "I1020 00:07:02.239518 4682925568 model_lib_v2.py:705] Step 18700 per-step time 0.839s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.056304168,\n",
      " 'Loss/localization_loss': 0.028128177,\n",
      " 'Loss/regularization_loss': 0.0856974,\n",
      " 'Loss/total_loss': 0.17012975,\n",
      " 'learning_rate': 0.05689181}\n",
      "I1020 00:07:02.241867 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.056304168,\n",
      " 'Loss/localization_loss': 0.028128177,\n",
      " 'Loss/regularization_loss': 0.0856974,\n",
      " 'Loss/total_loss': 0.17012975,\n",
      " 'learning_rate': 0.05689181}\n",
      "INFO:tensorflow:Step 18800 per-step time 0.860s\n",
      "I1020 00:08:28.199760 4682925568 model_lib_v2.py:705] Step 18800 per-step time 0.860s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08773655,\n",
      " 'Loss/localization_loss': 0.046841968,\n",
      " 'Loss/regularization_loss': 0.085448526,\n",
      " 'Loss/total_loss': 0.22002703,\n",
      " 'learning_rate': 0.056659}\n",
      "I1020 00:08:28.200038 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08773655,\n",
      " 'Loss/localization_loss': 0.046841968,\n",
      " 'Loss/regularization_loss': 0.085448526,\n",
      " 'Loss/total_loss': 0.22002703,\n",
      " 'learning_rate': 0.056659}\n",
      "INFO:tensorflow:Step 18900 per-step time 0.844s\n",
      "I1020 00:09:52.613229 4682925568 model_lib_v2.py:705] Step 18900 per-step time 0.844s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.084283754,\n",
      " 'Loss/localization_loss': 0.025809582,\n",
      " 'Loss/regularization_loss': 0.08520785,\n",
      " 'Loss/total_loss': 0.19530119,\n",
      " 'learning_rate': 0.056425504}\n",
      "I1020 00:09:52.613481 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.084283754,\n",
      " 'Loss/localization_loss': 0.025809582,\n",
      " 'Loss/regularization_loss': 0.08520785,\n",
      " 'Loss/total_loss': 0.19530119,\n",
      " 'learning_rate': 0.056425504}\n",
      "INFO:tensorflow:Step 19000 per-step time 0.832s\n",
      "I1020 00:11:15.839280 4682925568 model_lib_v2.py:705] Step 19000 per-step time 0.832s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.1261776,\n",
      " 'Loss/localization_loss': 0.037990373,\n",
      " 'Loss/regularization_loss': 0.08498364,\n",
      " 'Loss/total_loss': 0.24915162,\n",
      " 'learning_rate': 0.056191333}\n",
      "I1020 00:11:15.841658 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.1261776,\n",
      " 'Loss/localization_loss': 0.037990373,\n",
      " 'Loss/regularization_loss': 0.08498364,\n",
      " 'Loss/total_loss': 0.24915162,\n",
      " 'learning_rate': 0.056191333}\n",
      "INFO:tensorflow:Step 19100 per-step time 0.832s\n",
      "I1020 00:12:39.050014 4682925568 model_lib_v2.py:705] Step 19100 per-step time 0.832s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0648094,\n",
      " 'Loss/localization_loss': 0.027882302,\n",
      " 'Loss/regularization_loss': 0.08493654,\n",
      " 'Loss/total_loss': 0.17762823,\n",
      " 'learning_rate': 0.055956498}\n",
      "I1020 00:12:39.053223 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0648094,\n",
      " 'Loss/localization_loss': 0.027882302,\n",
      " 'Loss/regularization_loss': 0.08493654,\n",
      " 'Loss/total_loss': 0.17762823,\n",
      " 'learning_rate': 0.055956498}\n",
      "INFO:tensorflow:Step 19200 per-step time 0.822s\n",
      "I1020 00:14:01.206120 4682925568 model_lib_v2.py:705] Step 19200 per-step time 0.822s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.074117005,\n",
      " 'Loss/localization_loss': 0.017569276,\n",
      " 'Loss/regularization_loss': 0.084666856,\n",
      " 'Loss/total_loss': 0.17635313,\n",
      " 'learning_rate': 0.055721}\n",
      "I1020 00:14:01.208392 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.074117005,\n",
      " 'Loss/localization_loss': 0.017569276,\n",
      " 'Loss/regularization_loss': 0.084666856,\n",
      " 'Loss/total_loss': 0.17635313,\n",
      " 'learning_rate': 0.055721}\n",
      "INFO:tensorflow:Step 19300 per-step time 0.803s\n",
      "I1020 00:15:21.442687 4682925568 model_lib_v2.py:705] Step 19300 per-step time 0.803s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04850415,\n",
      " 'Loss/localization_loss': 0.017981369,\n",
      " 'Loss/regularization_loss': 0.08438945,\n",
      " 'Loss/total_loss': 0.15087497,\n",
      " 'learning_rate': 0.055484857}\n",
      "I1020 00:15:21.442917 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04850415,\n",
      " 'Loss/localization_loss': 0.017981369,\n",
      " 'Loss/regularization_loss': 0.08438945,\n",
      " 'Loss/total_loss': 0.15087497,\n",
      " 'learning_rate': 0.055484857}\n",
      "INFO:tensorflow:Step 19400 per-step time 0.802s\n",
      "I1020 00:16:41.685049 4682925568 model_lib_v2.py:705] Step 19400 per-step time 0.802s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.060165662,\n",
      " 'Loss/localization_loss': 0.048282247,\n",
      " 'Loss/regularization_loss': 0.08409652,\n",
      " 'Loss/total_loss': 0.19254443,\n",
      " 'learning_rate': 0.05524808}\n",
      "I1020 00:16:41.685285 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.060165662,\n",
      " 'Loss/localization_loss': 0.048282247,\n",
      " 'Loss/regularization_loss': 0.08409652,\n",
      " 'Loss/total_loss': 0.19254443,\n",
      " 'learning_rate': 0.05524808}\n",
      "INFO:tensorflow:Step 19500 per-step time 0.799s\n",
      "I1020 00:18:01.578085 4682925568 model_lib_v2.py:705] Step 19500 per-step time 0.799s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05410413,\n",
      " 'Loss/localization_loss': 0.023068283,\n",
      " 'Loss/regularization_loss': 0.08379821,\n",
      " 'Loss/total_loss': 0.16097063,\n",
      " 'learning_rate': 0.055010676}\n",
      "I1020 00:18:01.580339 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05410413,\n",
      " 'Loss/localization_loss': 0.023068283,\n",
      " 'Loss/regularization_loss': 0.08379821,\n",
      " 'Loss/total_loss': 0.16097063,\n",
      " 'learning_rate': 0.055010676}\n",
      "INFO:tensorflow:Step 19600 per-step time 0.799s\n",
      "I1020 00:19:21.454917 4682925568 model_lib_v2.py:705] Step 19600 per-step time 0.799s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.09031139,\n",
      " 'Loss/localization_loss': 0.024952821,\n",
      " 'Loss/regularization_loss': 0.08351218,\n",
      " 'Loss/total_loss': 0.1987764,\n",
      " 'learning_rate': 0.054772656}\n",
      "I1020 00:19:21.455224 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.09031139,\n",
      " 'Loss/localization_loss': 0.024952821,\n",
      " 'Loss/regularization_loss': 0.08351218,\n",
      " 'Loss/total_loss': 0.1987764,\n",
      " 'learning_rate': 0.054772656}\n",
      "INFO:tensorflow:Step 19700 per-step time 0.822s\n",
      "I1020 00:20:43.683732 4682925568 model_lib_v2.py:705] Step 19700 per-step time 0.822s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.2548839,\n",
      " 'Loss/localization_loss': 0.09209728,\n",
      " 'Loss/regularization_loss': 0.08325302,\n",
      " 'Loss/total_loss': 0.4302342,\n",
      " 'learning_rate': 0.054534025}\n",
      "I1020 00:20:43.686438 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.2548839,\n",
      " 'Loss/localization_loss': 0.09209728,\n",
      " 'Loss/regularization_loss': 0.08325302,\n",
      " 'Loss/total_loss': 0.4302342,\n",
      " 'learning_rate': 0.054534025}\n",
      "INFO:tensorflow:Step 19800 per-step time 0.797s\n",
      "I1020 00:22:03.391499 4682925568 model_lib_v2.py:705] Step 19800 per-step time 0.797s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.16071002,\n",
      " 'Loss/localization_loss': 0.09035079,\n",
      " 'Loss/regularization_loss': 0.08297526,\n",
      " 'Loss/total_loss': 0.33403608,\n",
      " 'learning_rate': 0.0542948}\n",
      "I1020 00:22:03.391738 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.16071002,\n",
      " 'Loss/localization_loss': 0.09035079,\n",
      " 'Loss/regularization_loss': 0.08297526,\n",
      " 'Loss/total_loss': 0.33403608,\n",
      " 'learning_rate': 0.0542948}\n",
      "INFO:tensorflow:Step 19900 per-step time 0.797s\n",
      "I1020 00:23:23.095354 4682925568 model_lib_v2.py:705] Step 19900 per-step time 0.797s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.090494655,\n",
      " 'Loss/localization_loss': 0.036430527,\n",
      " 'Loss/regularization_loss': 0.08270481,\n",
      " 'Loss/total_loss': 0.20963,\n",
      " 'learning_rate': 0.054054994}\n",
      "I1020 00:23:23.095600 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.090494655,\n",
      " 'Loss/localization_loss': 0.036430527,\n",
      " 'Loss/regularization_loss': 0.08270481,\n",
      " 'Loss/total_loss': 0.20963,\n",
      " 'learning_rate': 0.054054994}\n",
      "INFO:tensorflow:Step 20000 per-step time 0.794s\n",
      "I1020 00:24:42.488239 4682925568 model_lib_v2.py:705] Step 20000 per-step time 0.794s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.18026012,\n",
      " 'Loss/localization_loss': 0.11184752,\n",
      " 'Loss/regularization_loss': 0.082507215,\n",
      " 'Loss/total_loss': 0.37461486,\n",
      " 'learning_rate': 0.0538146}\n",
      "I1020 00:24:42.488474 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.18026012,\n",
      " 'Loss/localization_loss': 0.11184752,\n",
      " 'Loss/regularization_loss': 0.082507215,\n",
      " 'Loss/total_loss': 0.37461486,\n",
      " 'learning_rate': 0.0538146}\n",
      "INFO:tensorflow:Step 20100 per-step time 0.814s\n",
      "I1020 00:26:03.864355 4682925568 model_lib_v2.py:705] Step 20100 per-step time 0.814s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08946996,\n",
      " 'Loss/localization_loss': 0.033866204,\n",
      " 'Loss/regularization_loss': 0.082230896,\n",
      " 'Loss/total_loss': 0.20556706,\n",
      " 'learning_rate': 0.053573642}\n",
      "I1020 00:26:03.866518 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08946996,\n",
      " 'Loss/localization_loss': 0.033866204,\n",
      " 'Loss/regularization_loss': 0.082230896,\n",
      " 'Loss/total_loss': 0.20556706,\n",
      " 'learning_rate': 0.053573642}\n",
      "INFO:tensorflow:Step 20200 per-step time 0.847s\n",
      "I1020 00:27:28.562341 4682925568 model_lib_v2.py:705] Step 20200 per-step time 0.847s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.1126475,\n",
      " 'Loss/localization_loss': 0.032408375,\n",
      " 'Loss/regularization_loss': 0.08201122,\n",
      " 'Loss/total_loss': 0.2270671,\n",
      " 'learning_rate': 0.053332124}\n",
      "I1020 00:27:28.562585 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.1126475,\n",
      " 'Loss/localization_loss': 0.032408375,\n",
      " 'Loss/regularization_loss': 0.08201122,\n",
      " 'Loss/total_loss': 0.2270671,\n",
      " 'learning_rate': 0.053332124}\n",
      "INFO:tensorflow:Step 20300 per-step time 0.799s\n",
      "I1020 00:28:48.481595 4682925568 model_lib_v2.py:705] Step 20300 per-step time 0.799s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.2177851,\n",
      " 'Loss/localization_loss': 0.05227505,\n",
      " 'Loss/regularization_loss': 0.08188223,\n",
      " 'Loss/total_loss': 0.3519424,\n",
      " 'learning_rate': 0.053090062}\n",
      "I1020 00:28:48.481871 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.2177851,\n",
      " 'Loss/localization_loss': 0.05227505,\n",
      " 'Loss/regularization_loss': 0.08188223,\n",
      " 'Loss/total_loss': 0.3519424,\n",
      " 'learning_rate': 0.053090062}\n",
      "INFO:tensorflow:Step 20400 per-step time 0.818s\n",
      "I1020 00:30:10.329171 4682925568 model_lib_v2.py:705] Step 20400 per-step time 0.818s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.059747696,\n",
      " 'Loss/localization_loss': 0.01919296,\n",
      " 'Loss/regularization_loss': 0.081804864,\n",
      " 'Loss/total_loss': 0.16074553,\n",
      " 'learning_rate': 0.05284746}\n",
      "I1020 00:30:10.331493 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.059747696,\n",
      " 'Loss/localization_loss': 0.01919296,\n",
      " 'Loss/regularization_loss': 0.081804864,\n",
      " 'Loss/total_loss': 0.16074553,\n",
      " 'learning_rate': 0.05284746}\n",
      "INFO:tensorflow:Step 20500 per-step time 0.800s\n",
      "I1020 00:31:30.272948 4682925568 model_lib_v2.py:705] Step 20500 per-step time 0.800s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06489156,\n",
      " 'Loss/localization_loss': 0.02859745,\n",
      " 'Loss/regularization_loss': 0.08158866,\n",
      " 'Loss/total_loss': 0.17507768,\n",
      " 'learning_rate': 0.052604325}\n",
      "I1020 00:31:30.273202 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06489156,\n",
      " 'Loss/localization_loss': 0.02859745,\n",
      " 'Loss/regularization_loss': 0.08158866,\n",
      " 'Loss/total_loss': 0.17507768,\n",
      " 'learning_rate': 0.052604325}\n",
      "INFO:tensorflow:Step 20600 per-step time 0.801s\n",
      "I1020 00:32:50.336642 4682925568 model_lib_v2.py:705] Step 20600 per-step time 0.801s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.043293193,\n",
      " 'Loss/localization_loss': 0.007153854,\n",
      " 'Loss/regularization_loss': 0.081331626,\n",
      " 'Loss/total_loss': 0.13177867,\n",
      " 'learning_rate': 0.052360676}\n",
      "I1020 00:32:50.336888 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.043293193,\n",
      " 'Loss/localization_loss': 0.007153854,\n",
      " 'Loss/regularization_loss': 0.081331626,\n",
      " 'Loss/total_loss': 0.13177867,\n",
      " 'learning_rate': 0.052360676}\n",
      "INFO:tensorflow:Step 20700 per-step time 0.809s\n",
      "I1020 00:34:11.287920 4682925568 model_lib_v2.py:705] Step 20700 per-step time 0.809s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08296368,\n",
      " 'Loss/localization_loss': 0.022706818,\n",
      " 'Loss/regularization_loss': 0.08107193,\n",
      " 'Loss/total_loss': 0.18674242,\n",
      " 'learning_rate': 0.052116517}\n",
      "I1020 00:34:11.290771 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08296368,\n",
      " 'Loss/localization_loss': 0.022706818,\n",
      " 'Loss/regularization_loss': 0.08107193,\n",
      " 'Loss/total_loss': 0.18674242,\n",
      " 'learning_rate': 0.052116517}\n",
      "INFO:tensorflow:Step 20800 per-step time 0.793s\n",
      "I1020 00:35:30.610266 4682925568 model_lib_v2.py:705] Step 20800 per-step time 0.793s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08016539,\n",
      " 'Loss/localization_loss': 0.025777377,\n",
      " 'Loss/regularization_loss': 0.080791175,\n",
      " 'Loss/total_loss': 0.18673395,\n",
      " 'learning_rate': 0.05187187}\n",
      "I1020 00:35:30.610503 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08016539,\n",
      " 'Loss/localization_loss': 0.025777377,\n",
      " 'Loss/regularization_loss': 0.080791175,\n",
      " 'Loss/total_loss': 0.18673395,\n",
      " 'learning_rate': 0.05187187}\n",
      "INFO:tensorflow:Step 20900 per-step time 0.810s\n",
      "I1020 00:36:51.591439 4682925568 model_lib_v2.py:705] Step 20900 per-step time 0.810s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0727951,\n",
      " 'Loss/localization_loss': 0.01357165,\n",
      " 'Loss/regularization_loss': 0.08051393,\n",
      " 'Loss/total_loss': 0.16688068,\n",
      " 'learning_rate': 0.051626723}\n",
      "I1020 00:36:51.593760 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0727951,\n",
      " 'Loss/localization_loss': 0.01357165,\n",
      " 'Loss/regularization_loss': 0.08051393,\n",
      " 'Loss/total_loss': 0.16688068,\n",
      " 'learning_rate': 0.051626723}\n",
      "INFO:tensorflow:Step 21000 per-step time 0.793s\n",
      "I1020 00:38:10.857531 4682925568 model_lib_v2.py:705] Step 21000 per-step time 0.793s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06442827,\n",
      " 'Loss/localization_loss': 0.031905964,\n",
      " 'Loss/regularization_loss': 0.080254816,\n",
      " 'Loss/total_loss': 0.17658904,\n",
      " 'learning_rate': 0.0513811}\n",
      "I1020 00:38:10.857765 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06442827,\n",
      " 'Loss/localization_loss': 0.031905964,\n",
      " 'Loss/regularization_loss': 0.080254816,\n",
      " 'Loss/total_loss': 0.17658904,\n",
      " 'learning_rate': 0.0513811}\n",
      "INFO:tensorflow:Step 21100 per-step time 0.802s\n",
      "I1020 00:39:31.046618 4682925568 model_lib_v2.py:705] Step 21100 per-step time 0.802s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.09436052,\n",
      " 'Loss/localization_loss': 0.032261994,\n",
      " 'Loss/regularization_loss': 0.08000818,\n",
      " 'Loss/total_loss': 0.20663069,\n",
      " 'learning_rate': 0.051135015}\n",
      "I1020 00:39:31.046895 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.09436052,\n",
      " 'Loss/localization_loss': 0.032261994,\n",
      " 'Loss/regularization_loss': 0.08000818,\n",
      " 'Loss/total_loss': 0.20663069,\n",
      " 'learning_rate': 0.051135015}\n",
      "INFO:tensorflow:Step 21200 per-step time 0.806s\n",
      "I1020 00:40:51.628874 4682925568 model_lib_v2.py:705] Step 21200 per-step time 0.806s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06988157,\n",
      " 'Loss/localization_loss': 0.03505108,\n",
      " 'Loss/regularization_loss': 0.07980716,\n",
      " 'Loss/total_loss': 0.18473981,\n",
      " 'learning_rate': 0.050888464}\n",
      "I1020 00:40:51.631415 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06988157,\n",
      " 'Loss/localization_loss': 0.03505108,\n",
      " 'Loss/regularization_loss': 0.07980716,\n",
      " 'Loss/total_loss': 0.18473981,\n",
      " 'learning_rate': 0.050888464}\n",
      "INFO:tensorflow:Step 21300 per-step time 0.797s\n",
      "I1020 00:42:11.343053 4682925568 model_lib_v2.py:705] Step 21300 per-step time 0.797s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07287113,\n",
      " 'Loss/localization_loss': 0.04630226,\n",
      " 'Loss/regularization_loss': 0.07955019,\n",
      " 'Loss/total_loss': 0.19872358,\n",
      " 'learning_rate': 0.05064147}\n",
      "I1020 00:42:11.343304 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07287113,\n",
      " 'Loss/localization_loss': 0.04630226,\n",
      " 'Loss/regularization_loss': 0.07955019,\n",
      " 'Loss/total_loss': 0.19872358,\n",
      " 'learning_rate': 0.05064147}\n",
      "INFO:tensorflow:Step 21400 per-step time 0.813s\n",
      "I1020 00:43:32.619863 4682925568 model_lib_v2.py:705] Step 21400 per-step time 0.813s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.064689055,\n",
      " 'Loss/localization_loss': 0.051158026,\n",
      " 'Loss/regularization_loss': 0.07931294,\n",
      " 'Loss/total_loss': 0.19516003,\n",
      " 'learning_rate': 0.05039404}\n",
      "I1020 00:43:32.620094 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.064689055,\n",
      " 'Loss/localization_loss': 0.051158026,\n",
      " 'Loss/regularization_loss': 0.07931294,\n",
      " 'Loss/total_loss': 0.19516003,\n",
      " 'learning_rate': 0.05039404}\n",
      "INFO:tensorflow:Step 21500 per-step time 0.795s\n",
      "I1020 00:44:52.107333 4682925568 model_lib_v2.py:705] Step 21500 per-step time 0.795s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05704255,\n",
      " 'Loss/localization_loss': 0.029404644,\n",
      " 'Loss/regularization_loss': 0.07906157,\n",
      " 'Loss/total_loss': 0.16550876,\n",
      " 'learning_rate': 0.050146177}\n",
      "I1020 00:44:52.109926 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05704255,\n",
      " 'Loss/localization_loss': 0.029404644,\n",
      " 'Loss/regularization_loss': 0.07906157,\n",
      " 'Loss/total_loss': 0.16550876,\n",
      " 'learning_rate': 0.050146177}\n",
      "INFO:tensorflow:Step 21600 per-step time 0.797s\n",
      "I1020 00:46:11.768762 4682925568 model_lib_v2.py:705] Step 21600 per-step time 0.797s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08969789,\n",
      " 'Loss/localization_loss': 0.021348314,\n",
      " 'Loss/regularization_loss': 0.07884036,\n",
      " 'Loss/total_loss': 0.18988657,\n",
      " 'learning_rate': 0.04989791}\n",
      "I1020 00:46:11.769126 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08969789,\n",
      " 'Loss/localization_loss': 0.021348314,\n",
      " 'Loss/regularization_loss': 0.07884036,\n",
      " 'Loss/total_loss': 0.18988657,\n",
      " 'learning_rate': 0.04989791}\n",
      "INFO:tensorflow:Step 21700 per-step time 0.791s\n",
      "I1020 00:47:30.867761 4682925568 model_lib_v2.py:705] Step 21700 per-step time 0.791s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.095415816,\n",
      " 'Loss/localization_loss': 0.047717806,\n",
      " 'Loss/regularization_loss': 0.07859437,\n",
      " 'Loss/total_loss': 0.221728,\n",
      " 'learning_rate': 0.049649224}\n",
      "I1020 00:47:30.867987 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.095415816,\n",
      " 'Loss/localization_loss': 0.047717806,\n",
      " 'Loss/regularization_loss': 0.07859437,\n",
      " 'Loss/total_loss': 0.221728,\n",
      " 'learning_rate': 0.049649224}\n",
      "INFO:tensorflow:Step 21800 per-step time 0.792s\n",
      "I1020 00:48:50.094177 4682925568 model_lib_v2.py:705] Step 21800 per-step time 0.792s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.09984125,\n",
      " 'Loss/localization_loss': 0.07241783,\n",
      " 'Loss/regularization_loss': 0.07845989,\n",
      " 'Loss/total_loss': 0.25071898,\n",
      " 'learning_rate': 0.049400143}\n",
      "I1020 00:48:50.094415 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.09984125,\n",
      " 'Loss/localization_loss': 0.07241783,\n",
      " 'Loss/regularization_loss': 0.07845989,\n",
      " 'Loss/total_loss': 0.25071898,\n",
      " 'learning_rate': 0.049400143}\n",
      "INFO:tensorflow:Step 21900 per-step time 0.792s\n",
      "I1020 00:50:09.253324 4682925568 model_lib_v2.py:705] Step 21900 per-step time 0.792s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06532301,\n",
      " 'Loss/localization_loss': 0.026145682,\n",
      " 'Loss/regularization_loss': 0.078274176,\n",
      " 'Loss/total_loss': 0.16974287,\n",
      " 'learning_rate': 0.049150676}\n",
      "I1020 00:50:09.253629 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06532301,\n",
      " 'Loss/localization_loss': 0.026145682,\n",
      " 'Loss/regularization_loss': 0.078274176,\n",
      " 'Loss/total_loss': 0.16974287,\n",
      " 'learning_rate': 0.049150676}\n",
      "INFO:tensorflow:Step 22000 per-step time 0.789s\n",
      "I1020 00:51:28.196400 4682925568 model_lib_v2.py:705] Step 22000 per-step time 0.789s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.065488644,\n",
      " 'Loss/localization_loss': 0.021456085,\n",
      " 'Loss/regularization_loss': 0.07826858,\n",
      " 'Loss/total_loss': 0.16521332,\n",
      " 'learning_rate': 0.048900835}\n",
      "I1020 00:51:28.196632 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.065488644,\n",
      " 'Loss/localization_loss': 0.021456085,\n",
      " 'Loss/regularization_loss': 0.07826858,\n",
      " 'Loss/total_loss': 0.16521332,\n",
      " 'learning_rate': 0.048900835}\n",
      "INFO:tensorflow:Step 22100 per-step time 0.797s\n",
      "I1020 00:52:47.942199 4682925568 model_lib_v2.py:705] Step 22100 per-step time 0.797s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06977343,\n",
      " 'Loss/localization_loss': 0.024082536,\n",
      " 'Loss/regularization_loss': 0.07815393,\n",
      " 'Loss/total_loss': 0.17200989,\n",
      " 'learning_rate': 0.048650626}\n",
      "I1020 00:52:47.942492 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06977343,\n",
      " 'Loss/localization_loss': 0.024082536,\n",
      " 'Loss/regularization_loss': 0.07815393,\n",
      " 'Loss/total_loss': 0.17200989,\n",
      " 'learning_rate': 0.048650626}\n",
      "INFO:tensorflow:Step 22200 per-step time 0.787s\n",
      "I1020 00:54:06.667857 4682925568 model_lib_v2.py:705] Step 22200 per-step time 0.787s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.070843734,\n",
      " 'Loss/localization_loss': 0.014869645,\n",
      " 'Loss/regularization_loss': 0.07793857,\n",
      " 'Loss/total_loss': 0.16365194,\n",
      " 'learning_rate': 0.048400067}\n",
      "I1020 00:54:06.668087 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.070843734,\n",
      " 'Loss/localization_loss': 0.014869645,\n",
      " 'Loss/regularization_loss': 0.07793857,\n",
      " 'Loss/total_loss': 0.16365194,\n",
      " 'learning_rate': 0.048400067}\n",
      "INFO:tensorflow:Step 22300 per-step time 0.813s\n",
      "I1020 00:55:27.946249 4682925568 model_lib_v2.py:705] Step 22300 per-step time 0.813s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.037022743,\n",
      " 'Loss/localization_loss': 0.015787844,\n",
      " 'Loss/regularization_loss': 0.07777117,\n",
      " 'Loss/total_loss': 0.13058177,\n",
      " 'learning_rate': 0.048149157}\n",
      "I1020 00:55:27.948493 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.037022743,\n",
      " 'Loss/localization_loss': 0.015787844,\n",
      " 'Loss/regularization_loss': 0.07777117,\n",
      " 'Loss/total_loss': 0.13058177,\n",
      " 'learning_rate': 0.048149157}\n",
      "INFO:tensorflow:Step 22400 per-step time 0.791s\n",
      "I1020 00:56:47.080512 4682925568 model_lib_v2.py:705] Step 22400 per-step time 0.791s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.10480249,\n",
      " 'Loss/localization_loss': 0.042232417,\n",
      " 'Loss/regularization_loss': 0.077531286,\n",
      " 'Loss/total_loss': 0.22456619,\n",
      " 'learning_rate': 0.04789791}\n",
      "I1020 00:56:47.080759 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.10480249,\n",
      " 'Loss/localization_loss': 0.042232417,\n",
      " 'Loss/regularization_loss': 0.077531286,\n",
      " 'Loss/total_loss': 0.22456619,\n",
      " 'learning_rate': 0.04789791}\n",
      "INFO:tensorflow:Step 22500 per-step time 0.798s\n",
      "I1020 00:58:06.923281 4682925568 model_lib_v2.py:705] Step 22500 per-step time 0.798s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.050956156,\n",
      " 'Loss/localization_loss': 0.017313784,\n",
      " 'Loss/regularization_loss': 0.07730823,\n",
      " 'Loss/total_loss': 0.14557818,\n",
      " 'learning_rate': 0.04764634}\n",
      "I1020 00:58:06.923537 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.050956156,\n",
      " 'Loss/localization_loss': 0.017313784,\n",
      " 'Loss/regularization_loss': 0.07730823,\n",
      " 'Loss/total_loss': 0.14557818,\n",
      " 'learning_rate': 0.04764634}\n",
      "INFO:tensorflow:Step 22600 per-step time 0.804s\n",
      "I1020 00:59:27.312389 4682925568 model_lib_v2.py:705] Step 22600 per-step time 0.804s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05301162,\n",
      " 'Loss/localization_loss': 0.029901009,\n",
      " 'Loss/regularization_loss': 0.07721805,\n",
      " 'Loss/total_loss': 0.16013068,\n",
      " 'learning_rate': 0.047394462}\n",
      "I1020 00:59:27.314558 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05301162,\n",
      " 'Loss/localization_loss': 0.029901009,\n",
      " 'Loss/regularization_loss': 0.07721805,\n",
      " 'Loss/total_loss': 0.16013068,\n",
      " 'learning_rate': 0.047394462}\n",
      "INFO:tensorflow:Step 22700 per-step time 0.808s\n",
      "I1020 01:00:48.140988 4682925568 model_lib_v2.py:705] Step 22700 per-step time 0.808s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.040028416,\n",
      " 'Loss/localization_loss': 0.017570194,\n",
      " 'Loss/regularization_loss': 0.07699823,\n",
      " 'Loss/total_loss': 0.13459684,\n",
      " 'learning_rate': 0.04714227}\n",
      "I1020 01:00:48.141283 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.040028416,\n",
      " 'Loss/localization_loss': 0.017570194,\n",
      " 'Loss/regularization_loss': 0.07699823,\n",
      " 'Loss/total_loss': 0.13459684,\n",
      " 'learning_rate': 0.04714227}\n",
      "INFO:tensorflow:Step 22800 per-step time 0.825s\n",
      "I1020 01:02:10.714351 4682925568 model_lib_v2.py:705] Step 22800 per-step time 0.825s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08338396,\n",
      " 'Loss/localization_loss': 0.034117248,\n",
      " 'Loss/regularization_loss': 0.07677733,\n",
      " 'Loss/total_loss': 0.19427854,\n",
      " 'learning_rate': 0.046889797}\n",
      "I1020 01:02:10.717027 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08338396,\n",
      " 'Loss/localization_loss': 0.034117248,\n",
      " 'Loss/regularization_loss': 0.07677733,\n",
      " 'Loss/total_loss': 0.19427854,\n",
      " 'learning_rate': 0.046889797}\n",
      "INFO:tensorflow:Step 22900 per-step time 0.797s\n",
      "I1020 01:03:30.353167 4682925568 model_lib_v2.py:705] Step 22900 per-step time 0.797s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.056305304,\n",
      " 'Loss/localization_loss': 0.02877403,\n",
      " 'Loss/regularization_loss': 0.07655172,\n",
      " 'Loss/total_loss': 0.16163105,\n",
      " 'learning_rate': 0.04663703}\n",
      "I1020 01:03:30.353409 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.056305304,\n",
      " 'Loss/localization_loss': 0.02877403,\n",
      " 'Loss/regularization_loss': 0.07655172,\n",
      " 'Loss/total_loss': 0.16163105,\n",
      " 'learning_rate': 0.04663703}\n",
      "INFO:tensorflow:Step 23000 per-step time 0.795s\n",
      "I1020 01:04:49.876765 4682925568 model_lib_v2.py:705] Step 23000 per-step time 0.795s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04958364,\n",
      " 'Loss/localization_loss': 0.0292868,\n",
      " 'Loss/regularization_loss': 0.07633653,\n",
      " 'Loss/total_loss': 0.15520698,\n",
      " 'learning_rate': 0.046383996}\n",
      "I1020 01:04:49.877824 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04958364,\n",
      " 'Loss/localization_loss': 0.0292868,\n",
      " 'Loss/regularization_loss': 0.07633653,\n",
      " 'Loss/total_loss': 0.15520698,\n",
      " 'learning_rate': 0.046383996}\n",
      "INFO:tensorflow:Step 23100 per-step time 0.794s\n",
      "I1020 01:06:09.251491 4682925568 model_lib_v2.py:705] Step 23100 per-step time 0.794s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05282642,\n",
      " 'Loss/localization_loss': 0.023263847,\n",
      " 'Loss/regularization_loss': 0.07612292,\n",
      " 'Loss/total_loss': 0.15221319,\n",
      " 'learning_rate': 0.04613069}\n",
      "I1020 01:06:09.251778 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05282642,\n",
      " 'Loss/localization_loss': 0.023263847,\n",
      " 'Loss/regularization_loss': 0.07612292,\n",
      " 'Loss/total_loss': 0.15221319,\n",
      " 'learning_rate': 0.04613069}\n",
      "INFO:tensorflow:Step 23200 per-step time 0.789s\n",
      "I1020 01:07:28.196806 4682925568 model_lib_v2.py:705] Step 23200 per-step time 0.789s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07302118,\n",
      " 'Loss/localization_loss': 0.015191867,\n",
      " 'Loss/regularization_loss': 0.07590169,\n",
      " 'Loss/total_loss': 0.16411474,\n",
      " 'learning_rate': 0.04587714}\n",
      "I1020 01:07:28.199056 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07302118,\n",
      " 'Loss/localization_loss': 0.015191867,\n",
      " 'Loss/regularization_loss': 0.07590169,\n",
      " 'Loss/total_loss': 0.16411474,\n",
      " 'learning_rate': 0.04587714}\n",
      "INFO:tensorflow:Step 23300 per-step time 0.792s\n",
      "I1020 01:08:47.420084 4682925568 model_lib_v2.py:705] Step 23300 per-step time 0.792s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06365843,\n",
      " 'Loss/localization_loss': 0.016351137,\n",
      " 'Loss/regularization_loss': 0.07569931,\n",
      " 'Loss/total_loss': 0.15570888,\n",
      " 'learning_rate': 0.045623355}\n",
      "I1020 01:08:47.420336 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06365843,\n",
      " 'Loss/localization_loss': 0.016351137,\n",
      " 'Loss/regularization_loss': 0.07569931,\n",
      " 'Loss/total_loss': 0.15570888,\n",
      " 'learning_rate': 0.045623355}\n",
      "INFO:tensorflow:Step 23400 per-step time 0.790s\n",
      "I1020 01:10:06.385875 4682925568 model_lib_v2.py:705] Step 23400 per-step time 0.790s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.100450866,\n",
      " 'Loss/localization_loss': 0.04871167,\n",
      " 'Loss/regularization_loss': 0.07554349,\n",
      " 'Loss/total_loss': 0.22470602,\n",
      " 'learning_rate': 0.045369323}\n",
      "I1020 01:10:06.386135 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.100450866,\n",
      " 'Loss/localization_loss': 0.04871167,\n",
      " 'Loss/regularization_loss': 0.07554349,\n",
      " 'Loss/total_loss': 0.22470602,\n",
      " 'learning_rate': 0.045369323}\n",
      "INFO:tensorflow:Step 23500 per-step time 0.810s\n",
      "I1020 01:11:27.439373 4682925568 model_lib_v2.py:705] Step 23500 per-step time 0.810s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.12258494,\n",
      " 'Loss/localization_loss': 0.027492743,\n",
      " 'Loss/regularization_loss': 0.07536744,\n",
      " 'Loss/total_loss': 0.22544512,\n",
      " 'learning_rate': 0.045115083}\n",
      "I1020 01:11:27.441657 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.12258494,\n",
      " 'Loss/localization_loss': 0.027492743,\n",
      " 'Loss/regularization_loss': 0.07536744,\n",
      " 'Loss/total_loss': 0.22544512,\n",
      " 'learning_rate': 0.045115083}\n",
      "INFO:tensorflow:Step 23600 per-step time 0.795s\n",
      "I1020 01:12:46.946308 4682925568 model_lib_v2.py:705] Step 23600 per-step time 0.795s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.093104064,\n",
      " 'Loss/localization_loss': 0.02506319,\n",
      " 'Loss/regularization_loss': 0.075182535,\n",
      " 'Loss/total_loss': 0.19334978,\n",
      " 'learning_rate': 0.044860635}\n",
      "I1020 01:12:46.946538 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.093104064,\n",
      " 'Loss/localization_loss': 0.02506319,\n",
      " 'Loss/regularization_loss': 0.075182535,\n",
      " 'Loss/total_loss': 0.19334978,\n",
      " 'learning_rate': 0.044860635}\n",
      "INFO:tensorflow:Step 23700 per-step time 0.795s\n",
      "I1020 01:14:06.400946 4682925568 model_lib_v2.py:705] Step 23700 per-step time 0.795s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.11480181,\n",
      " 'Loss/localization_loss': 0.016314236,\n",
      " 'Loss/regularization_loss': 0.07501078,\n",
      " 'Loss/total_loss': 0.20612682,\n",
      " 'learning_rate': 0.044605974}\n",
      "I1020 01:14:06.401187 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.11480181,\n",
      " 'Loss/localization_loss': 0.016314236,\n",
      " 'Loss/regularization_loss': 0.07501078,\n",
      " 'Loss/total_loss': 0.20612682,\n",
      " 'learning_rate': 0.044605974}\n",
      "INFO:tensorflow:Step 23800 per-step time 0.792s\n",
      "I1020 01:15:25.604773 4682925568 model_lib_v2.py:705] Step 23800 per-step time 0.792s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.076025106,\n",
      " 'Loss/localization_loss': 0.016420063,\n",
      " 'Loss/regularization_loss': 0.07485284,\n",
      " 'Loss/total_loss': 0.167298,\n",
      " 'learning_rate': 0.044351134}\n",
      "I1020 01:15:25.605003 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.076025106,\n",
      " 'Loss/localization_loss': 0.016420063,\n",
      " 'Loss/regularization_loss': 0.07485284,\n",
      " 'Loss/total_loss': 0.167298,\n",
      " 'learning_rate': 0.044351134}\n",
      "INFO:tensorflow:Step 23900 per-step time 0.791s\n",
      "I1020 01:16:44.724998 4682925568 model_lib_v2.py:705] Step 23900 per-step time 0.791s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.097227916,\n",
      " 'Loss/localization_loss': 0.019709311,\n",
      " 'Loss/regularization_loss': 0.07470616,\n",
      " 'Loss/total_loss': 0.19164339,\n",
      " 'learning_rate': 0.044096105}\n",
      "I1020 01:16:44.725235 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.097227916,\n",
      " 'Loss/localization_loss': 0.019709311,\n",
      " 'Loss/regularization_loss': 0.07470616,\n",
      " 'Loss/total_loss': 0.19164339,\n",
      " 'learning_rate': 0.044096105}\n",
      "INFO:tensorflow:Step 24000 per-step time 0.800s\n",
      "I1020 01:18:04.701295 4682925568 model_lib_v2.py:705] Step 24000 per-step time 0.800s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07786582,\n",
      " 'Loss/localization_loss': 0.04254856,\n",
      " 'Loss/regularization_loss': 0.07450105,\n",
      " 'Loss/total_loss': 0.19491543,\n",
      " 'learning_rate': 0.043840922}\n",
      "I1020 01:18:04.703526 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07786582,\n",
      " 'Loss/localization_loss': 0.04254856,\n",
      " 'Loss/regularization_loss': 0.07450105,\n",
      " 'Loss/total_loss': 0.19491543,\n",
      " 'learning_rate': 0.043840922}\n",
      "INFO:tensorflow:Step 24100 per-step time 0.864s\n",
      "I1020 01:19:31.125945 4682925568 model_lib_v2.py:705] Step 24100 per-step time 0.864s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.047596637,\n",
      " 'Loss/localization_loss': 0.016782634,\n",
      " 'Loss/regularization_loss': 0.07437653,\n",
      " 'Loss/total_loss': 0.1387558,\n",
      " 'learning_rate': 0.043585572}\n",
      "I1020 01:19:31.126491 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.047596637,\n",
      " 'Loss/localization_loss': 0.016782634,\n",
      " 'Loss/regularization_loss': 0.07437653,\n",
      " 'Loss/total_loss': 0.1387558,\n",
      " 'learning_rate': 0.043585572}\n",
      "INFO:tensorflow:Step 24200 per-step time 0.842s\n",
      "I1020 01:20:55.317785 4682925568 model_lib_v2.py:705] Step 24200 per-step time 0.842s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.046058044,\n",
      " 'Loss/localization_loss': 0.02336499,\n",
      " 'Loss/regularization_loss': 0.074172206,\n",
      " 'Loss/total_loss': 0.14359525,\n",
      " 'learning_rate': 0.043330073}\n",
      "I1020 01:20:55.320024 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.046058044,\n",
      " 'Loss/localization_loss': 0.02336499,\n",
      " 'Loss/regularization_loss': 0.074172206,\n",
      " 'Loss/total_loss': 0.14359525,\n",
      " 'learning_rate': 0.043330073}\n",
      "INFO:tensorflow:Step 24300 per-step time 0.815s\n",
      "I1020 01:22:16.826266 4682925568 model_lib_v2.py:705] Step 24300 per-step time 0.815s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.105252266,\n",
      " 'Loss/localization_loss': 0.03637144,\n",
      " 'Loss/regularization_loss': 0.07396609,\n",
      " 'Loss/total_loss': 0.21558979,\n",
      " 'learning_rate': 0.04307444}\n",
      "I1020 01:22:16.829979 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.105252266,\n",
      " 'Loss/localization_loss': 0.03637144,\n",
      " 'Loss/regularization_loss': 0.07396609,\n",
      " 'Loss/total_loss': 0.21558979,\n",
      " 'learning_rate': 0.04307444}\n",
      "INFO:tensorflow:Step 24400 per-step time 0.793s\n",
      "I1020 01:23:36.133562 4682925568 model_lib_v2.py:705] Step 24400 per-step time 0.793s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03552471,\n",
      " 'Loss/localization_loss': 0.018837543,\n",
      " 'Loss/regularization_loss': 0.07379673,\n",
      " 'Loss/total_loss': 0.12815899,\n",
      " 'learning_rate': 0.04281868}\n",
      "I1020 01:23:36.133854 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03552471,\n",
      " 'Loss/localization_loss': 0.018837543,\n",
      " 'Loss/regularization_loss': 0.07379673,\n",
      " 'Loss/total_loss': 0.12815899,\n",
      " 'learning_rate': 0.04281868}\n",
      "INFO:tensorflow:Step 24500 per-step time 0.790s\n",
      "I1020 01:24:55.105976 4682925568 model_lib_v2.py:705] Step 24500 per-step time 0.790s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07855219,\n",
      " 'Loss/localization_loss': 0.03217076,\n",
      " 'Loss/regularization_loss': 0.07360629,\n",
      " 'Loss/total_loss': 0.18432924,\n",
      " 'learning_rate': 0.042562805}\n",
      "I1020 01:24:55.106220 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07855219,\n",
      " 'Loss/localization_loss': 0.03217076,\n",
      " 'Loss/regularization_loss': 0.07360629,\n",
      " 'Loss/total_loss': 0.18432924,\n",
      " 'learning_rate': 0.042562805}\n",
      "INFO:tensorflow:Step 24600 per-step time 0.806s\n",
      "I1020 01:26:15.725307 4682925568 model_lib_v2.py:705] Step 24600 per-step time 0.806s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.052340634,\n",
      " 'Loss/localization_loss': 0.013563607,\n",
      " 'Loss/regularization_loss': 0.07341556,\n",
      " 'Loss/total_loss': 0.1393198,\n",
      " 'learning_rate': 0.04230683}\n",
      "I1020 01:26:15.727638 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.052340634,\n",
      " 'Loss/localization_loss': 0.013563607,\n",
      " 'Loss/regularization_loss': 0.07341556,\n",
      " 'Loss/total_loss': 0.1393198,\n",
      " 'learning_rate': 0.04230683}\n",
      "INFO:tensorflow:Step 24700 per-step time 0.796s\n",
      "I1020 01:27:35.339633 4682925568 model_lib_v2.py:705] Step 24700 per-step time 0.796s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.100960806,\n",
      " 'Loss/localization_loss': 0.039680555,\n",
      " 'Loss/regularization_loss': 0.07330619,\n",
      " 'Loss/total_loss': 0.21394755,\n",
      " 'learning_rate': 0.042050745}\n",
      "I1020 01:27:35.339869 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.100960806,\n",
      " 'Loss/localization_loss': 0.039680555,\n",
      " 'Loss/regularization_loss': 0.07330619,\n",
      " 'Loss/total_loss': 0.21394755,\n",
      " 'learning_rate': 0.042050745}\n",
      "INFO:tensorflow:Step 24800 per-step time 0.796s\n",
      "I1020 01:28:54.973278 4682925568 model_lib_v2.py:705] Step 24800 per-step time 0.796s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.10064702,\n",
      " 'Loss/localization_loss': 0.02801881,\n",
      " 'Loss/regularization_loss': 0.07313783,\n",
      " 'Loss/total_loss': 0.20180365,\n",
      " 'learning_rate': 0.04179459}\n",
      "I1020 01:28:54.973510 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.10064702,\n",
      " 'Loss/localization_loss': 0.02801881,\n",
      " 'Loss/regularization_loss': 0.07313783,\n",
      " 'Loss/total_loss': 0.20180365,\n",
      " 'learning_rate': 0.04179459}\n",
      "INFO:tensorflow:Step 24900 per-step time 0.806s\n",
      "I1020 01:30:15.607871 4682925568 model_lib_v2.py:705] Step 24900 per-step time 0.806s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.060558096,\n",
      " 'Loss/localization_loss': 0.037635237,\n",
      " 'Loss/regularization_loss': 0.07295242,\n",
      " 'Loss/total_loss': 0.17114575,\n",
      " 'learning_rate': 0.04153835}\n",
      "I1020 01:30:15.610090 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.060558096,\n",
      " 'Loss/localization_loss': 0.037635237,\n",
      " 'Loss/regularization_loss': 0.07295242,\n",
      " 'Loss/total_loss': 0.17114575,\n",
      " 'learning_rate': 0.04153835}\n",
      "INFO:tensorflow:Step 25000 per-step time 0.798s\n",
      "I1020 01:31:35.366597 4682925568 model_lib_v2.py:705] Step 25000 per-step time 0.798s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.039133247,\n",
      " 'Loss/localization_loss': 0.016276784,\n",
      " 'Loss/regularization_loss': 0.07277275,\n",
      " 'Loss/total_loss': 0.12818278,\n",
      " 'learning_rate': 0.04128206}\n",
      "I1020 01:31:35.366836 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.039133247,\n",
      " 'Loss/localization_loss': 0.016276784,\n",
      " 'Loss/regularization_loss': 0.07277275,\n",
      " 'Loss/total_loss': 0.12818278,\n",
      " 'learning_rate': 0.04128206}\n",
      "INFO:tensorflow:Step 25100 per-step time 0.798s\n",
      "I1020 01:32:55.129017 4682925568 model_lib_v2.py:705] Step 25100 per-step time 0.798s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.106081486,\n",
      " 'Loss/localization_loss': 0.022422658,\n",
      " 'Loss/regularization_loss': 0.07259412,\n",
      " 'Loss/total_loss': 0.20109826,\n",
      " 'learning_rate': 0.041025713}\n",
      "I1020 01:32:55.129316 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.106081486,\n",
      " 'Loss/localization_loss': 0.022422658,\n",
      " 'Loss/regularization_loss': 0.07259412,\n",
      " 'Loss/total_loss': 0.20109826,\n",
      " 'learning_rate': 0.041025713}\n",
      "INFO:tensorflow:Step 25200 per-step time 0.800s\n",
      "I1020 01:34:15.135377 4682925568 model_lib_v2.py:705] Step 25200 per-step time 0.800s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.051620517,\n",
      " 'Loss/localization_loss': 0.007892628,\n",
      " 'Loss/regularization_loss': 0.0724205,\n",
      " 'Loss/total_loss': 0.13193364,\n",
      " 'learning_rate': 0.04076932}\n",
      "I1020 01:34:15.137612 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.051620517,\n",
      " 'Loss/localization_loss': 0.007892628,\n",
      " 'Loss/regularization_loss': 0.0724205,\n",
      " 'Loss/total_loss': 0.13193364,\n",
      " 'learning_rate': 0.04076932}\n",
      "INFO:tensorflow:Step 25300 per-step time 0.795s\n",
      "I1020 01:35:34.638363 4682925568 model_lib_v2.py:705] Step 25300 per-step time 0.795s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0833213,\n",
      " 'Loss/localization_loss': 0.018135583,\n",
      " 'Loss/regularization_loss': 0.07223377,\n",
      " 'Loss/total_loss': 0.17369065,\n",
      " 'learning_rate': 0.0405129}\n",
      "I1020 01:35:34.638615 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0833213,\n",
      " 'Loss/localization_loss': 0.018135583,\n",
      " 'Loss/regularization_loss': 0.07223377,\n",
      " 'Loss/total_loss': 0.17369065,\n",
      " 'learning_rate': 0.0405129}\n",
      "INFO:tensorflow:Step 25400 per-step time 0.791s\n",
      "I1020 01:36:53.724509 4682925568 model_lib_v2.py:705] Step 25400 per-step time 0.791s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.056516558,\n",
      " 'Loss/localization_loss': 0.019123325,\n",
      " 'Loss/regularization_loss': 0.072068155,\n",
      " 'Loss/total_loss': 0.14770803,\n",
      " 'learning_rate': 0.04025645}\n",
      "I1020 01:36:53.724756 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.056516558,\n",
      " 'Loss/localization_loss': 0.019123325,\n",
      " 'Loss/regularization_loss': 0.072068155,\n",
      " 'Loss/total_loss': 0.14770803,\n",
      " 'learning_rate': 0.04025645}\n",
      "INFO:tensorflow:Step 25500 per-step time 0.788s\n",
      "I1020 01:38:12.569390 4682925568 model_lib_v2.py:705] Step 25500 per-step time 0.788s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04518182,\n",
      " 'Loss/localization_loss': 0.014390753,\n",
      " 'Loss/regularization_loss': 0.07192076,\n",
      " 'Loss/total_loss': 0.13149333,\n",
      " 'learning_rate': 0.039999995}\n",
      "I1020 01:38:12.569657 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04518182,\n",
      " 'Loss/localization_loss': 0.014390753,\n",
      " 'Loss/regularization_loss': 0.07192076,\n",
      " 'Loss/total_loss': 0.13149333,\n",
      " 'learning_rate': 0.039999995}\n",
      "INFO:tensorflow:Step 25600 per-step time 0.791s\n",
      "I1020 01:39:31.689967 4682925568 model_lib_v2.py:705] Step 25600 per-step time 0.791s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.060745012,\n",
      " 'Loss/localization_loss': 0.0074154027,\n",
      " 'Loss/regularization_loss': 0.071741864,\n",
      " 'Loss/total_loss': 0.13990228,\n",
      " 'learning_rate': 0.039743543}\n",
      "I1020 01:39:31.692275 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.060745012,\n",
      " 'Loss/localization_loss': 0.0074154027,\n",
      " 'Loss/regularization_loss': 0.071741864,\n",
      " 'Loss/total_loss': 0.13990228,\n",
      " 'learning_rate': 0.039743543}\n",
      "INFO:tensorflow:Step 25700 per-step time 0.786s\n",
      "I1020 01:40:50.259794 4682925568 model_lib_v2.py:705] Step 25700 per-step time 0.786s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.036638267,\n",
      " 'Loss/localization_loss': 0.0190493,\n",
      " 'Loss/regularization_loss': 0.07155057,\n",
      " 'Loss/total_loss': 0.12723814,\n",
      " 'learning_rate': 0.039487097}\n",
      "I1020 01:40:50.260072 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.036638267,\n",
      " 'Loss/localization_loss': 0.0190493,\n",
      " 'Loss/regularization_loss': 0.07155057,\n",
      " 'Loss/total_loss': 0.12723814,\n",
      " 'learning_rate': 0.039487097}\n",
      "INFO:tensorflow:Step 25800 per-step time 0.786s\n",
      "I1020 01:42:08.802357 4682925568 model_lib_v2.py:705] Step 25800 per-step time 0.786s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04848816,\n",
      " 'Loss/localization_loss': 0.018250462,\n",
      " 'Loss/regularization_loss': 0.0713754,\n",
      " 'Loss/total_loss': 0.13811402,\n",
      " 'learning_rate': 0.03923067}\n",
      "I1020 01:42:08.802586 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04848816,\n",
      " 'Loss/localization_loss': 0.018250462,\n",
      " 'Loss/regularization_loss': 0.0713754,\n",
      " 'Loss/total_loss': 0.13811402,\n",
      " 'learning_rate': 0.03923067}\n",
      "INFO:tensorflow:Step 25900 per-step time 0.824s\n",
      "I1020 01:43:31.161540 4682925568 model_lib_v2.py:705] Step 25900 per-step time 0.824s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.11828607,\n",
      " 'Loss/localization_loss': 0.064017005,\n",
      " 'Loss/regularization_loss': 0.07119507,\n",
      " 'Loss/total_loss': 0.25349814,\n",
      " 'learning_rate': 0.038974285}\n",
      "I1020 01:43:31.163975 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.11828607,\n",
      " 'Loss/localization_loss': 0.064017005,\n",
      " 'Loss/regularization_loss': 0.07119507,\n",
      " 'Loss/total_loss': 0.25349814,\n",
      " 'learning_rate': 0.038974285}\n",
      "INFO:tensorflow:Step 26000 per-step time 0.809s\n",
      "I1020 01:44:52.015187 4682925568 model_lib_v2.py:705] Step 26000 per-step time 0.809s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.045566827,\n",
      " 'Loss/localization_loss': 0.059501883,\n",
      " 'Loss/regularization_loss': 0.071045205,\n",
      " 'Loss/total_loss': 0.17611392,\n",
      " 'learning_rate': 0.038717933}\n",
      "I1020 01:44:52.015427 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.045566827,\n",
      " 'Loss/localization_loss': 0.059501883,\n",
      " 'Loss/regularization_loss': 0.071045205,\n",
      " 'Loss/total_loss': 0.17611392,\n",
      " 'learning_rate': 0.038717933}\n",
      "INFO:tensorflow:Step 26100 per-step time 0.799s\n",
      "I1020 01:46:11.961199 4682925568 model_lib_v2.py:705] Step 26100 per-step time 0.799s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.089697614,\n",
      " 'Loss/localization_loss': 0.032799024,\n",
      " 'Loss/regularization_loss': 0.070868276,\n",
      " 'Loss/total_loss': 0.19336492,\n",
      " 'learning_rate': 0.038461637}\n",
      "I1020 01:46:11.961490 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.089697614,\n",
      " 'Loss/localization_loss': 0.032799024,\n",
      " 'Loss/regularization_loss': 0.070868276,\n",
      " 'Loss/total_loss': 0.19336492,\n",
      " 'learning_rate': 0.038461637}\n",
      "INFO:tensorflow:Step 26200 per-step time 0.810s\n",
      "I1020 01:47:32.990150 4682925568 model_lib_v2.py:705] Step 26200 per-step time 0.810s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07382243,\n",
      " 'Loss/localization_loss': 0.057373237,\n",
      " 'Loss/regularization_loss': 0.07069517,\n",
      " 'Loss/total_loss': 0.20189083,\n",
      " 'learning_rate': 0.038205404}\n",
      "I1020 01:47:32.992629 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07382243,\n",
      " 'Loss/localization_loss': 0.057373237,\n",
      " 'Loss/regularization_loss': 0.07069517,\n",
      " 'Loss/total_loss': 0.20189083,\n",
      " 'learning_rate': 0.038205404}\n",
      "INFO:tensorflow:Step 26300 per-step time 0.840s\n",
      "I1020 01:48:57.003048 4682925568 model_lib_v2.py:705] Step 26300 per-step time 0.840s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.11134323,\n",
      " 'Loss/localization_loss': 0.03979071,\n",
      " 'Loss/regularization_loss': 0.070580356,\n",
      " 'Loss/total_loss': 0.22171429,\n",
      " 'learning_rate': 0.03794924}\n",
      "I1020 01:48:57.003300 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.11134323,\n",
      " 'Loss/localization_loss': 0.03979071,\n",
      " 'Loss/regularization_loss': 0.070580356,\n",
      " 'Loss/total_loss': 0.22171429,\n",
      " 'learning_rate': 0.03794924}\n",
      "INFO:tensorflow:Step 26400 per-step time 0.787s\n",
      "I1020 01:50:15.745033 4682925568 model_lib_v2.py:705] Step 26400 per-step time 0.787s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.088347055,\n",
      " 'Loss/localization_loss': 0.028974455,\n",
      " 'Loss/regularization_loss': 0.07042257,\n",
      " 'Loss/total_loss': 0.18774408,\n",
      " 'learning_rate': 0.03769317}\n",
      "I1020 01:50:15.745264 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.088347055,\n",
      " 'Loss/localization_loss': 0.028974455,\n",
      " 'Loss/regularization_loss': 0.07042257,\n",
      " 'Loss/total_loss': 0.18774408,\n",
      " 'learning_rate': 0.03769317}\n",
      "INFO:tensorflow:Step 26500 per-step time 0.792s\n",
      "I1020 01:51:34.956358 4682925568 model_lib_v2.py:705] Step 26500 per-step time 0.792s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06294314,\n",
      " 'Loss/localization_loss': 0.0218977,\n",
      " 'Loss/regularization_loss': 0.07025002,\n",
      " 'Loss/total_loss': 0.15509085,\n",
      " 'learning_rate': 0.037437186}\n",
      "I1020 01:51:34.956598 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06294314,\n",
      " 'Loss/localization_loss': 0.0218977,\n",
      " 'Loss/regularization_loss': 0.07025002,\n",
      " 'Loss/total_loss': 0.15509085,\n",
      " 'learning_rate': 0.037437186}\n",
      "INFO:tensorflow:Step 26600 per-step time 0.799s\n",
      "I1020 01:52:54.858757 4682925568 model_lib_v2.py:705] Step 26600 per-step time 0.799s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0666847,\n",
      " 'Loss/localization_loss': 0.011336975,\n",
      " 'Loss/regularization_loss': 0.070075095,\n",
      " 'Loss/total_loss': 0.14809677,\n",
      " 'learning_rate': 0.037181314}\n",
      "I1020 01:52:54.860953 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0666847,\n",
      " 'Loss/localization_loss': 0.011336975,\n",
      " 'Loss/regularization_loss': 0.070075095,\n",
      " 'Loss/total_loss': 0.14809677,\n",
      " 'learning_rate': 0.037181314}\n",
      "INFO:tensorflow:Step 26700 per-step time 0.795s\n",
      "I1020 01:54:14.331317 4682925568 model_lib_v2.py:705] Step 26700 per-step time 0.795s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07472916,\n",
      " 'Loss/localization_loss': 0.044488315,\n",
      " 'Loss/regularization_loss': 0.069944754,\n",
      " 'Loss/total_loss': 0.18916222,\n",
      " 'learning_rate': 0.036925558}\n",
      "I1020 01:54:14.331553 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07472916,\n",
      " 'Loss/localization_loss': 0.044488315,\n",
      " 'Loss/regularization_loss': 0.069944754,\n",
      " 'Loss/total_loss': 0.18916222,\n",
      " 'learning_rate': 0.036925558}\n",
      "INFO:tensorflow:Step 26800 per-step time 0.794s\n",
      "I1020 01:55:33.780069 4682925568 model_lib_v2.py:705] Step 26800 per-step time 0.794s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.032339286,\n",
      " 'Loss/localization_loss': 0.013960927,\n",
      " 'Loss/regularization_loss': 0.069806345,\n",
      " 'Loss/total_loss': 0.116106555,\n",
      " 'learning_rate': 0.03666992}\n",
      "I1020 01:55:33.780354 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.032339286,\n",
      " 'Loss/localization_loss': 0.013960927,\n",
      " 'Loss/regularization_loss': 0.069806345,\n",
      " 'Loss/total_loss': 0.116106555,\n",
      " 'learning_rate': 0.03666992}\n",
      "INFO:tensorflow:Step 26900 per-step time 0.840s\n",
      "I1020 01:56:57.845545 4682925568 model_lib_v2.py:705] Step 26900 per-step time 0.840s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.054566354,\n",
      " 'Loss/localization_loss': 0.014492766,\n",
      " 'Loss/regularization_loss': 0.06965532,\n",
      " 'Loss/total_loss': 0.13871443,\n",
      " 'learning_rate': 0.03641443}\n",
      "I1020 01:56:57.847823 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.054566354,\n",
      " 'Loss/localization_loss': 0.014492766,\n",
      " 'Loss/regularization_loss': 0.06965532,\n",
      " 'Loss/total_loss': 0.13871443,\n",
      " 'learning_rate': 0.03641443}\n",
      "INFO:tensorflow:Step 27000 per-step time 0.799s\n",
      "I1020 01:58:17.699653 4682925568 model_lib_v2.py:705] Step 27000 per-step time 0.799s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.053103622,\n",
      " 'Loss/localization_loss': 0.032726627,\n",
      " 'Loss/regularization_loss': 0.06951932,\n",
      " 'Loss/total_loss': 0.15534957,\n",
      " 'learning_rate': 0.036159072}\n",
      "I1020 01:58:17.699889 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.053103622,\n",
      " 'Loss/localization_loss': 0.032726627,\n",
      " 'Loss/regularization_loss': 0.06951932,\n",
      " 'Loss/total_loss': 0.15534957,\n",
      " 'learning_rate': 0.036159072}\n",
      "INFO:tensorflow:Step 27100 per-step time 0.837s\n",
      "I1020 01:59:41.365005 4682925568 model_lib_v2.py:705] Step 27100 per-step time 0.837s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.17188549,\n",
      " 'Loss/localization_loss': 0.05785498,\n",
      " 'Loss/regularization_loss': 0.06937896,\n",
      " 'Loss/total_loss': 0.2991194,\n",
      " 'learning_rate': 0.035903886}\n",
      "I1020 01:59:41.367719 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.17188549,\n",
      " 'Loss/localization_loss': 0.05785498,\n",
      " 'Loss/regularization_loss': 0.06937896,\n",
      " 'Loss/total_loss': 0.2991194,\n",
      " 'learning_rate': 0.035903886}\n",
      "INFO:tensorflow:Step 27200 per-step time 0.809s\n",
      "I1020 02:01:02.228730 4682925568 model_lib_v2.py:705] Step 27200 per-step time 0.809s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05442818,\n",
      " 'Loss/localization_loss': 0.025869431,\n",
      " 'Loss/regularization_loss': 0.06922121,\n",
      " 'Loss/total_loss': 0.14951882,\n",
      " 'learning_rate': 0.035648864}\n",
      "I1020 02:01:02.231077 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05442818,\n",
      " 'Loss/localization_loss': 0.025869431,\n",
      " 'Loss/regularization_loss': 0.06922121,\n",
      " 'Loss/total_loss': 0.14951882,\n",
      " 'learning_rate': 0.035648864}\n",
      "INFO:tensorflow:Step 27300 per-step time 0.792s\n",
      "I1020 02:02:21.443509 4682925568 model_lib_v2.py:705] Step 27300 per-step time 0.792s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06319362,\n",
      " 'Loss/localization_loss': 0.020753875,\n",
      " 'Loss/regularization_loss': 0.06908081,\n",
      " 'Loss/total_loss': 0.15302831,\n",
      " 'learning_rate': 0.03539402}\n",
      "I1020 02:02:21.443761 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06319362,\n",
      " 'Loss/localization_loss': 0.020753875,\n",
      " 'Loss/regularization_loss': 0.06908081,\n",
      " 'Loss/total_loss': 0.15302831,\n",
      " 'learning_rate': 0.03539402}\n",
      "INFO:tensorflow:Step 27400 per-step time 0.795s\n",
      "I1020 02:03:40.908615 4682925568 model_lib_v2.py:705] Step 27400 per-step time 0.795s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06671675,\n",
      " 'Loss/localization_loss': 0.01649539,\n",
      " 'Loss/regularization_loss': 0.068925016,\n",
      " 'Loss/total_loss': 0.15213716,\n",
      " 'learning_rate': 0.035139363}\n",
      "I1020 02:03:40.908947 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06671675,\n",
      " 'Loss/localization_loss': 0.01649539,\n",
      " 'Loss/regularization_loss': 0.068925016,\n",
      " 'Loss/total_loss': 0.15213716,\n",
      " 'learning_rate': 0.035139363}\n",
      "INFO:tensorflow:Step 27500 per-step time 0.806s\n",
      "I1020 02:05:01.500745 4682925568 model_lib_v2.py:705] Step 27500 per-step time 0.806s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06815495,\n",
      " 'Loss/localization_loss': 0.020678416,\n",
      " 'Loss/regularization_loss': 0.06876253,\n",
      " 'Loss/total_loss': 0.1575959,\n",
      " 'learning_rate': 0.03488491}\n",
      "I1020 02:05:01.502972 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06815495,\n",
      " 'Loss/localization_loss': 0.020678416,\n",
      " 'Loss/regularization_loss': 0.06876253,\n",
      " 'Loss/total_loss': 0.1575959,\n",
      " 'learning_rate': 0.03488491}\n",
      "INFO:tensorflow:Step 27600 per-step time 0.793s\n",
      "I1020 02:06:20.810588 4682925568 model_lib_v2.py:705] Step 27600 per-step time 0.793s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.094488055,\n",
      " 'Loss/localization_loss': 0.09649496,\n",
      " 'Loss/regularization_loss': 0.06860188,\n",
      " 'Loss/total_loss': 0.2595849,\n",
      " 'learning_rate': 0.034630664}\n",
      "I1020 02:06:20.810821 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.094488055,\n",
      " 'Loss/localization_loss': 0.09649496,\n",
      " 'Loss/regularization_loss': 0.06860188,\n",
      " 'Loss/total_loss': 0.2595849,\n",
      " 'learning_rate': 0.034630664}\n",
      "INFO:tensorflow:Step 27700 per-step time 0.792s\n",
      "I1020 02:07:39.990831 4682925568 model_lib_v2.py:705] Step 27700 per-step time 0.792s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06270757,\n",
      " 'Loss/localization_loss': 0.017947758,\n",
      " 'Loss/regularization_loss': 0.06845156,\n",
      " 'Loss/total_loss': 0.14910689,\n",
      " 'learning_rate': 0.034376644}\n",
      "I1020 02:07:39.991060 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06270757,\n",
      " 'Loss/localization_loss': 0.017947758,\n",
      " 'Loss/regularization_loss': 0.06845156,\n",
      " 'Loss/total_loss': 0.14910689,\n",
      " 'learning_rate': 0.034376644}\n",
      "INFO:tensorflow:Step 27800 per-step time 0.793s\n",
      "I1020 02:08:59.262842 4682925568 model_lib_v2.py:705] Step 27800 per-step time 0.793s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08833846,\n",
      " 'Loss/localization_loss': 0.035296734,\n",
      " 'Loss/regularization_loss': 0.06830114,\n",
      " 'Loss/total_loss': 0.19193633,\n",
      " 'learning_rate': 0.034122847}\n",
      "I1020 02:08:59.263076 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08833846,\n",
      " 'Loss/localization_loss': 0.035296734,\n",
      " 'Loss/regularization_loss': 0.06830114,\n",
      " 'Loss/total_loss': 0.19193633,\n",
      " 'learning_rate': 0.034122847}\n",
      "INFO:tensorflow:Step 27900 per-step time 0.791s\n",
      "I1020 02:10:18.329108 4682925568 model_lib_v2.py:705] Step 27900 per-step time 0.791s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06485207,\n",
      " 'Loss/localization_loss': 0.051892184,\n",
      " 'Loss/regularization_loss': 0.06815096,\n",
      " 'Loss/total_loss': 0.18489522,\n",
      " 'learning_rate': 0.0338693}\n",
      "I1020 02:10:18.329361 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06485207,\n",
      " 'Loss/localization_loss': 0.051892184,\n",
      " 'Loss/regularization_loss': 0.06815096,\n",
      " 'Loss/total_loss': 0.18489522,\n",
      " 'learning_rate': 0.0338693}\n",
      "INFO:tensorflow:Step 28000 per-step time 0.784s\n",
      "I1020 02:11:36.742568 4682925568 model_lib_v2.py:705] Step 28000 per-step time 0.784s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08267367,\n",
      " 'Loss/localization_loss': 0.038139537,\n",
      " 'Loss/regularization_loss': 0.068004265,\n",
      " 'Loss/total_loss': 0.18881747,\n",
      " 'learning_rate': 0.033615995}\n",
      "I1020 02:11:36.742804 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08267367,\n",
      " 'Loss/localization_loss': 0.038139537,\n",
      " 'Loss/regularization_loss': 0.068004265,\n",
      " 'Loss/total_loss': 0.18881747,\n",
      " 'learning_rate': 0.033615995}\n",
      "INFO:tensorflow:Step 28100 per-step time 0.797s\n",
      "I1020 02:12:56.418456 4682925568 model_lib_v2.py:705] Step 28100 per-step time 0.797s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.061381,\n",
      " 'Loss/localization_loss': 0.020746546,\n",
      " 'Loss/regularization_loss': 0.06792966,\n",
      " 'Loss/total_loss': 0.15005721,\n",
      " 'learning_rate': 0.033362966}\n",
      "I1020 02:12:56.418813 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.061381,\n",
      " 'Loss/localization_loss': 0.020746546,\n",
      " 'Loss/regularization_loss': 0.06792966,\n",
      " 'Loss/total_loss': 0.15005721,\n",
      " 'learning_rate': 0.033362966}\n",
      "INFO:tensorflow:Step 28200 per-step time 0.792s\n",
      "I1020 02:14:15.617551 4682925568 model_lib_v2.py:705] Step 28200 per-step time 0.792s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.099509165,\n",
      " 'Loss/localization_loss': 0.022395354,\n",
      " 'Loss/regularization_loss': 0.067785576,\n",
      " 'Loss/total_loss': 0.1896901,\n",
      " 'learning_rate': 0.033110205}\n",
      "I1020 02:14:15.617921 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.099509165,\n",
      " 'Loss/localization_loss': 0.022395354,\n",
      " 'Loss/regularization_loss': 0.067785576,\n",
      " 'Loss/total_loss': 0.1896901,\n",
      " 'learning_rate': 0.033110205}\n",
      "INFO:tensorflow:Step 28300 per-step time 0.793s\n",
      "I1020 02:15:34.868350 4682925568 model_lib_v2.py:705] Step 28300 per-step time 0.793s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06846292,\n",
      " 'Loss/localization_loss': 0.034395617,\n",
      " 'Loss/regularization_loss': 0.06763861,\n",
      " 'Loss/total_loss': 0.17049715,\n",
      " 'learning_rate': 0.032857716}\n",
      "I1020 02:15:34.868572 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06846292,\n",
      " 'Loss/localization_loss': 0.034395617,\n",
      " 'Loss/regularization_loss': 0.06763861,\n",
      " 'Loss/total_loss': 0.17049715,\n",
      " 'learning_rate': 0.032857716}\n",
      "INFO:tensorflow:Step 28400 per-step time 0.814s\n",
      "I1020 02:16:56.306756 4682925568 model_lib_v2.py:705] Step 28400 per-step time 0.814s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.10411965,\n",
      " 'Loss/localization_loss': 0.032928552,\n",
      " 'Loss/regularization_loss': 0.067495085,\n",
      " 'Loss/total_loss': 0.20454329,\n",
      " 'learning_rate': 0.032605536}\n",
      "I1020 02:16:56.308933 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.10411965,\n",
      " 'Loss/localization_loss': 0.032928552,\n",
      " 'Loss/regularization_loss': 0.067495085,\n",
      " 'Loss/total_loss': 0.20454329,\n",
      " 'learning_rate': 0.032605536}\n",
      "INFO:tensorflow:Step 28500 per-step time 0.792s\n",
      "I1020 02:18:15.480975 4682925568 model_lib_v2.py:705] Step 28500 per-step time 0.792s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05988906,\n",
      " 'Loss/localization_loss': 0.016850155,\n",
      " 'Loss/regularization_loss': 0.06735664,\n",
      " 'Loss/total_loss': 0.14409585,\n",
      " 'learning_rate': 0.032353655}\n",
      "I1020 02:18:15.481270 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05988906,\n",
      " 'Loss/localization_loss': 0.016850155,\n",
      " 'Loss/regularization_loss': 0.06735664,\n",
      " 'Loss/total_loss': 0.14409585,\n",
      " 'learning_rate': 0.032353655}\n",
      "INFO:tensorflow:Step 28600 per-step time 0.818s\n",
      "I1020 02:19:37.261953 4682925568 model_lib_v2.py:705] Step 28600 per-step time 0.818s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.043765835,\n",
      " 'Loss/localization_loss': 0.034212425,\n",
      " 'Loss/regularization_loss': 0.067214414,\n",
      " 'Loss/total_loss': 0.14519268,\n",
      " 'learning_rate': 0.032102082}\n",
      "I1020 02:19:37.264197 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.043765835,\n",
      " 'Loss/localization_loss': 0.034212425,\n",
      " 'Loss/regularization_loss': 0.067214414,\n",
      " 'Loss/total_loss': 0.14519268,\n",
      " 'learning_rate': 0.032102082}\n",
      "INFO:tensorflow:Step 28700 per-step time 0.794s\n",
      "I1020 02:20:56.694351 4682925568 model_lib_v2.py:705] Step 28700 per-step time 0.794s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.050271682,\n",
      " 'Loss/localization_loss': 0.0052225916,\n",
      " 'Loss/regularization_loss': 0.067075014,\n",
      " 'Loss/total_loss': 0.12256929,\n",
      " 'learning_rate': 0.03185084}\n",
      "I1020 02:20:56.694586 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.050271682,\n",
      " 'Loss/localization_loss': 0.0052225916,\n",
      " 'Loss/regularization_loss': 0.067075014,\n",
      " 'Loss/total_loss': 0.12256929,\n",
      " 'learning_rate': 0.03185084}\n",
      "INFO:tensorflow:Step 28800 per-step time 0.794s\n",
      "I1020 02:22:16.115802 4682925568 model_lib_v2.py:705] Step 28800 per-step time 0.794s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07202767,\n",
      " 'Loss/localization_loss': 0.011906058,\n",
      " 'Loss/regularization_loss': 0.06693287,\n",
      " 'Loss/total_loss': 0.1508666,\n",
      " 'learning_rate': 0.031599928}\n",
      "I1020 02:22:16.116058 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07202767,\n",
      " 'Loss/localization_loss': 0.011906058,\n",
      " 'Loss/regularization_loss': 0.06693287,\n",
      " 'Loss/total_loss': 0.1508666,\n",
      " 'learning_rate': 0.031599928}\n",
      "INFO:tensorflow:Step 28900 per-step time 0.814s\n",
      "I1020 02:23:37.551333 4682925568 model_lib_v2.py:705] Step 28900 per-step time 0.814s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.040380646,\n",
      " 'Loss/localization_loss': 0.013418767,\n",
      " 'Loss/regularization_loss': 0.06678667,\n",
      " 'Loss/total_loss': 0.12058608,\n",
      " 'learning_rate': 0.03134937}\n",
      "I1020 02:23:37.554052 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.040380646,\n",
      " 'Loss/localization_loss': 0.013418767,\n",
      " 'Loss/regularization_loss': 0.06678667,\n",
      " 'Loss/total_loss': 0.12058608,\n",
      " 'learning_rate': 0.03134937}\n",
      "INFO:tensorflow:Step 29000 per-step time 0.807s\n",
      "I1020 02:24:58.252354 4682925568 model_lib_v2.py:705] Step 29000 per-step time 0.807s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.047167074,\n",
      " 'Loss/localization_loss': 0.026107185,\n",
      " 'Loss/regularization_loss': 0.06664398,\n",
      " 'Loss/total_loss': 0.13991824,\n",
      " 'learning_rate': 0.031099161}\n",
      "I1020 02:24:58.255275 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.047167074,\n",
      " 'Loss/localization_loss': 0.026107185,\n",
      " 'Loss/regularization_loss': 0.06664398,\n",
      " 'Loss/total_loss': 0.13991824,\n",
      " 'learning_rate': 0.031099161}\n",
      "INFO:tensorflow:Step 29100 per-step time 0.797s\n",
      "I1020 02:26:17.950773 4682925568 model_lib_v2.py:705] Step 29100 per-step time 0.797s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03654942,\n",
      " 'Loss/localization_loss': 0.015248718,\n",
      " 'Loss/regularization_loss': 0.06650589,\n",
      " 'Loss/total_loss': 0.11830402,\n",
      " 'learning_rate': 0.030849317}\n",
      "I1020 02:26:17.951066 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03654942,\n",
      " 'Loss/localization_loss': 0.015248718,\n",
      " 'Loss/regularization_loss': 0.06650589,\n",
      " 'Loss/total_loss': 0.11830402,\n",
      " 'learning_rate': 0.030849317}\n",
      "INFO:tensorflow:Step 29200 per-step time 0.792s\n",
      "I1020 02:27:37.114635 4682925568 model_lib_v2.py:705] Step 29200 per-step time 0.792s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06806076,\n",
      " 'Loss/localization_loss': 0.017139733,\n",
      " 'Loss/regularization_loss': 0.06636918,\n",
      " 'Loss/total_loss': 0.15156968,\n",
      " 'learning_rate': 0.030599851}\n",
      "I1020 02:27:37.114866 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06806076,\n",
      " 'Loss/localization_loss': 0.017139733,\n",
      " 'Loss/regularization_loss': 0.06636918,\n",
      " 'Loss/total_loss': 0.15156968,\n",
      " 'learning_rate': 0.030599851}\n",
      "INFO:tensorflow:Step 29300 per-step time 0.794s\n",
      "I1020 02:28:56.513264 4682925568 model_lib_v2.py:705] Step 29300 per-step time 0.794s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.062849835,\n",
      " 'Loss/localization_loss': 0.023723045,\n",
      " 'Loss/regularization_loss': 0.0662463,\n",
      " 'Loss/total_loss': 0.15281919,\n",
      " 'learning_rate': 0.030350773}\n",
      "I1020 02:28:56.513494 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.062849835,\n",
      " 'Loss/localization_loss': 0.023723045,\n",
      " 'Loss/regularization_loss': 0.0662463,\n",
      " 'Loss/total_loss': 0.15281919,\n",
      " 'learning_rate': 0.030350773}\n",
      "INFO:tensorflow:Step 29400 per-step time 0.798s\n",
      "I1020 02:30:16.351885 4682925568 model_lib_v2.py:705] Step 29400 per-step time 0.798s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07517535,\n",
      " 'Loss/localization_loss': 0.03590283,\n",
      " 'Loss/regularization_loss': 0.066118725,\n",
      " 'Loss/total_loss': 0.17719692,\n",
      " 'learning_rate': 0.030102087}\n",
      "I1020 02:30:16.352197 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07517535,\n",
      " 'Loss/localization_loss': 0.03590283,\n",
      " 'Loss/regularization_loss': 0.066118725,\n",
      " 'Loss/total_loss': 0.17719692,\n",
      " 'learning_rate': 0.030102087}\n",
      "INFO:tensorflow:Step 29500 per-step time 0.792s\n",
      "I1020 02:31:35.568357 4682925568 model_lib_v2.py:705] Step 29500 per-step time 0.792s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0559604,\n",
      " 'Loss/localization_loss': 0.009912311,\n",
      " 'Loss/regularization_loss': 0.06598996,\n",
      " 'Loss/total_loss': 0.13186267,\n",
      " 'learning_rate': 0.029853815}\n",
      "I1020 02:31:35.568588 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0559604,\n",
      " 'Loss/localization_loss': 0.009912311,\n",
      " 'Loss/regularization_loss': 0.06598996,\n",
      " 'Loss/total_loss': 0.13186267,\n",
      " 'learning_rate': 0.029853815}\n",
      "INFO:tensorflow:Step 29600 per-step time 0.794s\n",
      "I1020 02:32:54.938305 4682925568 model_lib_v2.py:705] Step 29600 per-step time 0.794s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06628843,\n",
      " 'Loss/localization_loss': 0.014771469,\n",
      " 'Loss/regularization_loss': 0.06586197,\n",
      " 'Loss/total_loss': 0.14692187,\n",
      " 'learning_rate': 0.029605951}\n",
      "I1020 02:32:54.938544 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06628843,\n",
      " 'Loss/localization_loss': 0.014771469,\n",
      " 'Loss/regularization_loss': 0.06586197,\n",
      " 'Loss/total_loss': 0.14692187,\n",
      " 'learning_rate': 0.029605951}\n",
      "INFO:tensorflow:Step 29700 per-step time 0.790s\n",
      "I1020 02:34:13.888643 4682925568 model_lib_v2.py:705] Step 29700 per-step time 0.790s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06251322,\n",
      " 'Loss/localization_loss': 0.023068586,\n",
      " 'Loss/regularization_loss': 0.06573482,\n",
      " 'Loss/total_loss': 0.15131661,\n",
      " 'learning_rate': 0.029358525}\n",
      "I1020 02:34:13.888885 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06251322,\n",
      " 'Loss/localization_loss': 0.023068586,\n",
      " 'Loss/regularization_loss': 0.06573482,\n",
      " 'Loss/total_loss': 0.15131661,\n",
      " 'learning_rate': 0.029358525}\n",
      "INFO:tensorflow:Step 29800 per-step time 0.818s\n",
      "I1020 02:35:35.701263 4682925568 model_lib_v2.py:705] Step 29800 per-step time 0.818s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.060373962,\n",
      " 'Loss/localization_loss': 0.020880902,\n",
      " 'Loss/regularization_loss': 0.0656034,\n",
      " 'Loss/total_loss': 0.14685826,\n",
      " 'learning_rate': 0.029111532}\n",
      "I1020 02:35:35.703472 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.060373962,\n",
      " 'Loss/localization_loss': 0.020880902,\n",
      " 'Loss/regularization_loss': 0.0656034,\n",
      " 'Loss/total_loss': 0.14685826,\n",
      " 'learning_rate': 0.029111532}\n",
      "INFO:tensorflow:Step 29900 per-step time 0.799s\n",
      "I1020 02:36:55.635864 4682925568 model_lib_v2.py:705] Step 29900 per-step time 0.799s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.051285267,\n",
      " 'Loss/localization_loss': 0.018929156,\n",
      " 'Loss/regularization_loss': 0.06547607,\n",
      " 'Loss/total_loss': 0.13569048,\n",
      " 'learning_rate': 0.028864982}\n",
      "I1020 02:36:55.636103 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.051285267,\n",
      " 'Loss/localization_loss': 0.018929156,\n",
      " 'Loss/regularization_loss': 0.06547607,\n",
      " 'Loss/total_loss': 0.13569048,\n",
      " 'learning_rate': 0.028864982}\n",
      "INFO:tensorflow:Step 30000 per-step time 0.808s\n",
      "I1020 02:38:16.408246 4682925568 model_lib_v2.py:705] Step 30000 per-step time 0.808s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05402556,\n",
      " 'Loss/localization_loss': 0.030196935,\n",
      " 'Loss/regularization_loss': 0.06535031,\n",
      " 'Loss/total_loss': 0.1495728,\n",
      " 'learning_rate': 0.028618898}\n",
      "I1020 02:38:16.410538 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05402556,\n",
      " 'Loss/localization_loss': 0.030196935,\n",
      " 'Loss/regularization_loss': 0.06535031,\n",
      " 'Loss/total_loss': 0.1495728,\n",
      " 'learning_rate': 0.028618898}\n",
      "INFO:tensorflow:Step 30100 per-step time 0.800s\n",
      "I1020 02:39:36.400202 4682925568 model_lib_v2.py:705] Step 30100 per-step time 0.800s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.13688903,\n",
      " 'Loss/localization_loss': 0.0282421,\n",
      " 'Loss/regularization_loss': 0.06523066,\n",
      " 'Loss/total_loss': 0.23036179,\n",
      " 'learning_rate': 0.02837327}\n",
      "I1020 02:39:36.400479 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.13688903,\n",
      " 'Loss/localization_loss': 0.0282421,\n",
      " 'Loss/regularization_loss': 0.06523066,\n",
      " 'Loss/total_loss': 0.23036179,\n",
      " 'learning_rate': 0.02837327}\n",
      "INFO:tensorflow:Step 30200 per-step time 0.798s\n",
      "I1020 02:40:56.158320 4682925568 model_lib_v2.py:705] Step 30200 per-step time 0.798s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06307742,\n",
      " 'Loss/localization_loss': 0.030559564,\n",
      " 'Loss/regularization_loss': 0.065114744,\n",
      " 'Loss/total_loss': 0.15875173,\n",
      " 'learning_rate': 0.028128132}\n",
      "I1020 02:40:56.158543 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06307742,\n",
      " 'Loss/localization_loss': 0.030559564,\n",
      " 'Loss/regularization_loss': 0.065114744,\n",
      " 'Loss/total_loss': 0.15875173,\n",
      " 'learning_rate': 0.028128132}\n",
      "INFO:tensorflow:Step 30300 per-step time 0.793s\n",
      "I1020 02:42:15.435450 4682925568 model_lib_v2.py:705] Step 30300 per-step time 0.793s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07599303,\n",
      " 'Loss/localization_loss': 0.015102763,\n",
      " 'Loss/regularization_loss': 0.06499263,\n",
      " 'Loss/total_loss': 0.15608841,\n",
      " 'learning_rate': 0.027883478}\n",
      "I1020 02:42:15.435816 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07599303,\n",
      " 'Loss/localization_loss': 0.015102763,\n",
      " 'Loss/regularization_loss': 0.06499263,\n",
      " 'Loss/total_loss': 0.15608841,\n",
      " 'learning_rate': 0.027883478}\n",
      "INFO:tensorflow:Step 30400 per-step time 0.796s\n",
      "I1020 02:43:35.046797 4682925568 model_lib_v2.py:705] Step 30400 per-step time 0.796s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06893309,\n",
      " 'Loss/localization_loss': 0.018705225,\n",
      " 'Loss/regularization_loss': 0.06487659,\n",
      " 'Loss/total_loss': 0.1525149,\n",
      " 'learning_rate': 0.027639316}\n",
      "I1020 02:43:35.047029 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06893309,\n",
      " 'Loss/localization_loss': 0.018705225,\n",
      " 'Loss/regularization_loss': 0.06487659,\n",
      " 'Loss/total_loss': 0.1525149,\n",
      " 'learning_rate': 0.027639316}\n",
      "INFO:tensorflow:Step 30500 per-step time 0.810s\n",
      "I1020 02:44:56.026940 4682925568 model_lib_v2.py:705] Step 30500 per-step time 0.810s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0665255,\n",
      " 'Loss/localization_loss': 0.03229313,\n",
      " 'Loss/regularization_loss': 0.064754,\n",
      " 'Loss/total_loss': 0.16357264,\n",
      " 'learning_rate': 0.027395668}\n",
      "I1020 02:44:56.029258 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0665255,\n",
      " 'Loss/localization_loss': 0.03229313,\n",
      " 'Loss/regularization_loss': 0.064754,\n",
      " 'Loss/total_loss': 0.16357264,\n",
      " 'learning_rate': 0.027395668}\n",
      "INFO:tensorflow:Step 30600 per-step time 0.805s\n",
      "I1020 02:46:16.493990 4682925568 model_lib_v2.py:705] Step 30600 per-step time 0.805s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.068302065,\n",
      " 'Loss/localization_loss': 0.023411188,\n",
      " 'Loss/regularization_loss': 0.064635955,\n",
      " 'Loss/total_loss': 0.15634921,\n",
      " 'learning_rate': 0.027152538}\n",
      "I1020 02:46:16.494222 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.068302065,\n",
      " 'Loss/localization_loss': 0.023411188,\n",
      " 'Loss/regularization_loss': 0.064635955,\n",
      " 'Loss/total_loss': 0.15634921,\n",
      " 'learning_rate': 0.027152538}\n",
      "INFO:tensorflow:Step 30700 per-step time 0.797s\n",
      "I1020 02:47:36.230660 4682925568 model_lib_v2.py:705] Step 30700 per-step time 0.797s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03616048,\n",
      " 'Loss/localization_loss': 0.0075823353,\n",
      " 'Loss/regularization_loss': 0.06451653,\n",
      " 'Loss/total_loss': 0.10825935,\n",
      " 'learning_rate': 0.026909932}\n",
      "I1020 02:47:36.230896 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03616048,\n",
      " 'Loss/localization_loss': 0.0075823353,\n",
      " 'Loss/regularization_loss': 0.06451653,\n",
      " 'Loss/total_loss': 0.10825935,\n",
      " 'learning_rate': 0.026909932}\n",
      "INFO:tensorflow:Step 30800 per-step time 0.812s\n",
      "I1020 02:48:57.431364 4682925568 model_lib_v2.py:705] Step 30800 per-step time 0.812s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04764149,\n",
      " 'Loss/localization_loss': 0.013608632,\n",
      " 'Loss/regularization_loss': 0.06439449,\n",
      " 'Loss/total_loss': 0.12564461,\n",
      " 'learning_rate': 0.02666787}\n",
      "I1020 02:48:57.433665 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04764149,\n",
      " 'Loss/localization_loss': 0.013608632,\n",
      " 'Loss/regularization_loss': 0.06439449,\n",
      " 'Loss/total_loss': 0.12564461,\n",
      " 'learning_rate': 0.02666787}\n",
      "INFO:tensorflow:Step 30900 per-step time 0.801s\n",
      "I1020 02:50:17.491822 4682925568 model_lib_v2.py:705] Step 30900 per-step time 0.801s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.066263005,\n",
      " 'Loss/localization_loss': 0.04067086,\n",
      " 'Loss/regularization_loss': 0.06427774,\n",
      " 'Loss/total_loss': 0.1712116,\n",
      " 'learning_rate': 0.02642635}\n",
      "I1020 02:50:17.492197 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.066263005,\n",
      " 'Loss/localization_loss': 0.04067086,\n",
      " 'Loss/regularization_loss': 0.06427774,\n",
      " 'Loss/total_loss': 0.1712116,\n",
      " 'learning_rate': 0.02642635}\n",
      "INFO:tensorflow:Step 31000 per-step time 0.794s\n",
      "I1020 02:51:36.874493 4682925568 model_lib_v2.py:705] Step 31000 per-step time 0.794s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.10886445,\n",
      " 'Loss/localization_loss': 0.04813111,\n",
      " 'Loss/regularization_loss': 0.064160086,\n",
      " 'Loss/total_loss': 0.22115564,\n",
      " 'learning_rate': 0.026185395}\n",
      "I1020 02:51:36.874727 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.10886445,\n",
      " 'Loss/localization_loss': 0.04813111,\n",
      " 'Loss/regularization_loss': 0.064160086,\n",
      " 'Loss/total_loss': 0.22115564,\n",
      " 'learning_rate': 0.026185395}\n",
      "INFO:tensorflow:Step 31100 per-step time 0.818s\n",
      "I1020 02:52:58.721763 4682925568 model_lib_v2.py:705] Step 31100 per-step time 0.818s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.10015791,\n",
      " 'Loss/localization_loss': 0.038480375,\n",
      " 'Loss/regularization_loss': 0.06404567,\n",
      " 'Loss/total_loss': 0.20268396,\n",
      " 'learning_rate': 0.025945}\n",
      "I1020 02:52:58.724503 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.10015791,\n",
      " 'Loss/localization_loss': 0.038480375,\n",
      " 'Loss/regularization_loss': 0.06404567,\n",
      " 'Loss/total_loss': 0.20268396,\n",
      " 'learning_rate': 0.025945}\n",
      "INFO:tensorflow:Step 31200 per-step time 0.803s\n",
      "I1020 02:54:19.010723 4682925568 model_lib_v2.py:705] Step 31200 per-step time 0.803s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.048147835,\n",
      " 'Loss/localization_loss': 0.02006644,\n",
      " 'Loss/regularization_loss': 0.06393473,\n",
      " 'Loss/total_loss': 0.13214901,\n",
      " 'learning_rate': 0.025705189}\n",
      "I1020 02:54:19.010962 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.048147835,\n",
      " 'Loss/localization_loss': 0.02006644,\n",
      " 'Loss/regularization_loss': 0.06393473,\n",
      " 'Loss/total_loss': 0.13214901,\n",
      " 'learning_rate': 0.025705189}\n",
      "INFO:tensorflow:Step 31300 per-step time 0.793s\n",
      "I1020 02:55:38.330362 4682925568 model_lib_v2.py:705] Step 31300 per-step time 0.793s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07062453,\n",
      " 'Loss/localization_loss': 0.04005902,\n",
      " 'Loss/regularization_loss': 0.063824,\n",
      " 'Loss/total_loss': 0.17450754,\n",
      " 'learning_rate': 0.025465969}\n",
      "I1020 02:55:38.330587 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07062453,\n",
      " 'Loss/localization_loss': 0.04005902,\n",
      " 'Loss/regularization_loss': 0.063824,\n",
      " 'Loss/total_loss': 0.17450754,\n",
      " 'learning_rate': 0.025465969}\n",
      "INFO:tensorflow:Step 31400 per-step time 0.797s\n",
      "I1020 02:56:58.033114 4682925568 model_lib_v2.py:705] Step 31400 per-step time 0.797s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04148083,\n",
      " 'Loss/localization_loss': 0.023270493,\n",
      " 'Loss/regularization_loss': 0.06371349,\n",
      " 'Loss/total_loss': 0.12846482,\n",
      " 'learning_rate': 0.025227336}\n",
      "I1020 02:56:58.033341 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04148083,\n",
      " 'Loss/localization_loss': 0.023270493,\n",
      " 'Loss/regularization_loss': 0.06371349,\n",
      " 'Loss/total_loss': 0.12846482,\n",
      " 'learning_rate': 0.025227336}\n",
      "INFO:tensorflow:Step 31500 per-step time 0.798s\n",
      "I1020 02:58:17.864737 4682925568 model_lib_v2.py:705] Step 31500 per-step time 0.798s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0439687,\n",
      " 'Loss/localization_loss': 0.014317329,\n",
      " 'Loss/regularization_loss': 0.06360613,\n",
      " 'Loss/total_loss': 0.121892154,\n",
      " 'learning_rate': 0.024989318}\n",
      "I1020 02:58:17.865039 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0439687,\n",
      " 'Loss/localization_loss': 0.014317329,\n",
      " 'Loss/regularization_loss': 0.06360613,\n",
      " 'Loss/total_loss': 0.121892154,\n",
      " 'learning_rate': 0.024989318}\n",
      "INFO:tensorflow:Step 31600 per-step time 0.812s\n",
      "I1020 02:59:39.055311 4682925568 model_lib_v2.py:705] Step 31600 per-step time 0.812s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.051927626,\n",
      " 'Loss/localization_loss': 0.008443626,\n",
      " 'Loss/regularization_loss': 0.06349729,\n",
      " 'Loss/total_loss': 0.12386854,\n",
      " 'learning_rate': 0.024751918}\n",
      "I1020 02:59:39.057670 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.051927626,\n",
      " 'Loss/localization_loss': 0.008443626,\n",
      " 'Loss/regularization_loss': 0.06349729,\n",
      " 'Loss/total_loss': 0.12386854,\n",
      " 'learning_rate': 0.024751918}\n",
      "INFO:tensorflow:Step 31700 per-step time 0.813s\n",
      "I1020 03:01:00.363510 4682925568 model_lib_v2.py:705] Step 31700 per-step time 0.813s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.086343974,\n",
      " 'Loss/localization_loss': 0.019021261,\n",
      " 'Loss/regularization_loss': 0.06342585,\n",
      " 'Loss/total_loss': 0.16879109,\n",
      " 'learning_rate': 0.024515137}\n",
      "I1020 03:01:00.363817 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.086343974,\n",
      " 'Loss/localization_loss': 0.019021261,\n",
      " 'Loss/regularization_loss': 0.06342585,\n",
      " 'Loss/total_loss': 0.16879109,\n",
      " 'learning_rate': 0.024515137}\n",
      "INFO:tensorflow:Step 31800 per-step time 0.807s\n",
      "I1020 03:02:21.023994 4682925568 model_lib_v2.py:705] Step 31800 per-step time 0.807s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.037676554,\n",
      " 'Loss/localization_loss': 0.0073390794,\n",
      " 'Loss/regularization_loss': 0.063326925,\n",
      " 'Loss/total_loss': 0.10834256,\n",
      " 'learning_rate': 0.024278998}\n",
      "I1020 03:02:21.024235 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.037676554,\n",
      " 'Loss/localization_loss': 0.0073390794,\n",
      " 'Loss/regularization_loss': 0.063326925,\n",
      " 'Loss/total_loss': 0.10834256,\n",
      " 'learning_rate': 0.024278998}\n",
      "INFO:tensorflow:Step 31900 per-step time 0.794s\n",
      "I1020 03:03:40.465406 4682925568 model_lib_v2.py:705] Step 31900 per-step time 0.794s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.029450616,\n",
      " 'Loss/localization_loss': 0.022901647,\n",
      " 'Loss/regularization_loss': 0.063221455,\n",
      " 'Loss/total_loss': 0.11557372,\n",
      " 'learning_rate': 0.0240435}\n",
      "I1020 03:03:40.465654 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.029450616,\n",
      " 'Loss/localization_loss': 0.022901647,\n",
      " 'Loss/regularization_loss': 0.063221455,\n",
      " 'Loss/total_loss': 0.11557372,\n",
      " 'learning_rate': 0.0240435}\n",
      "INFO:tensorflow:Step 32000 per-step time 0.799s\n",
      "I1020 03:05:00.378303 4682925568 model_lib_v2.py:705] Step 32000 per-step time 0.799s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.049821455,\n",
      " 'Loss/localization_loss': 0.010377317,\n",
      " 'Loss/regularization_loss': 0.06311602,\n",
      " 'Loss/total_loss': 0.1233148,\n",
      " 'learning_rate': 0.023808666}\n",
      "I1020 03:05:00.378540 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.049821455,\n",
      " 'Loss/localization_loss': 0.010377317,\n",
      " 'Loss/regularization_loss': 0.06311602,\n",
      " 'Loss/total_loss': 0.1233148,\n",
      " 'learning_rate': 0.023808666}\n",
      "INFO:tensorflow:Step 32100 per-step time 0.812s\n",
      "I1020 03:06:21.612825 4682925568 model_lib_v2.py:705] Step 32100 per-step time 0.812s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.09088163,\n",
      " 'Loss/localization_loss': 0.0269801,\n",
      " 'Loss/regularization_loss': 0.06301322,\n",
      " 'Loss/total_loss': 0.18087494,\n",
      " 'learning_rate': 0.023574492}\n",
      "I1020 03:06:21.614909 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.09088163,\n",
      " 'Loss/localization_loss': 0.0269801,\n",
      " 'Loss/regularization_loss': 0.06301322,\n",
      " 'Loss/total_loss': 0.18087494,\n",
      " 'learning_rate': 0.023574492}\n",
      "INFO:tensorflow:Step 32200 per-step time 0.811s\n",
      "I1020 03:07:42.690083 4682925568 model_lib_v2.py:705] Step 32200 per-step time 0.811s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04183235,\n",
      " 'Loss/localization_loss': 0.025401657,\n",
      " 'Loss/regularization_loss': 0.06290806,\n",
      " 'Loss/total_loss': 0.13014206,\n",
      " 'learning_rate': 0.023340987}\n",
      "I1020 03:07:42.690312 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04183235,\n",
      " 'Loss/localization_loss': 0.025401657,\n",
      " 'Loss/regularization_loss': 0.06290806,\n",
      " 'Loss/total_loss': 0.13014206,\n",
      " 'learning_rate': 0.023340987}\n",
      "INFO:tensorflow:Step 32300 per-step time 0.802s\n",
      "I1020 03:09:02.915908 4682925568 model_lib_v2.py:705] Step 32300 per-step time 0.802s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06343239,\n",
      " 'Loss/localization_loss': 0.019348646,\n",
      " 'Loss/regularization_loss': 0.06280542,\n",
      " 'Loss/total_loss': 0.14558646,\n",
      " 'learning_rate': 0.02310818}\n",
      "I1020 03:09:02.916214 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06343239,\n",
      " 'Loss/localization_loss': 0.019348646,\n",
      " 'Loss/regularization_loss': 0.06280542,\n",
      " 'Loss/total_loss': 0.14558646,\n",
      " 'learning_rate': 0.02310818}\n",
      "INFO:tensorflow:Step 32400 per-step time 0.802s\n",
      "I1020 03:10:23.147409 4682925568 model_lib_v2.py:705] Step 32400 per-step time 0.802s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04492331,\n",
      " 'Loss/localization_loss': 0.011911237,\n",
      " 'Loss/regularization_loss': 0.062703036,\n",
      " 'Loss/total_loss': 0.11953758,\n",
      " 'learning_rate': 0.022876058}\n",
      "I1020 03:10:23.147652 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04492331,\n",
      " 'Loss/localization_loss': 0.011911237,\n",
      " 'Loss/regularization_loss': 0.062703036,\n",
      " 'Loss/total_loss': 0.11953758,\n",
      " 'learning_rate': 0.022876058}\n",
      "INFO:tensorflow:Step 32500 per-step time 0.802s\n",
      "I1020 03:11:43.296666 4682925568 model_lib_v2.py:705] Step 32500 per-step time 0.802s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.052908827,\n",
      " 'Loss/localization_loss': 0.02798565,\n",
      " 'Loss/regularization_loss': 0.062609896,\n",
      " 'Loss/total_loss': 0.14350438,\n",
      " 'learning_rate': 0.022644646}\n",
      "I1020 03:11:43.296898 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.052908827,\n",
      " 'Loss/localization_loss': 0.02798565,\n",
      " 'Loss/regularization_loss': 0.062609896,\n",
      " 'Loss/total_loss': 0.14350438,\n",
      " 'learning_rate': 0.022644646}\n",
      "INFO:tensorflow:Step 32600 per-step time 0.807s\n",
      "I1020 03:13:04.020882 4682925568 model_lib_v2.py:705] Step 32600 per-step time 0.807s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03992872,\n",
      " 'Loss/localization_loss': 0.021695765,\n",
      " 'Loss/regularization_loss': 0.062521555,\n",
      " 'Loss/total_loss': 0.12414604,\n",
      " 'learning_rate': 0.022413949}\n",
      "I1020 03:13:04.023262 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03992872,\n",
      " 'Loss/localization_loss': 0.021695765,\n",
      " 'Loss/regularization_loss': 0.062521555,\n",
      " 'Loss/total_loss': 0.12414604,\n",
      " 'learning_rate': 0.022413949}\n",
      "INFO:tensorflow:Step 32700 per-step time 0.801s\n",
      "I1020 03:14:24.090247 4682925568 model_lib_v2.py:705] Step 32700 per-step time 0.801s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05230563,\n",
      " 'Loss/localization_loss': 0.018729271,\n",
      " 'Loss/regularization_loss': 0.062424965,\n",
      " 'Loss/total_loss': 0.13345987,\n",
      " 'learning_rate': 0.022183966}\n",
      "I1020 03:14:24.090483 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05230563,\n",
      " 'Loss/localization_loss': 0.018729271,\n",
      " 'Loss/regularization_loss': 0.062424965,\n",
      " 'Loss/total_loss': 0.13345987,\n",
      " 'learning_rate': 0.022183966}\n",
      "INFO:tensorflow:Step 32800 per-step time 0.829s\n",
      "I1020 03:15:47.018010 4682925568 model_lib_v2.py:705] Step 32800 per-step time 0.829s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06657288,\n",
      " 'Loss/localization_loss': 0.01595005,\n",
      " 'Loss/regularization_loss': 0.062328946,\n",
      " 'Loss/total_loss': 0.14485188,\n",
      " 'learning_rate': 0.021954723}\n",
      "I1020 03:15:47.020350 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06657288,\n",
      " 'Loss/localization_loss': 0.01595005,\n",
      " 'Loss/regularization_loss': 0.062328946,\n",
      " 'Loss/total_loss': 0.14485188,\n",
      " 'learning_rate': 0.021954723}\n",
      "INFO:tensorflow:Step 32900 per-step time 0.808s\n",
      "I1020 03:17:07.848464 4682925568 model_lib_v2.py:705] Step 32900 per-step time 0.808s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.060178567,\n",
      " 'Loss/localization_loss': 0.02945008,\n",
      " 'Loss/regularization_loss': 0.062238406,\n",
      " 'Loss/total_loss': 0.15186705,\n",
      " 'learning_rate': 0.02172622}\n",
      "I1020 03:17:07.848706 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.060178567,\n",
      " 'Loss/localization_loss': 0.02945008,\n",
      " 'Loss/regularization_loss': 0.062238406,\n",
      " 'Loss/total_loss': 0.15186705,\n",
      " 'learning_rate': 0.02172622}\n",
      "INFO:tensorflow:Step 33000 per-step time 0.800s\n",
      "I1020 03:18:27.808109 4682925568 model_lib_v2.py:705] Step 33000 per-step time 0.800s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0645288,\n",
      " 'Loss/localization_loss': 0.045657888,\n",
      " 'Loss/regularization_loss': 0.06214748,\n",
      " 'Loss/total_loss': 0.17233416,\n",
      " 'learning_rate': 0.021498466}\n",
      "I1020 03:18:27.808373 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0645288,\n",
      " 'Loss/localization_loss': 0.045657888,\n",
      " 'Loss/regularization_loss': 0.06214748,\n",
      " 'Loss/total_loss': 0.17233416,\n",
      " 'learning_rate': 0.021498466}\n",
      "INFO:tensorflow:Step 33100 per-step time 0.807s\n",
      "I1020 03:19:48.480287 4682925568 model_lib_v2.py:705] Step 33100 per-step time 0.807s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05030455,\n",
      " 'Loss/localization_loss': 0.015027403,\n",
      " 'Loss/regularization_loss': 0.062055975,\n",
      " 'Loss/total_loss': 0.12738793,\n",
      " 'learning_rate': 0.021271477}\n",
      "I1020 03:19:48.480629 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05030455,\n",
      " 'Loss/localization_loss': 0.015027403,\n",
      " 'Loss/regularization_loss': 0.062055975,\n",
      " 'Loss/total_loss': 0.12738793,\n",
      " 'learning_rate': 0.021271477}\n",
      "INFO:tensorflow:Step 33200 per-step time 0.804s\n",
      "I1020 03:21:08.868798 4682925568 model_lib_v2.py:705] Step 33200 per-step time 0.804s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06446735,\n",
      " 'Loss/localization_loss': 0.025897713,\n",
      " 'Loss/regularization_loss': 0.06196743,\n",
      " 'Loss/total_loss': 0.15233248,\n",
      " 'learning_rate': 0.021045249}\n",
      "I1020 03:21:08.869051 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06446735,\n",
      " 'Loss/localization_loss': 0.025897713,\n",
      " 'Loss/regularization_loss': 0.06196743,\n",
      " 'Loss/total_loss': 0.15233248,\n",
      " 'learning_rate': 0.021045249}\n",
      "INFO:tensorflow:Step 33300 per-step time 0.797s\n",
      "I1020 03:22:28.573402 4682925568 model_lib_v2.py:705] Step 33300 per-step time 0.797s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.02926679,\n",
      " 'Loss/localization_loss': 0.0087395655,\n",
      " 'Loss/regularization_loss': 0.061879348,\n",
      " 'Loss/total_loss': 0.0998857,\n",
      " 'learning_rate': 0.02081981}\n",
      "I1020 03:22:28.573645 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.02926679,\n",
      " 'Loss/localization_loss': 0.0087395655,\n",
      " 'Loss/regularization_loss': 0.061879348,\n",
      " 'Loss/total_loss': 0.0998857,\n",
      " 'learning_rate': 0.02081981}\n",
      "INFO:tensorflow:Step 33400 per-step time 0.799s\n",
      "I1020 03:23:48.456120 4682925568 model_lib_v2.py:705] Step 33400 per-step time 0.799s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.09190415,\n",
      " 'Loss/localization_loss': 0.036264516,\n",
      " 'Loss/regularization_loss': 0.061792135,\n",
      " 'Loss/total_loss': 0.18996081,\n",
      " 'learning_rate': 0.02059516}\n",
      "I1020 03:23:48.456437 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.09190415,\n",
      " 'Loss/localization_loss': 0.036264516,\n",
      " 'Loss/regularization_loss': 0.061792135,\n",
      " 'Loss/total_loss': 0.18996081,\n",
      " 'learning_rate': 0.02059516}\n",
      "INFO:tensorflow:Step 33500 per-step time 0.805s\n",
      "I1020 03:25:08.974546 4682925568 model_lib_v2.py:705] Step 33500 per-step time 0.805s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.020103674,\n",
      " 'Loss/localization_loss': 0.010474952,\n",
      " 'Loss/regularization_loss': 0.06170386,\n",
      " 'Loss/total_loss': 0.09228249,\n",
      " 'learning_rate': 0.020371294}\n",
      "I1020 03:25:08.974797 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.020103674,\n",
      " 'Loss/localization_loss': 0.010474952,\n",
      " 'Loss/regularization_loss': 0.06170386,\n",
      " 'Loss/total_loss': 0.09228249,\n",
      " 'learning_rate': 0.020371294}\n",
      "INFO:tensorflow:Step 33600 per-step time 0.800s\n",
      "I1020 03:26:28.967546 4682925568 model_lib_v2.py:705] Step 33600 per-step time 0.800s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.026198097,\n",
      " 'Loss/localization_loss': 0.007903629,\n",
      " 'Loss/regularization_loss': 0.061615348,\n",
      " 'Loss/total_loss': 0.09571707,\n",
      " 'learning_rate': 0.020148247}\n",
      "I1020 03:26:28.967789 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.026198097,\n",
      " 'Loss/localization_loss': 0.007903629,\n",
      " 'Loss/regularization_loss': 0.061615348,\n",
      " 'Loss/total_loss': 0.09571707,\n",
      " 'learning_rate': 0.020148247}\n",
      "INFO:tensorflow:Step 33700 per-step time 0.799s\n",
      "I1020 03:27:48.860381 4682925568 model_lib_v2.py:705] Step 33700 per-step time 0.799s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.069838405,\n",
      " 'Loss/localization_loss': 0.02022662,\n",
      " 'Loss/regularization_loss': 0.061532706,\n",
      " 'Loss/total_loss': 0.15159774,\n",
      " 'learning_rate': 0.019926008}\n",
      "I1020 03:27:48.860630 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.069838405,\n",
      " 'Loss/localization_loss': 0.02022662,\n",
      " 'Loss/regularization_loss': 0.061532706,\n",
      " 'Loss/total_loss': 0.15159774,\n",
      " 'learning_rate': 0.019926008}\n",
      "INFO:tensorflow:Step 33800 per-step time 0.797s\n",
      "I1020 03:29:08.561824 4682925568 model_lib_v2.py:705] Step 33800 per-step time 0.797s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0390867,\n",
      " 'Loss/localization_loss': 0.016508233,\n",
      " 'Loss/regularization_loss': 0.061446734,\n",
      " 'Loss/total_loss': 0.11704166,\n",
      " 'learning_rate': 0.019704605}\n",
      "I1020 03:29:08.562067 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0390867,\n",
      " 'Loss/localization_loss': 0.016508233,\n",
      " 'Loss/regularization_loss': 0.061446734,\n",
      " 'Loss/total_loss': 0.11704166,\n",
      " 'learning_rate': 0.019704605}\n",
      "INFO:tensorflow:Step 33900 per-step time 0.817s\n",
      "I1020 03:30:30.307008 4682925568 model_lib_v2.py:705] Step 33900 per-step time 0.817s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.035758283,\n",
      " 'Loss/localization_loss': 0.016379489,\n",
      " 'Loss/regularization_loss': 0.06136159,\n",
      " 'Loss/total_loss': 0.11349936,\n",
      " 'learning_rate': 0.01948403}\n",
      "I1020 03:30:30.309345 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.035758283,\n",
      " 'Loss/localization_loss': 0.016379489,\n",
      " 'Loss/regularization_loss': 0.06136159,\n",
      " 'Loss/total_loss': 0.11349936,\n",
      " 'learning_rate': 0.01948403}\n",
      "INFO:tensorflow:Step 34000 per-step time 0.802s\n",
      "I1020 03:31:50.455321 4682925568 model_lib_v2.py:705] Step 34000 per-step time 0.802s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.036596335,\n",
      " 'Loss/localization_loss': 0.010926103,\n",
      " 'Loss/regularization_loss': 0.061280224,\n",
      " 'Loss/total_loss': 0.10880266,\n",
      " 'learning_rate': 0.019264294}\n",
      "I1020 03:31:50.455546 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.036596335,\n",
      " 'Loss/localization_loss': 0.010926103,\n",
      " 'Loss/regularization_loss': 0.061280224,\n",
      " 'Loss/total_loss': 0.10880266,\n",
      " 'learning_rate': 0.019264294}\n",
      "INFO:tensorflow:Step 34100 per-step time 0.807s\n",
      "I1020 03:33:11.143482 4682925568 model_lib_v2.py:705] Step 34100 per-step time 0.807s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05288872,\n",
      " 'Loss/localization_loss': 0.023976626,\n",
      " 'Loss/regularization_loss': 0.0611965,\n",
      " 'Loss/total_loss': 0.13806185,\n",
      " 'learning_rate': 0.01904542}\n",
      "I1020 03:33:11.143784 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05288872,\n",
      " 'Loss/localization_loss': 0.023976626,\n",
      " 'Loss/regularization_loss': 0.0611965,\n",
      " 'Loss/total_loss': 0.13806185,\n",
      " 'learning_rate': 0.01904542}\n",
      "INFO:tensorflow:Step 34200 per-step time 0.801s\n",
      "I1020 03:34:31.221864 4682925568 model_lib_v2.py:705] Step 34200 per-step time 0.801s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.1473001,\n",
      " 'Loss/localization_loss': 0.05030299,\n",
      " 'Loss/regularization_loss': 0.06111636,\n",
      " 'Loss/total_loss': 0.25871944,\n",
      " 'learning_rate': 0.018827397}\n",
      "I1020 03:34:31.222100 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.1473001,\n",
      " 'Loss/localization_loss': 0.05030299,\n",
      " 'Loss/regularization_loss': 0.06111636,\n",
      " 'Loss/total_loss': 0.25871944,\n",
      " 'learning_rate': 0.018827397}\n",
      "INFO:tensorflow:Step 34300 per-step time 0.808s\n",
      "I1020 03:35:52.006890 4682925568 model_lib_v2.py:705] Step 34300 per-step time 0.808s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.060147993,\n",
      " 'Loss/localization_loss': 0.014941658,\n",
      " 'Loss/regularization_loss': 0.061041012,\n",
      " 'Loss/total_loss': 0.13613066,\n",
      " 'learning_rate': 0.018610252}\n",
      "I1020 03:35:52.009150 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.060147993,\n",
      " 'Loss/localization_loss': 0.014941658,\n",
      " 'Loss/regularization_loss': 0.061041012,\n",
      " 'Loss/total_loss': 0.13613066,\n",
      " 'learning_rate': 0.018610252}\n",
      "INFO:tensorflow:Step 34400 per-step time 0.802s\n",
      "I1020 03:37:12.238551 4682925568 model_lib_v2.py:705] Step 34400 per-step time 0.802s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06488586,\n",
      " 'Loss/localization_loss': 0.012372071,\n",
      " 'Loss/regularization_loss': 0.060961712,\n",
      " 'Loss/total_loss': 0.13821964,\n",
      " 'learning_rate': 0.018393978}\n",
      "I1020 03:37:12.238775 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06488586,\n",
      " 'Loss/localization_loss': 0.012372071,\n",
      " 'Loss/regularization_loss': 0.060961712,\n",
      " 'Loss/total_loss': 0.13821964,\n",
      " 'learning_rate': 0.018393978}\n",
      "INFO:tensorflow:Step 34500 per-step time 0.804s\n",
      "I1020 03:38:32.596871 4682925568 model_lib_v2.py:705] Step 34500 per-step time 0.804s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.043462496,\n",
      " 'Loss/localization_loss': 0.017303606,\n",
      " 'Loss/regularization_loss': 0.06088187,\n",
      " 'Loss/total_loss': 0.12164797,\n",
      " 'learning_rate': 0.018178603}\n",
      "I1020 03:38:32.597115 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.043462496,\n",
      " 'Loss/localization_loss': 0.017303606,\n",
      " 'Loss/regularization_loss': 0.06088187,\n",
      " 'Loss/total_loss': 0.12164797,\n",
      " 'learning_rate': 0.018178603}\n",
      "INFO:tensorflow:Step 34600 per-step time 0.813s\n",
      "I1020 03:39:53.880064 4682925568 model_lib_v2.py:705] Step 34600 per-step time 0.813s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07853478,\n",
      " 'Loss/localization_loss': 0.008555371,\n",
      " 'Loss/regularization_loss': 0.0608039,\n",
      " 'Loss/total_loss': 0.14789405,\n",
      " 'learning_rate': 0.017964115}\n",
      "I1020 03:39:53.882426 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07853478,\n",
      " 'Loss/localization_loss': 0.008555371,\n",
      " 'Loss/regularization_loss': 0.0608039,\n",
      " 'Loss/total_loss': 0.14789405,\n",
      " 'learning_rate': 0.017964115}\n",
      "INFO:tensorflow:Step 34700 per-step time 0.804s\n",
      "I1020 03:41:14.231909 4682925568 model_lib_v2.py:705] Step 34700 per-step time 0.804s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.060659483,\n",
      " 'Loss/localization_loss': 0.019549437,\n",
      " 'Loss/regularization_loss': 0.060728006,\n",
      " 'Loss/total_loss': 0.14093693,\n",
      " 'learning_rate': 0.017750543}\n",
      "I1020 03:41:14.232161 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.060659483,\n",
      " 'Loss/localization_loss': 0.019549437,\n",
      " 'Loss/regularization_loss': 0.060728006,\n",
      " 'Loss/total_loss': 0.14093693,\n",
      " 'learning_rate': 0.017750543}\n",
      "INFO:tensorflow:Step 34800 per-step time 0.805s\n",
      "I1020 03:42:34.779613 4682925568 model_lib_v2.py:705] Step 34800 per-step time 0.805s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.09821956,\n",
      " 'Loss/localization_loss': 0.022477275,\n",
      " 'Loss/regularization_loss': 0.06065462,\n",
      " 'Loss/total_loss': 0.18135145,\n",
      " 'learning_rate': 0.017537873}\n",
      "I1020 03:42:34.779992 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.09821956,\n",
      " 'Loss/localization_loss': 0.022477275,\n",
      " 'Loss/regularization_loss': 0.06065462,\n",
      " 'Loss/total_loss': 0.18135145,\n",
      " 'learning_rate': 0.017537873}\n",
      "INFO:tensorflow:Step 34900 per-step time 0.801s\n",
      "I1020 03:43:54.819833 4682925568 model_lib_v2.py:705] Step 34900 per-step time 0.801s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.060585454,\n",
      " 'Loss/localization_loss': 0.016126476,\n",
      " 'Loss/regularization_loss': 0.060580526,\n",
      " 'Loss/total_loss': 0.13729246,\n",
      " 'learning_rate': 0.017326137}\n",
      "I1020 03:43:54.820070 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.060585454,\n",
      " 'Loss/localization_loss': 0.016126476,\n",
      " 'Loss/regularization_loss': 0.060580526,\n",
      " 'Loss/total_loss': 0.13729246,\n",
      " 'learning_rate': 0.017326137}\n",
      "INFO:tensorflow:Step 35000 per-step time 0.819s\n",
      "I1020 03:45:16.765725 4682925568 model_lib_v2.py:705] Step 35000 per-step time 0.819s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03596425,\n",
      " 'Loss/localization_loss': 0.024725536,\n",
      " 'Loss/regularization_loss': 0.06050821,\n",
      " 'Loss/total_loss': 0.121198,\n",
      " 'learning_rate': 0.017115325}\n",
      "I1020 03:45:16.768102 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03596425,\n",
      " 'Loss/localization_loss': 0.024725536,\n",
      " 'Loss/regularization_loss': 0.06050821,\n",
      " 'Loss/total_loss': 0.121198,\n",
      " 'learning_rate': 0.017115325}\n",
      "INFO:tensorflow:Step 35100 per-step time 0.814s\n",
      "I1020 03:46:38.180396 4682925568 model_lib_v2.py:705] Step 35100 per-step time 0.814s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03691441,\n",
      " 'Loss/localization_loss': 0.010517274,\n",
      " 'Loss/regularization_loss': 0.060433447,\n",
      " 'Loss/total_loss': 0.107865125,\n",
      " 'learning_rate': 0.016905462}\n",
      "I1020 03:46:38.180668 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03691441,\n",
      " 'Loss/localization_loss': 0.010517274,\n",
      " 'Loss/regularization_loss': 0.060433447,\n",
      " 'Loss/total_loss': 0.107865125,\n",
      " 'learning_rate': 0.016905462}\n",
      "INFO:tensorflow:Step 35200 per-step time 0.806s\n",
      "I1020 03:47:58.762856 4682925568 model_lib_v2.py:705] Step 35200 per-step time 0.806s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.034957077,\n",
      " 'Loss/localization_loss': 0.007157755,\n",
      " 'Loss/regularization_loss': 0.060358964,\n",
      " 'Loss/total_loss': 0.102473795,\n",
      " 'learning_rate': 0.016696548}\n",
      "I1020 03:47:58.763097 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.034957077,\n",
      " 'Loss/localization_loss': 0.007157755,\n",
      " 'Loss/regularization_loss': 0.060358964,\n",
      " 'Loss/total_loss': 0.102473795,\n",
      " 'learning_rate': 0.016696548}\n",
      "INFO:tensorflow:Step 35300 per-step time 0.804s\n",
      "I1020 03:49:19.212543 4682925568 model_lib_v2.py:705] Step 35300 per-step time 0.804s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03003714,\n",
      " 'Loss/localization_loss': 0.008659544,\n",
      " 'Loss/regularization_loss': 0.060285352,\n",
      " 'Loss/total_loss': 0.098982036,\n",
      " 'learning_rate': 0.016488582}\n",
      "I1020 03:49:19.212882 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03003714,\n",
      " 'Loss/localization_loss': 0.008659544,\n",
      " 'Loss/regularization_loss': 0.060285352,\n",
      " 'Loss/total_loss': 0.098982036,\n",
      " 'learning_rate': 0.016488582}\n",
      "INFO:tensorflow:Step 35400 per-step time 0.805s\n",
      "I1020 03:50:39.733098 4682925568 model_lib_v2.py:705] Step 35400 per-step time 0.805s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05145911,\n",
      " 'Loss/localization_loss': 0.020797404,\n",
      " 'Loss/regularization_loss': 0.060214452,\n",
      " 'Loss/total_loss': 0.13247097,\n",
      " 'learning_rate': 0.016281595}\n",
      "I1020 03:50:39.733331 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05145911,\n",
      " 'Loss/localization_loss': 0.020797404,\n",
      " 'Loss/regularization_loss': 0.060214452,\n",
      " 'Loss/total_loss': 0.13247097,\n",
      " 'learning_rate': 0.016281595}\n",
      "INFO:tensorflow:Step 35500 per-step time 0.802s\n",
      "I1020 03:51:59.891791 4682925568 model_lib_v2.py:705] Step 35500 per-step time 0.802s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.037524547,\n",
      " 'Loss/localization_loss': 0.0077443495,\n",
      " 'Loss/regularization_loss': 0.060145937,\n",
      " 'Loss/total_loss': 0.10541484,\n",
      " 'learning_rate': 0.01607557}\n",
      "I1020 03:51:59.892036 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.037524547,\n",
      " 'Loss/localization_loss': 0.0077443495,\n",
      " 'Loss/regularization_loss': 0.060145937,\n",
      " 'Loss/total_loss': 0.10541484,\n",
      " 'learning_rate': 0.01607557}\n",
      "INFO:tensorflow:Step 35600 per-step time 0.833s\n",
      "I1020 03:53:23.209275 4682925568 model_lib_v2.py:705] Step 35600 per-step time 0.833s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04215276,\n",
      " 'Loss/localization_loss': 0.021780513,\n",
      " 'Loss/regularization_loss': 0.060078625,\n",
      " 'Loss/total_loss': 0.12401189,\n",
      " 'learning_rate': 0.01587054}\n",
      "I1020 03:53:23.211544 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04215276,\n",
      " 'Loss/localization_loss': 0.021780513,\n",
      " 'Loss/regularization_loss': 0.060078625,\n",
      " 'Loss/total_loss': 0.12401189,\n",
      " 'learning_rate': 0.01587054}\n",
      "INFO:tensorflow:Step 35700 per-step time 0.803s\n",
      "I1020 03:54:43.487100 4682925568 model_lib_v2.py:705] Step 35700 per-step time 0.803s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.056682236,\n",
      " 'Loss/localization_loss': 0.025624264,\n",
      " 'Loss/regularization_loss': 0.06001252,\n",
      " 'Loss/total_loss': 0.14231902,\n",
      " 'learning_rate': 0.015666502}\n",
      "I1020 03:54:43.487338 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.056682236,\n",
      " 'Loss/localization_loss': 0.025624264,\n",
      " 'Loss/regularization_loss': 0.06001252,\n",
      " 'Loss/total_loss': 0.14231902,\n",
      " 'learning_rate': 0.015666502}\n",
      "INFO:tensorflow:Step 35800 per-step time 0.807s\n",
      "I1020 03:56:04.157840 4682925568 model_lib_v2.py:705] Step 35800 per-step time 0.807s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08515011,\n",
      " 'Loss/localization_loss': 0.042058133,\n",
      " 'Loss/regularization_loss': 0.059947737,\n",
      " 'Loss/total_loss': 0.18715596,\n",
      " 'learning_rate': 0.015463455}\n",
      "I1020 03:56:04.158082 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08515011,\n",
      " 'Loss/localization_loss': 0.042058133,\n",
      " 'Loss/regularization_loss': 0.059947737,\n",
      " 'Loss/total_loss': 0.18715596,\n",
      " 'learning_rate': 0.015463455}\n",
      "INFO:tensorflow:Step 35900 per-step time 0.802s\n",
      "I1020 03:57:24.320719 4682925568 model_lib_v2.py:705] Step 35900 per-step time 0.802s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04627189,\n",
      " 'Loss/localization_loss': 0.019485079,\n",
      " 'Loss/regularization_loss': 0.05988108,\n",
      " 'Loss/total_loss': 0.12563805,\n",
      " 'learning_rate': 0.015261423}\n",
      "I1020 03:57:24.320956 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04627189,\n",
      " 'Loss/localization_loss': 0.019485079,\n",
      " 'Loss/regularization_loss': 0.05988108,\n",
      " 'Loss/total_loss': 0.12563805,\n",
      " 'learning_rate': 0.015261423}\n",
      "INFO:tensorflow:Step 36000 per-step time 0.803s\n",
      "I1020 03:58:44.577722 4682925568 model_lib_v2.py:705] Step 36000 per-step time 0.803s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.123528555,\n",
      " 'Loss/localization_loss': 0.04945781,\n",
      " 'Loss/regularization_loss': 0.0598154,\n",
      " 'Loss/total_loss': 0.23280177,\n",
      " 'learning_rate': 0.015060408}\n",
      "I1020 03:58:44.577965 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.123528555,\n",
      " 'Loss/localization_loss': 0.04945781,\n",
      " 'Loss/regularization_loss': 0.0598154,\n",
      " 'Loss/total_loss': 0.23280177,\n",
      " 'learning_rate': 0.015060408}\n",
      "INFO:tensorflow:Step 36100 per-step time 0.853s\n",
      "I1020 04:00:09.835462 4682925568 model_lib_v2.py:705] Step 36100 per-step time 0.853s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.062122587,\n",
      " 'Loss/localization_loss': 0.024731154,\n",
      " 'Loss/regularization_loss': 0.059752032,\n",
      " 'Loss/total_loss': 0.14660577,\n",
      " 'learning_rate': 0.014860413}\n",
      "I1020 04:00:09.835930 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.062122587,\n",
      " 'Loss/localization_loss': 0.024731154,\n",
      " 'Loss/regularization_loss': 0.059752032,\n",
      " 'Loss/total_loss': 0.14660577,\n",
      " 'learning_rate': 0.014860413}\n",
      "INFO:tensorflow:Step 36200 per-step time 0.863s\n",
      "I1020 04:01:36.137339 4682925568 model_lib_v2.py:705] Step 36200 per-step time 0.863s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.056300376,\n",
      " 'Loss/localization_loss': 0.019238643,\n",
      " 'Loss/regularization_loss': 0.059688855,\n",
      " 'Loss/total_loss': 0.13522787,\n",
      " 'learning_rate': 0.014661457}\n",
      "I1020 04:01:36.139455 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.056300376,\n",
      " 'Loss/localization_loss': 0.019238643,\n",
      " 'Loss/regularization_loss': 0.059688855,\n",
      " 'Loss/total_loss': 0.13522787,\n",
      " 'learning_rate': 0.014661457}\n",
      "INFO:tensorflow:Step 36300 per-step time 0.869s\n",
      "I1020 04:03:03.060184 4682925568 model_lib_v2.py:705] Step 36300 per-step time 0.869s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.080406025,\n",
      " 'Loss/localization_loss': 0.028356988,\n",
      " 'Loss/regularization_loss': 0.059626088,\n",
      " 'Loss/total_loss': 0.1683891,\n",
      " 'learning_rate': 0.014463536}\n",
      "I1020 04:03:03.062426 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.080406025,\n",
      " 'Loss/localization_loss': 0.028356988,\n",
      " 'Loss/regularization_loss': 0.059626088,\n",
      " 'Loss/total_loss': 0.1683891,\n",
      " 'learning_rate': 0.014463536}\n",
      "INFO:tensorflow:Step 36400 per-step time 0.836s\n",
      "I1020 04:04:26.619045 4682925568 model_lib_v2.py:705] Step 36400 per-step time 0.836s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.047691617,\n",
      " 'Loss/localization_loss': 0.008235925,\n",
      " 'Loss/regularization_loss': 0.059565686,\n",
      " 'Loss/total_loss': 0.11549322,\n",
      " 'learning_rate': 0.014266672}\n",
      "I1020 04:04:26.622534 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.047691617,\n",
      " 'Loss/localization_loss': 0.008235925,\n",
      " 'Loss/regularization_loss': 0.059565686,\n",
      " 'Loss/total_loss': 0.11549322,\n",
      " 'learning_rate': 0.014266672}\n",
      "INFO:tensorflow:Step 36500 per-step time 0.806s\n",
      "I1020 04:05:47.194360 4682925568 model_lib_v2.py:705] Step 36500 per-step time 0.806s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04580688,\n",
      " 'Loss/localization_loss': 0.020492509,\n",
      " 'Loss/regularization_loss': 0.059504565,\n",
      " 'Loss/total_loss': 0.12580396,\n",
      " 'learning_rate': 0.014070866}\n",
      "I1020 04:05:47.194616 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04580688,\n",
      " 'Loss/localization_loss': 0.020492509,\n",
      " 'Loss/regularization_loss': 0.059504565,\n",
      " 'Loss/total_loss': 0.12580396,\n",
      " 'learning_rate': 0.014070866}\n",
      "INFO:tensorflow:Step 36600 per-step time 0.816s\n",
      "I1020 04:07:08.841336 4682925568 model_lib_v2.py:705] Step 36600 per-step time 0.816s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0935515,\n",
      " 'Loss/localization_loss': 0.0153958965,\n",
      " 'Loss/regularization_loss': 0.059445594,\n",
      " 'Loss/total_loss': 0.16839299,\n",
      " 'learning_rate': 0.013876119}\n",
      "I1020 04:07:08.844424 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0935515,\n",
      " 'Loss/localization_loss': 0.0153958965,\n",
      " 'Loss/regularization_loss': 0.059445594,\n",
      " 'Loss/total_loss': 0.16839299,\n",
      " 'learning_rate': 0.013876119}\n",
      "INFO:tensorflow:Step 36700 per-step time 0.941s\n",
      "I1020 04:08:42.925768 4682925568 model_lib_v2.py:705] Step 36700 per-step time 0.941s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.024769682,\n",
      " 'Loss/localization_loss': 0.010252867,\n",
      " 'Loss/regularization_loss': 0.05938674,\n",
      " 'Loss/total_loss': 0.09440929,\n",
      " 'learning_rate': 0.013682451}\n",
      "I1020 04:08:42.929107 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.024769682,\n",
      " 'Loss/localization_loss': 0.010252867,\n",
      " 'Loss/regularization_loss': 0.05938674,\n",
      " 'Loss/total_loss': 0.09440929,\n",
      " 'learning_rate': 0.013682451}\n",
      "INFO:tensorflow:Step 36800 per-step time 0.823s\n",
      "I1020 04:10:05.266206 4682925568 model_lib_v2.py:705] Step 36800 per-step time 0.823s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0578624,\n",
      " 'Loss/localization_loss': 0.008406382,\n",
      " 'Loss/regularization_loss': 0.05932838,\n",
      " 'Loss/total_loss': 0.12559716,\n",
      " 'learning_rate': 0.013489859}\n",
      "I1020 04:10:05.267228 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0578624,\n",
      " 'Loss/localization_loss': 0.008406382,\n",
      " 'Loss/regularization_loss': 0.05932838,\n",
      " 'Loss/total_loss': 0.12559716,\n",
      " 'learning_rate': 0.013489859}\n",
      "INFO:tensorflow:Step 36900 per-step time 0.830s\n",
      "I1020 04:11:28.301770 4682925568 model_lib_v2.py:705] Step 36900 per-step time 0.830s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.054759767,\n",
      " 'Loss/localization_loss': 0.005081601,\n",
      " 'Loss/regularization_loss': 0.059269086,\n",
      " 'Loss/total_loss': 0.11911045,\n",
      " 'learning_rate': 0.013298363}\n",
      "I1020 04:11:28.303943 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.054759767,\n",
      " 'Loss/localization_loss': 0.005081601,\n",
      " 'Loss/regularization_loss': 0.059269086,\n",
      " 'Loss/total_loss': 0.11911045,\n",
      " 'learning_rate': 0.013298363}\n",
      "INFO:tensorflow:Step 37000 per-step time 0.807s\n",
      "I1020 04:12:49.030855 4682925568 model_lib_v2.py:705] Step 37000 per-step time 0.807s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.061648555,\n",
      " 'Loss/localization_loss': 0.028056009,\n",
      " 'Loss/regularization_loss': 0.059210777,\n",
      " 'Loss/total_loss': 0.14891535,\n",
      " 'learning_rate': 0.013107967}\n",
      "I1020 04:12:49.031090 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.061648555,\n",
      " 'Loss/localization_loss': 0.028056009,\n",
      " 'Loss/regularization_loss': 0.059210777,\n",
      " 'Loss/total_loss': 0.14891535,\n",
      " 'learning_rate': 0.013107967}\n",
      "INFO:tensorflow:Step 37100 per-step time 0.827s\n",
      "I1020 04:14:11.687606 4682925568 model_lib_v2.py:705] Step 37100 per-step time 0.827s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.045805067,\n",
      " 'Loss/localization_loss': 0.010034687,\n",
      " 'Loss/regularization_loss': 0.05915332,\n",
      " 'Loss/total_loss': 0.11499307,\n",
      " 'learning_rate': 0.012918668}\n",
      "I1020 04:14:11.687894 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.045805067,\n",
      " 'Loss/localization_loss': 0.010034687,\n",
      " 'Loss/regularization_loss': 0.05915332,\n",
      " 'Loss/total_loss': 0.11499307,\n",
      " 'learning_rate': 0.012918668}\n",
      "INFO:tensorflow:Step 37200 per-step time 0.808s\n",
      "I1020 04:15:32.458562 4682925568 model_lib_v2.py:705] Step 37200 per-step time 0.808s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.10362842,\n",
      " 'Loss/localization_loss': 0.044576596,\n",
      " 'Loss/regularization_loss': 0.059097826,\n",
      " 'Loss/total_loss': 0.20730284,\n",
      " 'learning_rate': 0.012730489}\n",
      "I1020 04:15:32.458795 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.10362842,\n",
      " 'Loss/localization_loss': 0.044576596,\n",
      " 'Loss/regularization_loss': 0.059097826,\n",
      " 'Loss/total_loss': 0.20730284,\n",
      " 'learning_rate': 0.012730489}\n",
      "INFO:tensorflow:Step 37300 per-step time 0.847s\n",
      "I1020 04:16:57.197021 4682925568 model_lib_v2.py:705] Step 37300 per-step time 0.847s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.09494657,\n",
      " 'Loss/localization_loss': 0.02245816,\n",
      " 'Loss/regularization_loss': 0.059043214,\n",
      " 'Loss/total_loss': 0.17644794,\n",
      " 'learning_rate': 0.012543423}\n",
      "I1020 04:16:57.197264 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.09494657,\n",
      " 'Loss/localization_loss': 0.02245816,\n",
      " 'Loss/regularization_loss': 0.059043214,\n",
      " 'Loss/total_loss': 0.17644794,\n",
      " 'learning_rate': 0.012543423}\n",
      "INFO:tensorflow:Step 37400 per-step time 0.831s\n",
      "I1020 04:18:20.324571 4682925568 model_lib_v2.py:705] Step 37400 per-step time 0.831s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0820338,\n",
      " 'Loss/localization_loss': 0.03192466,\n",
      " 'Loss/regularization_loss': 0.05899763,\n",
      " 'Loss/total_loss': 0.1729561,\n",
      " 'learning_rate': 0.012357495}\n",
      "I1020 04:18:20.326833 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0820338,\n",
      " 'Loss/localization_loss': 0.03192466,\n",
      " 'Loss/regularization_loss': 0.05899763,\n",
      " 'Loss/total_loss': 0.1729561,\n",
      " 'learning_rate': 0.012357495}\n",
      "INFO:tensorflow:Step 37500 per-step time 0.808s\n",
      "I1020 04:19:41.125142 4682925568 model_lib_v2.py:705] Step 37500 per-step time 0.808s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06954317,\n",
      " 'Loss/localization_loss': 0.02160727,\n",
      " 'Loss/regularization_loss': 0.05894724,\n",
      " 'Loss/total_loss': 0.15009768,\n",
      " 'learning_rate': 0.012172694}\n",
      "I1020 04:19:41.125434 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06954317,\n",
      " 'Loss/localization_loss': 0.02160727,\n",
      " 'Loss/regularization_loss': 0.05894724,\n",
      " 'Loss/total_loss': 0.15009768,\n",
      " 'learning_rate': 0.012172694}\n",
      "INFO:tensorflow:Step 37600 per-step time 0.803s\n",
      "I1020 04:21:01.459610 4682925568 model_lib_v2.py:705] Step 37600 per-step time 0.803s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.042809352,\n",
      " 'Loss/localization_loss': 0.02639372,\n",
      " 'Loss/regularization_loss': 0.05889501,\n",
      " 'Loss/total_loss': 0.12809809,\n",
      " 'learning_rate': 0.011989045}\n",
      "I1020 04:21:01.459838 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.042809352,\n",
      " 'Loss/localization_loss': 0.02639372,\n",
      " 'Loss/regularization_loss': 0.05889501,\n",
      " 'Loss/total_loss': 0.12809809,\n",
      " 'learning_rate': 0.011989045}\n",
      "INFO:tensorflow:Step 37700 per-step time 0.832s\n",
      "I1020 04:22:24.739212 4682925568 model_lib_v2.py:705] Step 37700 per-step time 0.832s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.036933545,\n",
      " 'Loss/localization_loss': 0.009521121,\n",
      " 'Loss/regularization_loss': 0.058844153,\n",
      " 'Loss/total_loss': 0.10529882,\n",
      " 'learning_rate': 0.01180654}\n",
      "I1020 04:22:24.741765 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.036933545,\n",
      " 'Loss/localization_loss': 0.009521121,\n",
      " 'Loss/regularization_loss': 0.058844153,\n",
      " 'Loss/total_loss': 0.10529882,\n",
      " 'learning_rate': 0.01180654}\n",
      "INFO:tensorflow:Step 37800 per-step time 0.808s\n",
      "I1020 04:23:45.537108 4682925568 model_lib_v2.py:705] Step 37800 per-step time 0.808s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.059151344,\n",
      " 'Loss/localization_loss': 0.031939186,\n",
      " 'Loss/regularization_loss': 0.05879668,\n",
      " 'Loss/total_loss': 0.1498872,\n",
      " 'learning_rate': 0.011625201}\n",
      "I1020 04:23:45.537374 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.059151344,\n",
      " 'Loss/localization_loss': 0.031939186,\n",
      " 'Loss/regularization_loss': 0.05879668,\n",
      " 'Loss/total_loss': 0.1498872,\n",
      " 'learning_rate': 0.011625201}\n",
      "INFO:tensorflow:Step 37900 per-step time 0.829s\n",
      "I1020 04:25:08.444267 4682925568 model_lib_v2.py:705] Step 37900 per-step time 0.829s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06654,\n",
      " 'Loss/localization_loss': 0.03987966,\n",
      " 'Loss/regularization_loss': 0.05874842,\n",
      " 'Loss/total_loss': 0.16516809,\n",
      " 'learning_rate': 0.011445029}\n",
      "I1020 04:25:08.446725 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06654,\n",
      " 'Loss/localization_loss': 0.03987966,\n",
      " 'Loss/regularization_loss': 0.05874842,\n",
      " 'Loss/total_loss': 0.16516809,\n",
      " 'learning_rate': 0.011445029}\n",
      "INFO:tensorflow:Step 38000 per-step time 0.817s\n",
      "I1020 04:26:30.168602 4682925568 model_lib_v2.py:705] Step 38000 per-step time 0.817s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.041181955,\n",
      " 'Loss/localization_loss': 0.010360465,\n",
      " 'Loss/regularization_loss': 0.05869918,\n",
      " 'Loss/total_loss': 0.1102416,\n",
      " 'learning_rate': 0.011266021}\n",
      "I1020 04:26:30.168843 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.041181955,\n",
      " 'Loss/localization_loss': 0.010360465,\n",
      " 'Loss/regularization_loss': 0.05869918,\n",
      " 'Loss/total_loss': 0.1102416,\n",
      " 'learning_rate': 0.011266021}\n",
      "INFO:tensorflow:Step 38100 per-step time 0.815s\n",
      "I1020 04:27:51.672950 4682925568 model_lib_v2.py:705] Step 38100 per-step time 0.815s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.050549276,\n",
      " 'Loss/localization_loss': 0.012980159,\n",
      " 'Loss/regularization_loss': 0.05865131,\n",
      " 'Loss/total_loss': 0.122180745,\n",
      " 'learning_rate': 0.011088205}\n",
      "I1020 04:27:51.673522 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.050549276,\n",
      " 'Loss/localization_loss': 0.012980159,\n",
      " 'Loss/regularization_loss': 0.05865131,\n",
      " 'Loss/total_loss': 0.122180745,\n",
      " 'learning_rate': 0.011088205}\n",
      "INFO:tensorflow:Step 38200 per-step time 0.804s\n",
      "I1020 04:29:12.033185 4682925568 model_lib_v2.py:705] Step 38200 per-step time 0.804s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.048306692,\n",
      " 'Loss/localization_loss': 0.015976805,\n",
      " 'Loss/regularization_loss': 0.058603644,\n",
      " 'Loss/total_loss': 0.12288714,\n",
      " 'learning_rate': 0.010911569}\n",
      "I1020 04:29:12.033414 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.048306692,\n",
      " 'Loss/localization_loss': 0.015976805,\n",
      " 'Loss/regularization_loss': 0.058603644,\n",
      " 'Loss/total_loss': 0.12288714,\n",
      " 'learning_rate': 0.010911569}\n",
      "INFO:tensorflow:Step 38300 per-step time 0.811s\n",
      "I1020 04:30:33.100009 4682925568 model_lib_v2.py:705] Step 38300 per-step time 0.811s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.054340426,\n",
      " 'Loss/localization_loss': 0.008805784,\n",
      " 'Loss/regularization_loss': 0.0585569,\n",
      " 'Loss/total_loss': 0.12170311,\n",
      " 'learning_rate': 0.010736136}\n",
      "I1020 04:30:33.100262 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.054340426,\n",
      " 'Loss/localization_loss': 0.008805784,\n",
      " 'Loss/regularization_loss': 0.0585569,\n",
      " 'Loss/total_loss': 0.12170311,\n",
      " 'learning_rate': 0.010736136}\n",
      "INFO:tensorflow:Step 38400 per-step time 0.801s\n",
      "I1020 04:31:53.218847 4682925568 model_lib_v2.py:705] Step 38400 per-step time 0.801s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.031367645,\n",
      " 'Loss/localization_loss': 0.006793474,\n",
      " 'Loss/regularization_loss': 0.058511052,\n",
      " 'Loss/total_loss': 0.09667217,\n",
      " 'learning_rate': 0.010561897}\n",
      "I1020 04:31:53.219079 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.031367645,\n",
      " 'Loss/localization_loss': 0.006793474,\n",
      " 'Loss/regularization_loss': 0.058511052,\n",
      " 'Loss/total_loss': 0.09667217,\n",
      " 'learning_rate': 0.010561897}\n",
      "INFO:tensorflow:Step 38500 per-step time 0.830s\n",
      "I1020 04:33:16.274234 4682925568 model_lib_v2.py:705] Step 38500 per-step time 0.830s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.051874045,\n",
      " 'Loss/localization_loss': 0.012668068,\n",
      " 'Loss/regularization_loss': 0.058465067,\n",
      " 'Loss/total_loss': 0.12300718,\n",
      " 'learning_rate': 0.010388877}\n",
      "I1020 04:33:16.276540 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.051874045,\n",
      " 'Loss/localization_loss': 0.012668068,\n",
      " 'Loss/regularization_loss': 0.058465067,\n",
      " 'Loss/total_loss': 0.12300718,\n",
      " 'learning_rate': 0.010388877}\n",
      "INFO:tensorflow:Step 38600 per-step time 0.810s\n",
      "I1020 04:34:37.259639 4682925568 model_lib_v2.py:705] Step 38600 per-step time 0.810s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.060641017,\n",
      " 'Loss/localization_loss': 0.0050294274,\n",
      " 'Loss/regularization_loss': 0.05842008,\n",
      " 'Loss/total_loss': 0.12409052,\n",
      " 'learning_rate': 0.010217068}\n",
      "I1020 04:34:37.259916 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.060641017,\n",
      " 'Loss/localization_loss': 0.0050294274,\n",
      " 'Loss/regularization_loss': 0.05842008,\n",
      " 'Loss/total_loss': 0.12409052,\n",
      " 'learning_rate': 0.010217068}\n",
      "INFO:tensorflow:Step 38700 per-step time 0.809s\n",
      "I1020 04:35:58.157962 4682925568 model_lib_v2.py:705] Step 38700 per-step time 0.809s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.028963571,\n",
      " 'Loss/localization_loss': 0.010599452,\n",
      " 'Loss/regularization_loss': 0.058378246,\n",
      " 'Loss/total_loss': 0.097941265,\n",
      " 'learning_rate': 0.010046489}\n",
      "I1020 04:35:58.158202 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.028963571,\n",
      " 'Loss/localization_loss': 0.010599452,\n",
      " 'Loss/regularization_loss': 0.058378246,\n",
      " 'Loss/total_loss': 0.097941265,\n",
      " 'learning_rate': 0.010046489}\n",
      "INFO:tensorflow:Step 38800 per-step time 0.826s\n",
      "I1020 04:37:20.801728 4682925568 model_lib_v2.py:705] Step 38800 per-step time 0.826s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.046873562,\n",
      " 'Loss/localization_loss': 0.012038165,\n",
      " 'Loss/regularization_loss': 0.058334548,\n",
      " 'Loss/total_loss': 0.11724627,\n",
      " 'learning_rate': 0.009877141}\n",
      "I1020 04:37:20.804037 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.046873562,\n",
      " 'Loss/localization_loss': 0.012038165,\n",
      " 'Loss/regularization_loss': 0.058334548,\n",
      " 'Loss/total_loss': 0.11724627,\n",
      " 'learning_rate': 0.009877141}\n",
      "INFO:tensorflow:Step 38900 per-step time 0.802s\n",
      "I1020 04:38:40.973743 4682925568 model_lib_v2.py:705] Step 38900 per-step time 0.802s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.053499956,\n",
      " 'Loss/localization_loss': 0.028951174,\n",
      " 'Loss/regularization_loss': 0.058291823,\n",
      " 'Loss/total_loss': 0.14074296,\n",
      " 'learning_rate': 0.009709024}\n",
      "I1020 04:38:40.973993 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.053499956,\n",
      " 'Loss/localization_loss': 0.028951174,\n",
      " 'Loss/regularization_loss': 0.058291823,\n",
      " 'Loss/total_loss': 0.14074296,\n",
      " 'learning_rate': 0.009709024}\n",
      "INFO:tensorflow:Step 39000 per-step time 0.824s\n",
      "I1020 04:40:03.425608 4682925568 model_lib_v2.py:705] Step 39000 per-step time 0.824s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.047980256,\n",
      " 'Loss/localization_loss': 0.02158466,\n",
      " 'Loss/regularization_loss': 0.058250207,\n",
      " 'Loss/total_loss': 0.12781513,\n",
      " 'learning_rate': 0.00954216}\n",
      "I1020 04:40:03.428021 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.047980256,\n",
      " 'Loss/localization_loss': 0.02158466,\n",
      " 'Loss/regularization_loss': 0.058250207,\n",
      " 'Loss/total_loss': 0.12781513,\n",
      " 'learning_rate': 0.00954216}\n",
      "INFO:tensorflow:Step 39100 per-step time 0.830s\n",
      "I1020 04:41:26.404904 4682925568 model_lib_v2.py:705] Step 39100 per-step time 0.830s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.044626743,\n",
      " 'Loss/localization_loss': 0.022720233,\n",
      " 'Loss/regularization_loss': 0.058209583,\n",
      " 'Loss/total_loss': 0.12555656,\n",
      " 'learning_rate': 0.009376547}\n",
      "I1020 04:41:26.407418 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.044626743,\n",
      " 'Loss/localization_loss': 0.022720233,\n",
      " 'Loss/regularization_loss': 0.058209583,\n",
      " 'Loss/total_loss': 0.12555656,\n",
      " 'learning_rate': 0.009376547}\n",
      "INFO:tensorflow:Step 39200 per-step time 0.806s\n",
      "I1020 04:42:46.974590 4682925568 model_lib_v2.py:705] Step 39200 per-step time 0.806s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.037540346,\n",
      " 'Loss/localization_loss': 0.016867572,\n",
      " 'Loss/regularization_loss': 0.058168724,\n",
      " 'Loss/total_loss': 0.11257664,\n",
      " 'learning_rate': 0.009212187}\n",
      "I1020 04:42:46.974821 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.037540346,\n",
      " 'Loss/localization_loss': 0.016867572,\n",
      " 'Loss/regularization_loss': 0.058168724,\n",
      " 'Loss/total_loss': 0.11257664,\n",
      " 'learning_rate': 0.009212187}\n",
      "INFO:tensorflow:Step 39300 per-step time 0.805s\n",
      "I1020 04:44:07.519351 4682925568 model_lib_v2.py:705] Step 39300 per-step time 0.805s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.027097289,\n",
      " 'Loss/localization_loss': 0.0071186135,\n",
      " 'Loss/regularization_loss': 0.05812891,\n",
      " 'Loss/total_loss': 0.092344806,\n",
      " 'learning_rate': 0.009049098}\n",
      "I1020 04:44:07.519660 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.027097289,\n",
      " 'Loss/localization_loss': 0.0071186135,\n",
      " 'Loss/regularization_loss': 0.05812891,\n",
      " 'Loss/total_loss': 0.092344806,\n",
      " 'learning_rate': 0.009049098}\n",
      "INFO:tensorflow:Step 39400 per-step time 0.800s\n",
      "I1020 04:45:27.514579 4682925568 model_lib_v2.py:705] Step 39400 per-step time 0.800s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.16727914,\n",
      " 'Loss/localization_loss': 0.06083078,\n",
      " 'Loss/regularization_loss': 0.058089737,\n",
      " 'Loss/total_loss': 0.28619966,\n",
      " 'learning_rate': 0.008887276}\n",
      "I1020 04:45:27.514816 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.16727914,\n",
      " 'Loss/localization_loss': 0.06083078,\n",
      " 'Loss/regularization_loss': 0.058089737,\n",
      " 'Loss/total_loss': 0.28619966,\n",
      " 'learning_rate': 0.008887276}\n",
      "INFO:tensorflow:Step 39500 per-step time 0.808s\n",
      "I1020 04:46:48.360638 4682925568 model_lib_v2.py:705] Step 39500 per-step time 0.808s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.051990096,\n",
      " 'Loss/localization_loss': 0.021486409,\n",
      " 'Loss/regularization_loss': 0.058050986,\n",
      " 'Loss/total_loss': 0.1315275,\n",
      " 'learning_rate': 0.008726739}\n",
      "I1020 04:46:48.360870 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.051990096,\n",
      " 'Loss/localization_loss': 0.021486409,\n",
      " 'Loss/regularization_loss': 0.058050986,\n",
      " 'Loss/total_loss': 0.1315275,\n",
      " 'learning_rate': 0.008726739}\n",
      "INFO:tensorflow:Step 39600 per-step time 0.812s\n",
      "I1020 04:48:09.555372 4682925568 model_lib_v2.py:705] Step 39600 per-step time 0.812s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04475114,\n",
      " 'Loss/localization_loss': 0.014689634,\n",
      " 'Loss/regularization_loss': 0.05801348,\n",
      " 'Loss/total_loss': 0.11745426,\n",
      " 'learning_rate': 0.008567488}\n",
      "I1020 04:48:09.564177 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04475114,\n",
      " 'Loss/localization_loss': 0.014689634,\n",
      " 'Loss/regularization_loss': 0.05801348,\n",
      " 'Loss/total_loss': 0.11745426,\n",
      " 'learning_rate': 0.008567488}\n",
      "INFO:tensorflow:Step 39700 per-step time 0.820s\n",
      "I1020 04:49:31.475349 4682925568 model_lib_v2.py:705] Step 39700 per-step time 0.820s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.059257235,\n",
      " 'Loss/localization_loss': 0.008723424,\n",
      " 'Loss/regularization_loss': 0.057976376,\n",
      " 'Loss/total_loss': 0.12595704,\n",
      " 'learning_rate': 0.008409522}\n",
      "I1020 04:49:31.477660 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.059257235,\n",
      " 'Loss/localization_loss': 0.008723424,\n",
      " 'Loss/regularization_loss': 0.057976376,\n",
      " 'Loss/total_loss': 0.12595704,\n",
      " 'learning_rate': 0.008409522}\n",
      "INFO:tensorflow:Step 39800 per-step time 0.813s\n",
      "I1020 04:50:52.733731 4682925568 model_lib_v2.py:705] Step 39800 per-step time 0.813s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0271725,\n",
      " 'Loss/localization_loss': 0.0051175207,\n",
      " 'Loss/regularization_loss': 0.057940453,\n",
      " 'Loss/total_loss': 0.09023047,\n",
      " 'learning_rate': 0.008252861}\n",
      "I1020 04:50:52.733973 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0271725,\n",
      " 'Loss/localization_loss': 0.0051175207,\n",
      " 'Loss/regularization_loss': 0.057940453,\n",
      " 'Loss/total_loss': 0.09023047,\n",
      " 'learning_rate': 0.008252861}\n",
      "INFO:tensorflow:Step 39900 per-step time 0.800s\n",
      "I1020 04:52:12.761137 4682925568 model_lib_v2.py:705] Step 39900 per-step time 0.800s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04138997,\n",
      " 'Loss/localization_loss': 0.011329823,\n",
      " 'Loss/regularization_loss': 0.057904664,\n",
      " 'Loss/total_loss': 0.110624455,\n",
      " 'learning_rate': 0.008097499}\n",
      "I1020 04:52:12.761374 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04138997,\n",
      " 'Loss/localization_loss': 0.011329823,\n",
      " 'Loss/regularization_loss': 0.057904664,\n",
      " 'Loss/total_loss': 0.110624455,\n",
      " 'learning_rate': 0.008097499}\n",
      "INFO:tensorflow:Step 40000 per-step time 0.819s\n",
      "I1020 04:53:34.679642 4682925568 model_lib_v2.py:705] Step 40000 per-step time 0.819s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04928576,\n",
      " 'Loss/localization_loss': 0.014009611,\n",
      " 'Loss/regularization_loss': 0.057869446,\n",
      " 'Loss/total_loss': 0.12116481,\n",
      " 'learning_rate': 0.007943453}\n",
      "I1020 04:53:34.681951 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04928576,\n",
      " 'Loss/localization_loss': 0.014009611,\n",
      " 'Loss/regularization_loss': 0.057869446,\n",
      " 'Loss/total_loss': 0.12116481,\n",
      " 'learning_rate': 0.007943453}\n",
      "INFO:tensorflow:Step 40100 per-step time 0.811s\n",
      "I1020 04:54:55.786973 4682925568 model_lib_v2.py:705] Step 40100 per-step time 0.811s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05357662,\n",
      " 'Loss/localization_loss': 0.01487352,\n",
      " 'Loss/regularization_loss': 0.05783523,\n",
      " 'Loss/total_loss': 0.12628537,\n",
      " 'learning_rate': 0.0077907275}\n",
      "I1020 04:54:55.787271 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05357662,\n",
      " 'Loss/localization_loss': 0.01487352,\n",
      " 'Loss/regularization_loss': 0.05783523,\n",
      " 'Loss/total_loss': 0.12628537,\n",
      " 'learning_rate': 0.0077907275}\n",
      "INFO:tensorflow:Step 40200 per-step time 0.800s\n",
      "I1020 04:56:15.767143 4682925568 model_lib_v2.py:705] Step 40200 per-step time 0.800s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06829498,\n",
      " 'Loss/localization_loss': 0.030060586,\n",
      " 'Loss/regularization_loss': 0.05780205,\n",
      " 'Loss/total_loss': 0.15615761,\n",
      " 'learning_rate': 0.0076393173}\n",
      "I1020 04:56:15.767515 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06829498,\n",
      " 'Loss/localization_loss': 0.030060586,\n",
      " 'Loss/regularization_loss': 0.05780205,\n",
      " 'Loss/total_loss': 0.15615761,\n",
      " 'learning_rate': 0.0076393173}\n",
      "INFO:tensorflow:Step 40300 per-step time 0.810s\n",
      "I1020 04:57:36.722471 4682925568 model_lib_v2.py:705] Step 40300 per-step time 0.810s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.041241027,\n",
      " 'Loss/localization_loss': 0.011848886,\n",
      " 'Loss/regularization_loss': 0.057769187,\n",
      " 'Loss/total_loss': 0.110859096,\n",
      " 'learning_rate': 0.007489245}\n",
      "I1020 04:57:36.722703 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.041241027,\n",
      " 'Loss/localization_loss': 0.011848886,\n",
      " 'Loss/regularization_loss': 0.057769187,\n",
      " 'Loss/total_loss': 0.110859096,\n",
      " 'learning_rate': 0.007489245}\n",
      "INFO:tensorflow:Step 40400 per-step time 0.802s\n",
      "I1020 04:58:56.905375 4682925568 model_lib_v2.py:705] Step 40400 per-step time 0.802s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08976288,\n",
      " 'Loss/localization_loss': 0.025596263,\n",
      " 'Loss/regularization_loss': 0.05773664,\n",
      " 'Loss/total_loss': 0.17309578,\n",
      " 'learning_rate': 0.0073405025}\n",
      "I1020 04:58:56.905685 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08976288,\n",
      " 'Loss/localization_loss': 0.025596263,\n",
      " 'Loss/regularization_loss': 0.05773664,\n",
      " 'Loss/total_loss': 0.17309578,\n",
      " 'learning_rate': 0.0073405025}\n",
      "INFO:tensorflow:Step 40500 per-step time 0.803s\n",
      "I1020 05:00:17.153748 4682925568 model_lib_v2.py:705] Step 40500 per-step time 0.803s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.057835016,\n",
      " 'Loss/localization_loss': 0.018410746,\n",
      " 'Loss/regularization_loss': 0.057705007,\n",
      " 'Loss/total_loss': 0.13395077,\n",
      " 'learning_rate': 0.00719311}\n",
      "I1020 05:00:17.154002 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.057835016,\n",
      " 'Loss/localization_loss': 0.018410746,\n",
      " 'Loss/regularization_loss': 0.057705007,\n",
      " 'Loss/total_loss': 0.13395077,\n",
      " 'learning_rate': 0.00719311}\n",
      "INFO:tensorflow:Step 40600 per-step time 0.822s\n",
      "I1020 05:01:39.360459 4682925568 model_lib_v2.py:705] Step 40600 per-step time 0.822s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.079785734,\n",
      " 'Loss/localization_loss': 0.030796515,\n",
      " 'Loss/regularization_loss': 0.057673465,\n",
      " 'Loss/total_loss': 0.16825572,\n",
      " 'learning_rate': 0.007047064}\n",
      "I1020 05:01:39.362842 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.079785734,\n",
      " 'Loss/localization_loss': 0.030796515,\n",
      " 'Loss/regularization_loss': 0.057673465,\n",
      " 'Loss/total_loss': 0.16825572,\n",
      " 'learning_rate': 0.007047064}\n",
      "INFO:tensorflow:Step 40700 per-step time 0.826s\n",
      "I1020 05:03:01.924034 4682925568 model_lib_v2.py:705] Step 40700 per-step time 0.826s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06097712,\n",
      " 'Loss/localization_loss': 0.02387021,\n",
      " 'Loss/regularization_loss': 0.057643056,\n",
      " 'Loss/total_loss': 0.14249039,\n",
      " 'learning_rate': 0.006902368}\n",
      "I1020 05:03:01.926270 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06097712,\n",
      " 'Loss/localization_loss': 0.02387021,\n",
      " 'Loss/regularization_loss': 0.057643056,\n",
      " 'Loss/total_loss': 0.14249039,\n",
      " 'learning_rate': 0.006902368}\n",
      "INFO:tensorflow:Step 40800 per-step time 0.811s\n",
      "I1020 05:04:22.964406 4682925568 model_lib_v2.py:705] Step 40800 per-step time 0.811s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.059084512,\n",
      " 'Loss/localization_loss': 0.020924317,\n",
      " 'Loss/regularization_loss': 0.057612784,\n",
      " 'Loss/total_loss': 0.13762161,\n",
      " 'learning_rate': 0.0067590377}\n",
      "I1020 05:04:22.964644 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.059084512,\n",
      " 'Loss/localization_loss': 0.020924317,\n",
      " 'Loss/regularization_loss': 0.057612784,\n",
      " 'Loss/total_loss': 0.13762161,\n",
      " 'learning_rate': 0.0067590377}\n",
      "INFO:tensorflow:Step 40900 per-step time 0.804s\n",
      "I1020 05:05:43.381533 4682925568 model_lib_v2.py:705] Step 40900 per-step time 0.804s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0291898,\n",
      " 'Loss/localization_loss': 0.0057738917,\n",
      " 'Loss/regularization_loss': 0.057583667,\n",
      " 'Loss/total_loss': 0.09254736,\n",
      " 'learning_rate': 0.006617067}\n",
      "I1020 05:05:43.381845 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0291898,\n",
      " 'Loss/localization_loss': 0.0057738917,\n",
      " 'Loss/regularization_loss': 0.057583667,\n",
      " 'Loss/total_loss': 0.09254736,\n",
      " 'learning_rate': 0.006617067}\n",
      "INFO:tensorflow:Step 41000 per-step time 0.804s\n",
      "I1020 05:07:03.769959 4682925568 model_lib_v2.py:705] Step 41000 per-step time 0.804s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05992132,\n",
      " 'Loss/localization_loss': 0.031294182,\n",
      " 'Loss/regularization_loss': 0.057554945,\n",
      " 'Loss/total_loss': 0.14877045,\n",
      " 'learning_rate': 0.006476476}\n",
      "I1020 05:07:03.770203 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05992132,\n",
      " 'Loss/localization_loss': 0.031294182,\n",
      " 'Loss/regularization_loss': 0.057554945,\n",
      " 'Loss/total_loss': 0.14877045,\n",
      " 'learning_rate': 0.006476476}\n",
      "INFO:tensorflow:Step 41100 per-step time 0.807s\n",
      "I1020 05:08:24.446061 4682925568 model_lib_v2.py:705] Step 41100 per-step time 0.807s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.058993373,\n",
      " 'Loss/localization_loss': 0.036385875,\n",
      " 'Loss/regularization_loss': 0.05752697,\n",
      " 'Loss/total_loss': 0.15290621,\n",
      " 'learning_rate': 0.006337256}\n",
      "I1020 05:08:24.446361 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.058993373,\n",
      " 'Loss/localization_loss': 0.036385875,\n",
      " 'Loss/regularization_loss': 0.05752697,\n",
      " 'Loss/total_loss': 0.15290621,\n",
      " 'learning_rate': 0.006337256}\n",
      "INFO:tensorflow:Step 41200 per-step time 0.806s\n",
      "I1020 05:09:45.093951 4682925568 model_lib_v2.py:705] Step 41200 per-step time 0.806s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06203599,\n",
      " 'Loss/localization_loss': 0.019566298,\n",
      " 'Loss/regularization_loss': 0.05749938,\n",
      " 'Loss/total_loss': 0.13910167,\n",
      " 'learning_rate': 0.0061994265}\n",
      "I1020 05:09:45.096235 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06203599,\n",
      " 'Loss/localization_loss': 0.019566298,\n",
      " 'Loss/regularization_loss': 0.05749938,\n",
      " 'Loss/total_loss': 0.13910167,\n",
      " 'learning_rate': 0.0061994265}\n",
      "INFO:tensorflow:Step 41300 per-step time 0.827s\n",
      "I1020 05:11:07.753333 4682925568 model_lib_v2.py:705] Step 41300 per-step time 0.827s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03890912,\n",
      " 'Loss/localization_loss': 0.00560919,\n",
      " 'Loss/regularization_loss': 0.057472672,\n",
      " 'Loss/total_loss': 0.10199098,\n",
      " 'learning_rate': 0.006062982}\n",
      "I1020 05:11:07.755642 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03890912,\n",
      " 'Loss/localization_loss': 0.00560919,\n",
      " 'Loss/regularization_loss': 0.057472672,\n",
      " 'Loss/total_loss': 0.10199098,\n",
      " 'learning_rate': 0.006062982}\n",
      "INFO:tensorflow:Step 41400 per-step time 0.832s\n",
      "I1020 05:12:30.945044 4682925568 model_lib_v2.py:705] Step 41400 per-step time 0.832s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05681845,\n",
      " 'Loss/localization_loss': 0.012528576,\n",
      " 'Loss/regularization_loss': 0.057446282,\n",
      " 'Loss/total_loss': 0.12679331,\n",
      " 'learning_rate': 0.0059279343}\n",
      "I1020 05:12:30.947561 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05681845,\n",
      " 'Loss/localization_loss': 0.012528576,\n",
      " 'Loss/regularization_loss': 0.057446282,\n",
      " 'Loss/total_loss': 0.12679331,\n",
      " 'learning_rate': 0.0059279343}\n",
      "INFO:tensorflow:Step 41500 per-step time 0.799s\n",
      "I1020 05:13:50.828871 4682925568 model_lib_v2.py:705] Step 41500 per-step time 0.799s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.043493178,\n",
      " 'Loss/localization_loss': 0.0064419326,\n",
      " 'Loss/regularization_loss': 0.057420403,\n",
      " 'Loss/total_loss': 0.10735551,\n",
      " 'learning_rate': 0.0057942844}\n",
      "I1020 05:13:50.829221 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.043493178,\n",
      " 'Loss/localization_loss': 0.0064419326,\n",
      " 'Loss/regularization_loss': 0.057420403,\n",
      " 'Loss/total_loss': 0.10735551,\n",
      " 'learning_rate': 0.0057942844}\n",
      "INFO:tensorflow:Step 41600 per-step time 0.822s\n",
      "I1020 05:15:13.024672 4682925568 model_lib_v2.py:705] Step 41600 per-step time 0.822s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.035015024,\n",
      " 'Loss/localization_loss': 0.017348554,\n",
      " 'Loss/regularization_loss': 0.057394873,\n",
      " 'Loss/total_loss': 0.10975845,\n",
      " 'learning_rate': 0.0056620454}\n",
      "I1020 05:15:13.027148 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.035015024,\n",
      " 'Loss/localization_loss': 0.017348554,\n",
      " 'Loss/regularization_loss': 0.057394873,\n",
      " 'Loss/total_loss': 0.10975845,\n",
      " 'learning_rate': 0.0056620454}\n",
      "INFO:tensorflow:Step 41700 per-step time 0.805s\n",
      "I1020 05:16:33.555135 4682925568 model_lib_v2.py:705] Step 41700 per-step time 0.805s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.036380626,\n",
      " 'Loss/localization_loss': 0.012236215,\n",
      " 'Loss/regularization_loss': 0.05737006,\n",
      " 'Loss/total_loss': 0.1059869,\n",
      " 'learning_rate': 0.0055312132}\n",
      "I1020 05:16:33.555371 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.036380626,\n",
      " 'Loss/localization_loss': 0.012236215,\n",
      " 'Loss/regularization_loss': 0.05737006,\n",
      " 'Loss/total_loss': 0.1059869,\n",
      " 'learning_rate': 0.0055312132}\n",
      "INFO:tensorflow:Step 41800 per-step time 0.800s\n",
      "I1020 05:17:53.595551 4682925568 model_lib_v2.py:705] Step 41800 per-step time 0.800s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03917807,\n",
      " 'Loss/localization_loss': 0.020454315,\n",
      " 'Loss/regularization_loss': 0.057345793,\n",
      " 'Loss/total_loss': 0.116978176,\n",
      " 'learning_rate': 0.005401802}\n",
      "I1020 05:17:53.595798 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03917807,\n",
      " 'Loss/localization_loss': 0.020454315,\n",
      " 'Loss/regularization_loss': 0.057345793,\n",
      " 'Loss/total_loss': 0.116978176,\n",
      " 'learning_rate': 0.005401802}\n",
      "INFO:tensorflow:Step 41900 per-step time 0.802s\n",
      "I1020 05:19:13.847576 4682925568 model_lib_v2.py:705] Step 41900 per-step time 0.802s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04749869,\n",
      " 'Loss/localization_loss': 0.012686944,\n",
      " 'Loss/regularization_loss': 0.05732286,\n",
      " 'Loss/total_loss': 0.11750849,\n",
      " 'learning_rate': 0.0052738143}\n",
      "I1020 05:19:13.847819 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04749869,\n",
      " 'Loss/localization_loss': 0.012686944,\n",
      " 'Loss/regularization_loss': 0.05732286,\n",
      " 'Loss/total_loss': 0.11750849,\n",
      " 'learning_rate': 0.0052738143}\n",
      "INFO:tensorflow:Step 42000 per-step time 0.796s\n",
      "I1020 05:20:33.434659 4682925568 model_lib_v2.py:705] Step 42000 per-step time 0.796s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.023129353,\n",
      " 'Loss/localization_loss': 0.01780617,\n",
      " 'Loss/regularization_loss': 0.05730004,\n",
      " 'Loss/total_loss': 0.09823556,\n",
      " 'learning_rate': 0.005147247}\n",
      "I1020 05:20:33.435066 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.023129353,\n",
      " 'Loss/localization_loss': 0.01780617,\n",
      " 'Loss/regularization_loss': 0.05730004,\n",
      " 'Loss/total_loss': 0.09823556,\n",
      " 'learning_rate': 0.005147247}\n",
      "INFO:tensorflow:Step 42100 per-step time 0.811s\n",
      "I1020 05:21:54.511628 4682925568 model_lib_v2.py:705] Step 42100 per-step time 0.811s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.043055315,\n",
      " 'Loss/localization_loss': 0.00567636,\n",
      " 'Loss/regularization_loss': 0.057277523,\n",
      " 'Loss/total_loss': 0.1060092,\n",
      " 'learning_rate': 0.00502212}\n",
      "I1020 05:21:54.511907 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.043055315,\n",
      " 'Loss/localization_loss': 0.00567636,\n",
      " 'Loss/regularization_loss': 0.057277523,\n",
      " 'Loss/total_loss': 0.1060092,\n",
      " 'learning_rate': 0.00502212}\n",
      "INFO:tensorflow:Step 42200 per-step time 0.801s\n",
      "I1020 05:23:14.659641 4682925568 model_lib_v2.py:705] Step 42200 per-step time 0.801s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.043791234,\n",
      " 'Loss/localization_loss': 0.012928801,\n",
      " 'Loss/regularization_loss': 0.057255495,\n",
      " 'Loss/total_loss': 0.113975525,\n",
      " 'learning_rate': 0.004898429}\n",
      "I1020 05:23:14.661834 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.043791234,\n",
      " 'Loss/localization_loss': 0.012928801,\n",
      " 'Loss/regularization_loss': 0.057255495,\n",
      " 'Loss/total_loss': 0.113975525,\n",
      " 'learning_rate': 0.004898429}\n",
      "INFO:tensorflow:Step 42300 per-step time 0.800s\n",
      "I1020 05:24:34.666423 4682925568 model_lib_v2.py:705] Step 42300 per-step time 0.800s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.113929495,\n",
      " 'Loss/localization_loss': 0.02895631,\n",
      " 'Loss/regularization_loss': 0.057234176,\n",
      " 'Loss/total_loss': 0.20011997,\n",
      " 'learning_rate': 0.0047761775}\n",
      "I1020 05:24:34.666666 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.113929495,\n",
      " 'Loss/localization_loss': 0.02895631,\n",
      " 'Loss/regularization_loss': 0.057234176,\n",
      " 'Loss/total_loss': 0.20011997,\n",
      " 'learning_rate': 0.0047761775}\n",
      "INFO:tensorflow:Step 42400 per-step time 0.801s\n",
      "I1020 05:25:54.749567 4682925568 model_lib_v2.py:705] Step 42400 per-step time 0.801s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.058203537,\n",
      " 'Loss/localization_loss': 0.022073058,\n",
      " 'Loss/regularization_loss': 0.057213165,\n",
      " 'Loss/total_loss': 0.13748977,\n",
      " 'learning_rate': 0.004655375}\n",
      "I1020 05:25:54.749808 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.058203537,\n",
      " 'Loss/localization_loss': 0.022073058,\n",
      " 'Loss/regularization_loss': 0.057213165,\n",
      " 'Loss/total_loss': 0.13748977,\n",
      " 'learning_rate': 0.004655375}\n",
      "INFO:tensorflow:Step 42500 per-step time 0.804s\n",
      "I1020 05:27:15.163780 4682925568 model_lib_v2.py:705] Step 42500 per-step time 0.804s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04366416,\n",
      " 'Loss/localization_loss': 0.0104909055,\n",
      " 'Loss/regularization_loss': 0.05719287,\n",
      " 'Loss/total_loss': 0.111347936,\n",
      " 'learning_rate': 0.004536025}\n",
      "I1020 05:27:15.165995 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04366416,\n",
      " 'Loss/localization_loss': 0.0104909055,\n",
      " 'Loss/regularization_loss': 0.05719287,\n",
      " 'Loss/total_loss': 0.111347936,\n",
      " 'learning_rate': 0.004536025}\n",
      "INFO:tensorflow:Step 42600 per-step time 0.794s\n",
      "I1020 05:28:34.542574 4682925568 model_lib_v2.py:705] Step 42600 per-step time 0.794s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07563172,\n",
      " 'Loss/localization_loss': 0.054777253,\n",
      " 'Loss/regularization_loss': 0.057173252,\n",
      " 'Loss/total_loss': 0.18758222,\n",
      " 'learning_rate': 0.0044181347}\n",
      "I1020 05:28:34.542809 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07563172,\n",
      " 'Loss/localization_loss': 0.054777253,\n",
      " 'Loss/regularization_loss': 0.057173252,\n",
      " 'Loss/total_loss': 0.18758222,\n",
      " 'learning_rate': 0.0044181347}\n",
      "INFO:tensorflow:Step 42700 per-step time 0.798s\n",
      "I1020 05:29:54.337221 4682925568 model_lib_v2.py:705] Step 42700 per-step time 0.798s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.23803079,\n",
      " 'Loss/localization_loss': 0.08214237,\n",
      " 'Loss/regularization_loss': 0.057153944,\n",
      " 'Loss/total_loss': 0.37732708,\n",
      " 'learning_rate': 0.0043017077}\n",
      "I1020 05:29:54.337444 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.23803079,\n",
      " 'Loss/localization_loss': 0.08214237,\n",
      " 'Loss/regularization_loss': 0.057153944,\n",
      " 'Loss/total_loss': 0.37732708,\n",
      " 'learning_rate': 0.0043017077}\n",
      "INFO:tensorflow:Step 42800 per-step time 0.802s\n",
      "I1020 05:31:14.524008 4682925568 model_lib_v2.py:705] Step 42800 per-step time 0.802s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.023307499,\n",
      " 'Loss/localization_loss': 0.0074360785,\n",
      " 'Loss/regularization_loss': 0.057135377,\n",
      " 'Loss/total_loss': 0.08787896,\n",
      " 'learning_rate': 0.004186745}\n",
      "I1020 05:31:14.524245 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.023307499,\n",
      " 'Loss/localization_loss': 0.0074360785,\n",
      " 'Loss/regularization_loss': 0.057135377,\n",
      " 'Loss/total_loss': 0.08787896,\n",
      " 'learning_rate': 0.004186745}\n",
      "INFO:tensorflow:Step 42900 per-step time 0.818s\n",
      "I1020 05:32:36.367594 4682925568 model_lib_v2.py:705] Step 42900 per-step time 0.818s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03669789,\n",
      " 'Loss/localization_loss': 0.008667463,\n",
      " 'Loss/regularization_loss': 0.057117134,\n",
      " 'Loss/total_loss': 0.10248249,\n",
      " 'learning_rate': 0.0040732576}\n",
      "I1020 05:32:36.369874 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03669789,\n",
      " 'Loss/localization_loss': 0.008667463,\n",
      " 'Loss/regularization_loss': 0.057117134,\n",
      " 'Loss/total_loss': 0.10248249,\n",
      " 'learning_rate': 0.0040732576}\n",
      "INFO:tensorflow:Step 43000 per-step time 0.794s\n",
      "I1020 05:33:55.744240 4682925568 model_lib_v2.py:705] Step 43000 per-step time 0.794s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.052107435,\n",
      " 'Loss/localization_loss': 0.006322851,\n",
      " 'Loss/regularization_loss': 0.05709924,\n",
      " 'Loss/total_loss': 0.11552952,\n",
      " 'learning_rate': 0.003961246}\n",
      "I1020 05:33:55.744467 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.052107435,\n",
      " 'Loss/localization_loss': 0.006322851,\n",
      " 'Loss/regularization_loss': 0.05709924,\n",
      " 'Loss/total_loss': 0.11552952,\n",
      " 'learning_rate': 0.003961246}\n",
      "INFO:tensorflow:Step 43100 per-step time 0.809s\n",
      "I1020 05:35:16.615940 4682925568 model_lib_v2.py:705] Step 43100 per-step time 0.809s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04860493,\n",
      " 'Loss/localization_loss': 0.011993486,\n",
      " 'Loss/regularization_loss': 0.057081956,\n",
      " 'Loss/total_loss': 0.11768037,\n",
      " 'learning_rate': 0.0038507176}\n",
      "I1020 05:35:16.616220 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04860493,\n",
      " 'Loss/localization_loss': 0.011993486,\n",
      " 'Loss/regularization_loss': 0.057081956,\n",
      " 'Loss/total_loss': 0.11768037,\n",
      " 'learning_rate': 0.0038507176}\n",
      "INFO:tensorflow:Step 43200 per-step time 0.800s\n",
      "I1020 05:36:36.595135 4682925568 model_lib_v2.py:705] Step 43200 per-step time 0.800s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03423225,\n",
      " 'Loss/localization_loss': 0.0121851135,\n",
      " 'Loss/regularization_loss': 0.05706502,\n",
      " 'Loss/total_loss': 0.10348238,\n",
      " 'learning_rate': 0.0037416671}\n",
      "I1020 05:36:36.595373 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03423225,\n",
      " 'Loss/localization_loss': 0.0121851135,\n",
      " 'Loss/regularization_loss': 0.05706502,\n",
      " 'Loss/total_loss': 0.10348238,\n",
      " 'learning_rate': 0.0037416671}\n",
      "INFO:tensorflow:Step 43300 per-step time 0.812s\n",
      "I1020 05:37:57.802393 4682925568 model_lib_v2.py:705] Step 43300 per-step time 0.812s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05998854,\n",
      " 'Loss/localization_loss': 0.017010484,\n",
      " 'Loss/regularization_loss': 0.05704863,\n",
      " 'Loss/total_loss': 0.13404766,\n",
      " 'learning_rate': 0.0036341143}\n",
      "I1020 05:37:57.804707 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05998854,\n",
      " 'Loss/localization_loss': 0.017010484,\n",
      " 'Loss/regularization_loss': 0.05704863,\n",
      " 'Loss/total_loss': 0.13404766,\n",
      " 'learning_rate': 0.0036341143}\n",
      "INFO:tensorflow:Step 43400 per-step time 0.802s\n",
      "I1020 05:39:18.004232 4682925568 model_lib_v2.py:705] Step 43400 per-step time 0.802s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05506752,\n",
      " 'Loss/localization_loss': 0.0049496703,\n",
      " 'Loss/regularization_loss': 0.057032622,\n",
      " 'Loss/total_loss': 0.11704981,\n",
      " 'learning_rate': 0.003528056}\n",
      "I1020 05:39:18.004468 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05506752,\n",
      " 'Loss/localization_loss': 0.0049496703,\n",
      " 'Loss/regularization_loss': 0.057032622,\n",
      " 'Loss/total_loss': 0.11704981,\n",
      " 'learning_rate': 0.003528056}\n",
      "INFO:tensorflow:Step 43500 per-step time 0.798s\n",
      "I1020 05:40:37.814541 4682925568 model_lib_v2.py:705] Step 43500 per-step time 0.798s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04120172,\n",
      " 'Loss/localization_loss': 0.017037205,\n",
      " 'Loss/regularization_loss': 0.057016976,\n",
      " 'Loss/total_loss': 0.1152559,\n",
      " 'learning_rate': 0.0034234975}\n",
      "I1020 05:40:37.814779 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04120172,\n",
      " 'Loss/localization_loss': 0.017037205,\n",
      " 'Loss/regularization_loss': 0.057016976,\n",
      " 'Loss/total_loss': 0.1152559,\n",
      " 'learning_rate': 0.0034234975}\n",
      "INFO:tensorflow:Step 43600 per-step time 0.798s\n",
      "I1020 05:41:57.643402 4682925568 model_lib_v2.py:705] Step 43600 per-step time 0.798s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03409201,\n",
      " 'Loss/localization_loss': 0.012611753,\n",
      " 'Loss/regularization_loss': 0.057001945,\n",
      " 'Loss/total_loss': 0.103705704,\n",
      " 'learning_rate': 0.0033204388}\n",
      "I1020 05:41:57.643646 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03409201,\n",
      " 'Loss/localization_loss': 0.012611753,\n",
      " 'Loss/regularization_loss': 0.057001945,\n",
      " 'Loss/total_loss': 0.103705704,\n",
      " 'learning_rate': 0.0033204388}\n",
      "INFO:tensorflow:Step 43700 per-step time 0.833s\n",
      "I1020 05:43:21.019342 4682925568 model_lib_v2.py:705] Step 43700 per-step time 0.833s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.061065365,\n",
      " 'Loss/localization_loss': 0.020519514,\n",
      " 'Loss/regularization_loss': 0.056987356,\n",
      " 'Loss/total_loss': 0.13857223,\n",
      " 'learning_rate': 0.0032188867}\n",
      "I1020 05:43:21.021638 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.061065365,\n",
      " 'Loss/localization_loss': 0.020519514,\n",
      " 'Loss/regularization_loss': 0.056987356,\n",
      " 'Loss/total_loss': 0.13857223,\n",
      " 'learning_rate': 0.0032188867}\n",
      "INFO:tensorflow:Step 43800 per-step time 0.796s\n",
      "I1020 05:44:40.560850 4682925568 model_lib_v2.py:705] Step 43800 per-step time 0.796s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07962263,\n",
      " 'Loss/localization_loss': 0.030596014,\n",
      " 'Loss/regularization_loss': 0.05697321,\n",
      " 'Loss/total_loss': 0.16719186,\n",
      " 'learning_rate': 0.0031188512}\n",
      "I1020 05:44:40.561100 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07962263,\n",
      " 'Loss/localization_loss': 0.030596014,\n",
      " 'Loss/regularization_loss': 0.05697321,\n",
      " 'Loss/total_loss': 0.16719186,\n",
      " 'learning_rate': 0.0031188512}\n",
      "INFO:tensorflow:Step 43900 per-step time 0.800s\n",
      "I1020 05:46:00.536298 4682925568 model_lib_v2.py:705] Step 43900 per-step time 0.800s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.044395737,\n",
      " 'Loss/localization_loss': 0.016193716,\n",
      " 'Loss/regularization_loss': 0.05695953,\n",
      " 'Loss/total_loss': 0.11754899,\n",
      " 'learning_rate': 0.0030203317}\n",
      "I1020 05:46:00.536523 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.044395737,\n",
      " 'Loss/localization_loss': 0.016193716,\n",
      " 'Loss/regularization_loss': 0.05695953,\n",
      " 'Loss/total_loss': 0.11754899,\n",
      " 'learning_rate': 0.0030203317}\n",
      "INFO:tensorflow:Step 44000 per-step time 0.812s\n",
      "I1020 05:47:21.789800 4682925568 model_lib_v2.py:705] Step 44000 per-step time 0.812s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03661752,\n",
      " 'Loss/localization_loss': 0.027804947,\n",
      " 'Loss/regularization_loss': 0.05694623,\n",
      " 'Loss/total_loss': 0.12136869,\n",
      " 'learning_rate': 0.002923329}\n",
      "I1020 05:47:21.792598 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03661752,\n",
      " 'Loss/localization_loss': 0.027804947,\n",
      " 'Loss/regularization_loss': 0.05694623,\n",
      " 'Loss/total_loss': 0.12136869,\n",
      " 'learning_rate': 0.002923329}\n",
      "INFO:tensorflow:Step 44100 per-step time 0.814s\n",
      "I1020 05:48:43.182041 4682925568 model_lib_v2.py:705] Step 44100 per-step time 0.814s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.054589808,\n",
      " 'Loss/localization_loss': 0.016041802,\n",
      " 'Loss/regularization_loss': 0.056933455,\n",
      " 'Loss/total_loss': 0.12756506,\n",
      " 'learning_rate': 0.0028278518}\n",
      "I1020 05:48:43.182336 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.054589808,\n",
      " 'Loss/localization_loss': 0.016041802,\n",
      " 'Loss/regularization_loss': 0.056933455,\n",
      " 'Loss/total_loss': 0.12756506,\n",
      " 'learning_rate': 0.0028278518}\n",
      "INFO:tensorflow:Step 44200 per-step time 0.799s\n",
      "I1020 05:50:03.079962 4682925568 model_lib_v2.py:705] Step 44200 per-step time 0.799s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06404644,\n",
      " 'Loss/localization_loss': 0.013637031,\n",
      " 'Loss/regularization_loss': 0.056921,\n",
      " 'Loss/total_loss': 0.13460447,\n",
      " 'learning_rate': 0.0027339004}\n",
      "I1020 05:50:03.080197 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06404644,\n",
      " 'Loss/localization_loss': 0.013637031,\n",
      " 'Loss/regularization_loss': 0.056921,\n",
      " 'Loss/total_loss': 0.13460447,\n",
      " 'learning_rate': 0.0027339004}\n",
      "INFO:tensorflow:Step 44300 per-step time 0.798s\n",
      "I1020 05:51:22.907590 4682925568 model_lib_v2.py:705] Step 44300 per-step time 0.798s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06769118,\n",
      " 'Loss/localization_loss': 0.008568467,\n",
      " 'Loss/regularization_loss': 0.05690897,\n",
      " 'Loss/total_loss': 0.13316861,\n",
      " 'learning_rate': 0.0026414872}\n",
      "I1020 05:51:22.907830 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06769118,\n",
      " 'Loss/localization_loss': 0.008568467,\n",
      " 'Loss/regularization_loss': 0.05690897,\n",
      " 'Loss/total_loss': 0.13316861,\n",
      " 'learning_rate': 0.0026414872}\n",
      "INFO:tensorflow:Step 44400 per-step time 0.798s\n",
      "I1020 05:52:42.714589 4682925568 model_lib_v2.py:705] Step 44400 per-step time 0.798s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.036760923,\n",
      " 'Loss/localization_loss': 0.011586,\n",
      " 'Loss/regularization_loss': 0.056897454,\n",
      " 'Loss/total_loss': 0.105244376,\n",
      " 'learning_rate': 0.002550602}\n",
      "I1020 05:52:42.718486 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.036760923,\n",
      " 'Loss/localization_loss': 0.011586,\n",
      " 'Loss/regularization_loss': 0.056897454,\n",
      " 'Loss/total_loss': 0.105244376,\n",
      " 'learning_rate': 0.002550602}\n",
      "INFO:tensorflow:Step 44500 per-step time 0.803s\n",
      "I1020 05:54:03.036186 4682925568 model_lib_v2.py:705] Step 44500 per-step time 0.803s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.045623116,\n",
      " 'Loss/localization_loss': 0.012529399,\n",
      " 'Loss/regularization_loss': 0.05688613,\n",
      " 'Loss/total_loss': 0.11503865,\n",
      " 'learning_rate': 0.0024612616}\n",
      "I1020 05:54:03.036420 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.045623116,\n",
      " 'Loss/localization_loss': 0.012529399,\n",
      " 'Loss/regularization_loss': 0.05688613,\n",
      " 'Loss/total_loss': 0.11503865,\n",
      " 'learning_rate': 0.0024612616}\n",
      "INFO:tensorflow:Step 44600 per-step time 0.797s\n",
      "I1020 05:55:22.698404 4682925568 model_lib_v2.py:705] Step 44600 per-step time 0.797s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.02958022,\n",
      " 'Loss/localization_loss': 0.008737608,\n",
      " 'Loss/regularization_loss': 0.056875337,\n",
      " 'Loss/total_loss': 0.09519316,\n",
      " 'learning_rate': 0.002373464}\n",
      "I1020 05:55:22.698650 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.02958022,\n",
      " 'Loss/localization_loss': 0.008737608,\n",
      " 'Loss/regularization_loss': 0.056875337,\n",
      " 'Loss/total_loss': 0.09519316,\n",
      " 'learning_rate': 0.002373464}\n",
      "INFO:tensorflow:Step 44700 per-step time 0.799s\n",
      "I1020 05:56:42.616895 4682925568 model_lib_v2.py:705] Step 44700 per-step time 0.799s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.026808402,\n",
      " 'Loss/localization_loss': 0.008809308,\n",
      " 'Loss/regularization_loss': 0.05686503,\n",
      " 'Loss/total_loss': 0.09248274,\n",
      " 'learning_rate': 0.0022872114}\n",
      "I1020 05:56:42.617135 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.026808402,\n",
      " 'Loss/localization_loss': 0.008809308,\n",
      " 'Loss/regularization_loss': 0.05686503,\n",
      " 'Loss/total_loss': 0.09248274,\n",
      " 'learning_rate': 0.0022872114}\n",
      "INFO:tensorflow:Step 44800 per-step time 0.811s\n",
      "I1020 05:58:03.703402 4682925568 model_lib_v2.py:705] Step 44800 per-step time 0.811s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.045493666,\n",
      " 'Loss/localization_loss': 0.018238235,\n",
      " 'Loss/regularization_loss': 0.056854956,\n",
      " 'Loss/total_loss': 0.12058686,\n",
      " 'learning_rate': 0.002202506}\n",
      "I1020 05:58:03.705578 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.045493666,\n",
      " 'Loss/localization_loss': 0.018238235,\n",
      " 'Loss/regularization_loss': 0.056854956,\n",
      " 'Loss/total_loss': 0.12058686,\n",
      " 'learning_rate': 0.002202506}\n",
      "INFO:tensorflow:Step 44900 per-step time 0.820s\n",
      "I1020 05:59:25.753039 4682925568 model_lib_v2.py:705] Step 44900 per-step time 0.820s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03837091,\n",
      " 'Loss/localization_loss': 0.011230504,\n",
      " 'Loss/regularization_loss': 0.056845184,\n",
      " 'Loss/total_loss': 0.106446594,\n",
      " 'learning_rate': 0.00211936}\n",
      "I1020 05:59:25.755183 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03837091,\n",
      " 'Loss/localization_loss': 0.011230504,\n",
      " 'Loss/regularization_loss': 0.056845184,\n",
      " 'Loss/total_loss': 0.106446594,\n",
      " 'learning_rate': 0.00211936}\n",
      "INFO:tensorflow:Step 45000 per-step time 0.797s\n",
      "I1020 06:00:45.481379 4682925568 model_lib_v2.py:705] Step 45000 per-step time 0.797s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.022743378,\n",
      " 'Loss/localization_loss': 0.0055296025,\n",
      " 'Loss/regularization_loss': 0.056835838,\n",
      " 'Loss/total_loss': 0.08510882,\n",
      " 'learning_rate': 0.0020377683}\n",
      "I1020 06:00:45.481620 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.022743378,\n",
      " 'Loss/localization_loss': 0.0055296025,\n",
      " 'Loss/regularization_loss': 0.056835838,\n",
      " 'Loss/total_loss': 0.08510882,\n",
      " 'learning_rate': 0.0020377683}\n",
      "INFO:tensorflow:Step 45100 per-step time 0.831s\n",
      "I1020 06:02:08.622539 4682925568 model_lib_v2.py:705] Step 45100 per-step time 0.831s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07376705,\n",
      " 'Loss/localization_loss': 0.025011117,\n",
      " 'Loss/regularization_loss': 0.05682688,\n",
      " 'Loss/total_loss': 0.15560505,\n",
      " 'learning_rate': 0.0019577406}\n",
      "I1020 06:02:08.625082 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07376705,\n",
      " 'Loss/localization_loss': 0.025011117,\n",
      " 'Loss/regularization_loss': 0.05682688,\n",
      " 'Loss/total_loss': 0.15560505,\n",
      " 'learning_rate': 0.0019577406}\n",
      "INFO:tensorflow:Step 45200 per-step time 0.830s\n",
      "I1020 06:03:31.578207 4682925568 model_lib_v2.py:705] Step 45200 per-step time 0.830s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.059128106,\n",
      " 'Loss/localization_loss': 0.010481422,\n",
      " 'Loss/regularization_loss': 0.056818217,\n",
      " 'Loss/total_loss': 0.12642774,\n",
      " 'learning_rate': 0.0018792676}\n",
      "I1020 06:03:31.580485 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.059128106,\n",
      " 'Loss/localization_loss': 0.010481422,\n",
      " 'Loss/regularization_loss': 0.056818217,\n",
      " 'Loss/total_loss': 0.12642774,\n",
      " 'learning_rate': 0.0018792676}\n",
      "INFO:tensorflow:Step 45300 per-step time 0.803s\n",
      "I1020 06:04:51.828897 4682925568 model_lib_v2.py:705] Step 45300 per-step time 0.803s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.080651596,\n",
      " 'Loss/localization_loss': 0.038925797,\n",
      " 'Loss/regularization_loss': 0.05680991,\n",
      " 'Loss/total_loss': 0.17638731,\n",
      " 'learning_rate': 0.0018023705}\n",
      "I1020 06:04:51.829293 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.080651596,\n",
      " 'Loss/localization_loss': 0.038925797,\n",
      " 'Loss/regularization_loss': 0.05680991,\n",
      " 'Loss/total_loss': 0.17638731,\n",
      " 'learning_rate': 0.0018023705}\n",
      "INFO:tensorflow:Step 45400 per-step time 0.799s\n",
      "I1020 06:06:11.709428 4682925568 model_lib_v2.py:705] Step 45400 per-step time 0.799s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04911701,\n",
      " 'Loss/localization_loss': 0.021458264,\n",
      " 'Loss/regularization_loss': 0.056802,\n",
      " 'Loss/total_loss': 0.12737727,\n",
      " 'learning_rate': 0.0017270398}\n",
      "I1020 06:06:11.709685 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04911701,\n",
      " 'Loss/localization_loss': 0.021458264,\n",
      " 'Loss/regularization_loss': 0.056802,\n",
      " 'Loss/total_loss': 0.12737727,\n",
      " 'learning_rate': 0.0017270398}\n",
      "INFO:tensorflow:Step 45500 per-step time 0.801s\n",
      "I1020 06:07:31.865652 4682925568 model_lib_v2.py:705] Step 45500 per-step time 0.801s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.043139677,\n",
      " 'Loss/localization_loss': 0.016177649,\n",
      " 'Loss/regularization_loss': 0.056794345,\n",
      " 'Loss/total_loss': 0.11611167,\n",
      " 'learning_rate': 0.001653285}\n",
      "I1020 06:07:31.865978 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.043139677,\n",
      " 'Loss/localization_loss': 0.016177649,\n",
      " 'Loss/regularization_loss': 0.056794345,\n",
      " 'Loss/total_loss': 0.11611167,\n",
      " 'learning_rate': 0.001653285}\n",
      "INFO:tensorflow:Step 45600 per-step time 0.796s\n",
      "I1020 06:08:51.413285 4682925568 model_lib_v2.py:705] Step 45600 per-step time 0.796s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.051175956,\n",
      " 'Loss/localization_loss': 0.012646785,\n",
      " 'Loss/regularization_loss': 0.056787033,\n",
      " 'Loss/total_loss': 0.120609775,\n",
      " 'learning_rate': 0.0015811061}\n",
      "I1020 06:08:51.413530 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.051175956,\n",
      " 'Loss/localization_loss': 0.012646785,\n",
      " 'Loss/regularization_loss': 0.056787033,\n",
      " 'Loss/total_loss': 0.120609775,\n",
      " 'learning_rate': 0.0015811061}\n",
      "INFO:tensorflow:Step 45700 per-step time 0.795s\n",
      "I1020 06:10:10.892131 4682925568 model_lib_v2.py:705] Step 45700 per-step time 0.795s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.062428232,\n",
      " 'Loss/localization_loss': 0.010804137,\n",
      " 'Loss/regularization_loss': 0.056780104,\n",
      " 'Loss/total_loss': 0.13001247,\n",
      " 'learning_rate': 0.0015105009}\n",
      "I1020 06:10:10.892442 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.062428232,\n",
      " 'Loss/localization_loss': 0.010804137,\n",
      " 'Loss/regularization_loss': 0.056780104,\n",
      " 'Loss/total_loss': 0.13001247,\n",
      " 'learning_rate': 0.0015105009}\n",
      "INFO:tensorflow:Step 45800 per-step time 0.807s\n",
      "I1020 06:11:31.600325 4682925568 model_lib_v2.py:705] Step 45800 per-step time 0.807s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.09862937,\n",
      " 'Loss/localization_loss': 0.081157334,\n",
      " 'Loss/regularization_loss': 0.056773443,\n",
      " 'Loss/total_loss': 0.23656015,\n",
      " 'learning_rate': 0.0014414835}\n",
      "I1020 06:11:31.600581 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.09862937,\n",
      " 'Loss/localization_loss': 0.081157334,\n",
      " 'Loss/regularization_loss': 0.056773443,\n",
      " 'Loss/total_loss': 0.23656015,\n",
      " 'learning_rate': 0.0014414835}\n",
      "INFO:tensorflow:Step 45900 per-step time 0.810s\n",
      "I1020 06:12:52.638089 4682925568 model_lib_v2.py:705] Step 45900 per-step time 0.810s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06668113,\n",
      " 'Loss/localization_loss': 0.02705064,\n",
      " 'Loss/regularization_loss': 0.05676709,\n",
      " 'Loss/total_loss': 0.15049887,\n",
      " 'learning_rate': 0.0013740516}\n",
      "I1020 06:12:52.640313 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06668113,\n",
      " 'Loss/localization_loss': 0.02705064,\n",
      " 'Loss/regularization_loss': 0.05676709,\n",
      " 'Loss/total_loss': 0.15049887,\n",
      " 'learning_rate': 0.0013740516}\n",
      "INFO:tensorflow:Step 46000 per-step time 0.804s\n",
      "I1020 06:14:13.004425 4682925568 model_lib_v2.py:705] Step 46000 per-step time 0.804s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.026283653,\n",
      " 'Loss/localization_loss': 0.009140053,\n",
      " 'Loss/regularization_loss': 0.05676126,\n",
      " 'Loss/total_loss': 0.09218497,\n",
      " 'learning_rate': 0.0013082051}\n",
      "I1020 06:14:13.004800 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.026283653,\n",
      " 'Loss/localization_loss': 0.009140053,\n",
      " 'Loss/regularization_loss': 0.05676126,\n",
      " 'Loss/total_loss': 0.09218497,\n",
      " 'learning_rate': 0.0013082051}\n",
      "INFO:tensorflow:Step 46100 per-step time 0.808s\n",
      "I1020 06:15:33.848209 4682925568 model_lib_v2.py:705] Step 46100 per-step time 0.808s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03749233,\n",
      " 'Loss/localization_loss': 0.012949886,\n",
      " 'Loss/regularization_loss': 0.056755546,\n",
      " 'Loss/total_loss': 0.10719776,\n",
      " 'learning_rate': 0.0012439513}\n",
      "I1020 06:15:33.848514 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03749233,\n",
      " 'Loss/localization_loss': 0.012949886,\n",
      " 'Loss/regularization_loss': 0.056755546,\n",
      " 'Loss/total_loss': 0.10719776,\n",
      " 'learning_rate': 0.0012439513}\n",
      "INFO:tensorflow:Step 46200 per-step time 0.811s\n",
      "I1020 06:16:54.959815 4682925568 model_lib_v2.py:705] Step 46200 per-step time 0.811s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.060739353,\n",
      " 'Loss/localization_loss': 0.014495632,\n",
      " 'Loss/regularization_loss': 0.05675008,\n",
      " 'Loss/total_loss': 0.13198507,\n",
      " 'learning_rate': 0.0011812877}\n",
      "I1020 06:16:54.963409 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.060739353,\n",
      " 'Loss/localization_loss': 0.014495632,\n",
      " 'Loss/regularization_loss': 0.05675008,\n",
      " 'Loss/total_loss': 0.13198507,\n",
      " 'learning_rate': 0.0011812877}\n",
      "INFO:tensorflow:Step 46300 per-step time 0.798s\n",
      "I1020 06:18:14.776715 4682925568 model_lib_v2.py:705] Step 46300 per-step time 0.798s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03746197,\n",
      " 'Loss/localization_loss': 0.004461292,\n",
      " 'Loss/regularization_loss': 0.05674488,\n",
      " 'Loss/total_loss': 0.09866814,\n",
      " 'learning_rate': 0.0011202192}\n",
      "I1020 06:18:14.776959 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03746197,\n",
      " 'Loss/localization_loss': 0.004461292,\n",
      " 'Loss/regularization_loss': 0.05674488,\n",
      " 'Loss/total_loss': 0.09866814,\n",
      " 'learning_rate': 0.0011202192}\n",
      "INFO:tensorflow:Step 46400 per-step time 0.804s\n",
      "I1020 06:19:35.134933 4682925568 model_lib_v2.py:705] Step 46400 per-step time 0.804s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06585603,\n",
      " 'Loss/localization_loss': 0.008375729,\n",
      " 'Loss/regularization_loss': 0.056739938,\n",
      " 'Loss/total_loss': 0.1309717,\n",
      " 'learning_rate': 0.0010607505}\n",
      "I1020 06:19:35.135204 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06585603,\n",
      " 'Loss/localization_loss': 0.008375729,\n",
      " 'Loss/regularization_loss': 0.056739938,\n",
      " 'Loss/total_loss': 0.1309717,\n",
      " 'learning_rate': 0.0010607505}\n",
      "INFO:tensorflow:Step 46500 per-step time 0.821s\n",
      "I1020 06:20:57.263475 4682925568 model_lib_v2.py:705] Step 46500 per-step time 0.821s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03659586,\n",
      " 'Loss/localization_loss': 0.023530189,\n",
      " 'Loss/regularization_loss': 0.05673523,\n",
      " 'Loss/total_loss': 0.11686128,\n",
      " 'learning_rate': 0.0010028839}\n",
      "I1020 06:20:57.266999 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03659586,\n",
      " 'Loss/localization_loss': 0.023530189,\n",
      " 'Loss/regularization_loss': 0.05673523,\n",
      " 'Loss/total_loss': 0.11686128,\n",
      " 'learning_rate': 0.0010028839}\n",
      "INFO:tensorflow:Step 46600 per-step time 0.819s\n",
      "I1020 06:22:19.188439 4682925568 model_lib_v2.py:705] Step 46600 per-step time 0.819s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08790064,\n",
      " 'Loss/localization_loss': 0.025956701,\n",
      " 'Loss/regularization_loss': 0.05673079,\n",
      " 'Loss/total_loss': 0.17058814,\n",
      " 'learning_rate': 0.0009466195}\n",
      "I1020 06:22:19.191709 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08790064,\n",
      " 'Loss/localization_loss': 0.025956701,\n",
      " 'Loss/regularization_loss': 0.05673079,\n",
      " 'Loss/total_loss': 0.17058814,\n",
      " 'learning_rate': 0.0009466195}\n",
      "INFO:tensorflow:Step 46700 per-step time 0.801s\n",
      "I1020 06:23:39.236774 4682925568 model_lib_v2.py:705] Step 46700 per-step time 0.801s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05570675,\n",
      " 'Loss/localization_loss': 0.011633711,\n",
      " 'Loss/regularization_loss': 0.05672662,\n",
      " 'Loss/total_loss': 0.12406708,\n",
      " 'learning_rate': 0.00089195726}\n",
      "I1020 06:23:39.237042 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05570675,\n",
      " 'Loss/localization_loss': 0.011633711,\n",
      " 'Loss/regularization_loss': 0.05672662,\n",
      " 'Loss/total_loss': 0.12406708,\n",
      " 'learning_rate': 0.00089195726}\n",
      "INFO:tensorflow:Step 46800 per-step time 0.825s\n",
      "I1020 06:25:01.715115 4682925568 model_lib_v2.py:705] Step 46800 per-step time 0.825s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.078762546,\n",
      " 'Loss/localization_loss': 0.022853069,\n",
      " 'Loss/regularization_loss': 0.0567227,\n",
      " 'Loss/total_loss': 0.15833831,\n",
      " 'learning_rate': 0.00083890435}\n",
      "I1020 06:25:01.717466 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.078762546,\n",
      " 'Loss/localization_loss': 0.022853069,\n",
      " 'Loss/regularization_loss': 0.0567227,\n",
      " 'Loss/total_loss': 0.15833831,\n",
      " 'learning_rate': 0.00083890435}\n",
      "INFO:tensorflow:Step 46900 per-step time 0.843s\n",
      "I1020 06:26:26.014150 4682925568 model_lib_v2.py:705] Step 46900 per-step time 0.843s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.056132056,\n",
      " 'Loss/localization_loss': 0.011567804,\n",
      " 'Loss/regularization_loss': 0.056718994,\n",
      " 'Loss/total_loss': 0.124418855,\n",
      " 'learning_rate': 0.00078746316}\n",
      "I1020 06:26:26.014392 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.056132056,\n",
      " 'Loss/localization_loss': 0.011567804,\n",
      " 'Loss/regularization_loss': 0.056718994,\n",
      " 'Loss/total_loss': 0.124418855,\n",
      " 'learning_rate': 0.00078746316}\n",
      "INFO:tensorflow:Step 47000 per-step time 0.831s\n",
      "I1020 06:27:49.137011 4682925568 model_lib_v2.py:705] Step 47000 per-step time 0.831s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.027707746,\n",
      " 'Loss/localization_loss': 0.015388867,\n",
      " 'Loss/regularization_loss': 0.05671555,\n",
      " 'Loss/total_loss': 0.099812165,\n",
      " 'learning_rate': 0.0007376337}\n",
      "I1020 06:27:49.139740 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.027707746,\n",
      " 'Loss/localization_loss': 0.015388867,\n",
      " 'Loss/regularization_loss': 0.05671555,\n",
      " 'Loss/total_loss': 0.099812165,\n",
      " 'learning_rate': 0.0007376337}\n",
      "INFO:tensorflow:Step 47100 per-step time 0.811s\n",
      "I1020 06:29:10.184581 4682925568 model_lib_v2.py:705] Step 47100 per-step time 0.811s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05286526,\n",
      " 'Loss/localization_loss': 0.0112437485,\n",
      " 'Loss/regularization_loss': 0.05671232,\n",
      " 'Loss/total_loss': 0.12082133,\n",
      " 'learning_rate': 0.0006894159}\n",
      "I1020 06:29:10.184877 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05286526,\n",
      " 'Loss/localization_loss': 0.0112437485,\n",
      " 'Loss/regularization_loss': 0.05671232,\n",
      " 'Loss/total_loss': 0.12082133,\n",
      " 'learning_rate': 0.0006894159}\n",
      "INFO:tensorflow:Step 47200 per-step time 0.802s\n",
      "I1020 06:30:30.404893 4682925568 model_lib_v2.py:705] Step 47200 per-step time 0.802s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.049793947,\n",
      " 'Loss/localization_loss': 0.008480165,\n",
      " 'Loss/regularization_loss': 0.056709286,\n",
      " 'Loss/total_loss': 0.114983395,\n",
      " 'learning_rate': 0.000642817}\n",
      "I1020 06:30:30.405119 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.049793947,\n",
      " 'Loss/localization_loss': 0.008480165,\n",
      " 'Loss/regularization_loss': 0.056709286,\n",
      " 'Loss/total_loss': 0.114983395,\n",
      " 'learning_rate': 0.000642817}\n",
      "INFO:tensorflow:Step 47300 per-step time 0.798s\n",
      "I1020 06:31:50.207031 4682925568 model_lib_v2.py:705] Step 47300 per-step time 0.798s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.051951572,\n",
      " 'Loss/localization_loss': 0.006118933,\n",
      " 'Loss/regularization_loss': 0.05670646,\n",
      " 'Loss/total_loss': 0.11477697,\n",
      " 'learning_rate': 0.0005978322}\n",
      "I1020 06:31:50.207259 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.051951572,\n",
      " 'Loss/localization_loss': 0.006118933,\n",
      " 'Loss/regularization_loss': 0.05670646,\n",
      " 'Loss/total_loss': 0.11477697,\n",
      " 'learning_rate': 0.0005978322}\n",
      "INFO:tensorflow:Step 47400 per-step time 0.800s\n",
      "I1020 06:33:10.230228 4682925568 model_lib_v2.py:705] Step 47400 per-step time 0.800s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.086393766,\n",
      " 'Loss/localization_loss': 0.021604186,\n",
      " 'Loss/regularization_loss': 0.05670383,\n",
      " 'Loss/total_loss': 0.16470179,\n",
      " 'learning_rate': 0.000554471}\n",
      "I1020 06:33:10.232452 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.086393766,\n",
      " 'Loss/localization_loss': 0.021604186,\n",
      " 'Loss/regularization_loss': 0.05670383,\n",
      " 'Loss/total_loss': 0.16470179,\n",
      " 'learning_rate': 0.000554471}\n",
      "INFO:tensorflow:Step 47500 per-step time 0.811s\n",
      "I1020 06:34:31.305664 4682925568 model_lib_v2.py:705] Step 47500 per-step time 0.811s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.047948148,\n",
      " 'Loss/localization_loss': 0.026254918,\n",
      " 'Loss/regularization_loss': 0.056701396,\n",
      " 'Loss/total_loss': 0.13090447,\n",
      " 'learning_rate': 0.0005127263}\n",
      "I1020 06:34:31.308016 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.047948148,\n",
      " 'Loss/localization_loss': 0.026254918,\n",
      " 'Loss/regularization_loss': 0.056701396,\n",
      " 'Loss/total_loss': 0.13090447,\n",
      " 'learning_rate': 0.0005127263}\n",
      "INFO:tensorflow:Step 47600 per-step time 0.820s\n",
      "I1020 06:35:53.259231 4682925568 model_lib_v2.py:705] Step 47600 per-step time 0.820s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06557706,\n",
      " 'Loss/localization_loss': 0.0105736265,\n",
      " 'Loss/regularization_loss': 0.05669914,\n",
      " 'Loss/total_loss': 0.13284983,\n",
      " 'learning_rate': 0.00047260997}\n",
      "I1020 06:35:53.259485 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06557706,\n",
      " 'Loss/localization_loss': 0.0105736265,\n",
      " 'Loss/regularization_loss': 0.05669914,\n",
      " 'Loss/total_loss': 0.13284983,\n",
      " 'learning_rate': 0.00047260997}\n",
      "INFO:tensorflow:Step 47700 per-step time 0.803s\n",
      "I1020 06:37:13.552062 4682925568 model_lib_v2.py:705] Step 47700 per-step time 0.803s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05534276,\n",
      " 'Loss/localization_loss': 0.016274463,\n",
      " 'Loss/regularization_loss': 0.056697074,\n",
      " 'Loss/total_loss': 0.1283143,\n",
      " 'learning_rate': 0.0004341173}\n",
      "I1020 06:37:13.552426 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05534276,\n",
      " 'Loss/localization_loss': 0.016274463,\n",
      " 'Loss/regularization_loss': 0.056697074,\n",
      " 'Loss/total_loss': 0.1283143,\n",
      " 'learning_rate': 0.0004341173}\n",
      "INFO:tensorflow:Step 47800 per-step time 0.806s\n",
      "I1020 06:38:34.143229 4682925568 model_lib_v2.py:705] Step 47800 per-step time 0.806s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.34494257,\n",
      " 'Loss/localization_loss': 0.034996193,\n",
      " 'Loss/regularization_loss': 0.05669518,\n",
      " 'Loss/total_loss': 0.43663394,\n",
      " 'learning_rate': 0.00039724825}\n",
      "I1020 06:38:34.145809 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.34494257,\n",
      " 'Loss/localization_loss': 0.034996193,\n",
      " 'Loss/regularization_loss': 0.05669518,\n",
      " 'Loss/total_loss': 0.43663394,\n",
      " 'learning_rate': 0.00039724825}\n",
      "INFO:tensorflow:Step 47900 per-step time 0.799s\n",
      "I1020 06:39:54.054517 4682925568 model_lib_v2.py:705] Step 47900 per-step time 0.799s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.037100453,\n",
      " 'Loss/localization_loss': 0.014182705,\n",
      " 'Loss/regularization_loss': 0.056693487,\n",
      " 'Loss/total_loss': 0.107976645,\n",
      " 'learning_rate': 0.00036200762}\n",
      "I1020 06:39:54.054744 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.037100453,\n",
      " 'Loss/localization_loss': 0.014182705,\n",
      " 'Loss/regularization_loss': 0.056693487,\n",
      " 'Loss/total_loss': 0.107976645,\n",
      " 'learning_rate': 0.00036200762}\n",
      "INFO:tensorflow:Step 48000 per-step time 0.876s\n",
      "I1020 06:41:21.679636 4682925568 model_lib_v2.py:705] Step 48000 per-step time 0.876s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06937852,\n",
      " 'Loss/localization_loss': 0.0066532725,\n",
      " 'Loss/regularization_loss': 0.056691907,\n",
      " 'Loss/total_loss': 0.13272369,\n",
      " 'learning_rate': 0.00032839776}\n",
      "I1020 06:41:21.681958 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06937852,\n",
      " 'Loss/localization_loss': 0.0066532725,\n",
      " 'Loss/regularization_loss': 0.056691907,\n",
      " 'Loss/total_loss': 0.13272369,\n",
      " 'learning_rate': 0.00032839776}\n",
      "INFO:tensorflow:Step 48100 per-step time 0.821s\n",
      "I1020 06:42:43.738913 4682925568 model_lib_v2.py:705] Step 48100 per-step time 0.821s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.050609495,\n",
      " 'Loss/localization_loss': 0.007100205,\n",
      " 'Loss/regularization_loss': 0.05669048,\n",
      " 'Loss/total_loss': 0.11440018,\n",
      " 'learning_rate': 0.00029641867}\n",
      "I1020 06:42:43.741151 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.050609495,\n",
      " 'Loss/localization_loss': 0.007100205,\n",
      " 'Loss/regularization_loss': 0.05669048,\n",
      " 'Loss/total_loss': 0.11440018,\n",
      " 'learning_rate': 0.00029641867}\n",
      "INFO:tensorflow:Step 48200 per-step time 0.803s\n",
      "I1020 06:44:04.039837 4682925568 model_lib_v2.py:705] Step 48200 per-step time 0.803s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0669341,\n",
      " 'Loss/localization_loss': 0.022464307,\n",
      " 'Loss/regularization_loss': 0.05668919,\n",
      " 'Loss/total_loss': 0.1460876,\n",
      " 'learning_rate': 0.00026607275}\n",
      "I1020 06:44:04.040080 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0669341,\n",
      " 'Loss/localization_loss': 0.022464307,\n",
      " 'Loss/regularization_loss': 0.05668919,\n",
      " 'Loss/total_loss': 0.1460876,\n",
      " 'learning_rate': 0.00026607275}\n",
      "INFO:tensorflow:Step 48300 per-step time 0.794s\n",
      "I1020 06:45:23.447510 4682925568 model_lib_v2.py:705] Step 48300 per-step time 0.794s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.034472603,\n",
      " 'Loss/localization_loss': 0.0051145675,\n",
      " 'Loss/regularization_loss': 0.05668804,\n",
      " 'Loss/total_loss': 0.09627521,\n",
      " 'learning_rate': 0.00023735761}\n",
      "I1020 06:45:23.447887 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.034472603,\n",
      " 'Loss/localization_loss': 0.0051145675,\n",
      " 'Loss/regularization_loss': 0.05668804,\n",
      " 'Loss/total_loss': 0.09627521,\n",
      " 'learning_rate': 0.00023735761}\n",
      "INFO:tensorflow:Step 48400 per-step time 0.807s\n",
      "I1020 06:46:44.089649 4682925568 model_lib_v2.py:705] Step 48400 per-step time 0.807s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.078533255,\n",
      " 'Loss/localization_loss': 0.012395059,\n",
      " 'Loss/regularization_loss': 0.056687016,\n",
      " 'Loss/total_loss': 0.14761533,\n",
      " 'learning_rate': 0.00021027803}\n",
      "I1020 06:46:44.089946 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.078533255,\n",
      " 'Loss/localization_loss': 0.012395059,\n",
      " 'Loss/regularization_loss': 0.056687016,\n",
      " 'Loss/total_loss': 0.14761533,\n",
      " 'learning_rate': 0.00021027803}\n",
      "INFO:tensorflow:Step 48500 per-step time 0.798s\n",
      "I1020 06:48:03.884425 4682925568 model_lib_v2.py:705] Step 48500 per-step time 0.798s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06635327,\n",
      " 'Loss/localization_loss': 0.031324856,\n",
      " 'Loss/regularization_loss': 0.056686115,\n",
      " 'Loss/total_loss': 0.15436424,\n",
      " 'learning_rate': 0.000184834}\n",
      "I1020 06:48:03.884667 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06635327,\n",
      " 'Loss/localization_loss': 0.031324856,\n",
      " 'Loss/regularization_loss': 0.056686115,\n",
      " 'Loss/total_loss': 0.15436424,\n",
      " 'learning_rate': 0.000184834}\n",
      "INFO:tensorflow:Step 48600 per-step time 0.808s\n",
      "I1020 06:49:24.679386 4682925568 model_lib_v2.py:705] Step 48600 per-step time 0.808s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05273812,\n",
      " 'Loss/localization_loss': 0.019797789,\n",
      " 'Loss/regularization_loss': 0.05668533,\n",
      " 'Loss/total_loss': 0.12922123,\n",
      " 'learning_rate': 0.0001610279}\n",
      "I1020 06:49:24.679616 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05273812,\n",
      " 'Loss/localization_loss': 0.019797789,\n",
      " 'Loss/regularization_loss': 0.05668533,\n",
      " 'Loss/total_loss': 0.12922123,\n",
      " 'learning_rate': 0.0001610279}\n",
      "INFO:tensorflow:Step 48700 per-step time 0.866s\n",
      "I1020 06:50:51.305266 4682925568 model_lib_v2.py:705] Step 48700 per-step time 0.866s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07162466,\n",
      " 'Loss/localization_loss': 0.024700232,\n",
      " 'Loss/regularization_loss': 0.056684636,\n",
      " 'Loss/total_loss': 0.15300953,\n",
      " 'learning_rate': 0.00013885974}\n",
      "I1020 06:50:51.307788 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07162466,\n",
      " 'Loss/localization_loss': 0.024700232,\n",
      " 'Loss/regularization_loss': 0.056684636,\n",
      " 'Loss/total_loss': 0.15300953,\n",
      " 'learning_rate': 0.00013885974}\n",
      "INFO:tensorflow:Step 48800 per-step time 0.941s\n",
      "I1020 06:52:25.480058 4682925568 model_lib_v2.py:705] Step 48800 per-step time 0.941s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07211733,\n",
      " 'Loss/localization_loss': 0.033169013,\n",
      " 'Loss/regularization_loss': 0.05668404,\n",
      " 'Loss/total_loss': 0.16197038,\n",
      " 'learning_rate': 0.00011832714}\n",
      "I1020 06:52:25.483877 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07211733,\n",
      " 'Loss/localization_loss': 0.033169013,\n",
      " 'Loss/regularization_loss': 0.05668404,\n",
      " 'Loss/total_loss': 0.16197038,\n",
      " 'learning_rate': 0.00011832714}\n",
      "INFO:tensorflow:Step 48900 per-step time 0.819s\n",
      "I1020 06:53:47.296773 4682925568 model_lib_v2.py:705] Step 48900 per-step time 0.819s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.051952343,\n",
      " 'Loss/localization_loss': 0.041447315,\n",
      " 'Loss/regularization_loss': 0.05668354,\n",
      " 'Loss/total_loss': 0.1500832,\n",
      " 'learning_rate': 9.943485e-05}\n",
      "I1020 06:53:47.297023 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.051952343,\n",
      " 'Loss/localization_loss': 0.041447315,\n",
      " 'Loss/regularization_loss': 0.05668354,\n",
      " 'Loss/total_loss': 0.1500832,\n",
      " 'learning_rate': 9.943485e-05}\n",
      "INFO:tensorflow:Step 49000 per-step time 0.795s\n",
      "I1020 06:55:06.834620 4682925568 model_lib_v2.py:705] Step 49000 per-step time 0.795s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06288141,\n",
      " 'Loss/localization_loss': 0.027035907,\n",
      " 'Loss/regularization_loss': 0.05668312,\n",
      " 'Loss/total_loss': 0.14660044,\n",
      " 'learning_rate': 8.218288e-05}\n",
      "I1020 06:55:06.834861 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06288141,\n",
      " 'Loss/localization_loss': 0.027035907,\n",
      " 'Loss/regularization_loss': 0.05668312,\n",
      " 'Loss/total_loss': 0.14660044,\n",
      " 'learning_rate': 8.218288e-05}\n",
      "INFO:tensorflow:Step 49100 per-step time 0.824s\n",
      "I1020 06:56:29.259979 4682925568 model_lib_v2.py:705] Step 49100 per-step time 0.824s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04890265,\n",
      " 'Loss/localization_loss': 0.019234337,\n",
      " 'Loss/regularization_loss': 0.056682777,\n",
      " 'Loss/total_loss': 0.12481977,\n",
      " 'learning_rate': 6.6573615e-05}\n",
      "I1020 06:56:29.260282 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04890265,\n",
      " 'Loss/localization_loss': 0.019234337,\n",
      " 'Loss/regularization_loss': 0.056682777,\n",
      " 'Loss/total_loss': 0.12481977,\n",
      " 'learning_rate': 6.6573615e-05}\n",
      "INFO:tensorflow:Step 49200 per-step time 0.804s\n",
      "I1020 06:57:49.631717 4682925568 model_lib_v2.py:705] Step 49200 per-step time 0.804s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03647953,\n",
      " 'Loss/localization_loss': 0.0073140278,\n",
      " 'Loss/regularization_loss': 0.056682497,\n",
      " 'Loss/total_loss': 0.10047606,\n",
      " 'learning_rate': 5.2604675e-05}\n",
      "I1020 06:57:49.631955 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03647953,\n",
      " 'Loss/localization_loss': 0.0073140278,\n",
      " 'Loss/regularization_loss': 0.056682497,\n",
      " 'Loss/total_loss': 0.10047606,\n",
      " 'learning_rate': 5.2604675e-05}\n",
      "INFO:tensorflow:Step 49300 per-step time 0.799s\n",
      "I1020 06:59:09.505582 4682925568 model_lib_v2.py:705] Step 49300 per-step time 0.799s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06544267,\n",
      " 'Loss/localization_loss': 0.010764626,\n",
      " 'Loss/regularization_loss': 0.05668228,\n",
      " 'Loss/total_loss': 0.13288957,\n",
      " 'learning_rate': 4.0278435e-05}\n",
      "I1020 06:59:09.505815 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06544267,\n",
      " 'Loss/localization_loss': 0.010764626,\n",
      " 'Loss/regularization_loss': 0.05668228,\n",
      " 'Loss/total_loss': 0.13288957,\n",
      " 'learning_rate': 4.0278435e-05}\n",
      "INFO:tensorflow:Step 49400 per-step time 0.804s\n",
      "I1020 07:00:29.873381 4682925568 model_lib_v2.py:705] Step 49400 per-step time 0.804s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.038018495,\n",
      " 'Loss/localization_loss': 0.008543463,\n",
      " 'Loss/regularization_loss': 0.05668213,\n",
      " 'Loss/total_loss': 0.10324408,\n",
      " 'learning_rate': 2.9592513e-05}\n",
      "I1020 07:00:29.873686 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.038018495,\n",
      " 'Loss/localization_loss': 0.008543463,\n",
      " 'Loss/regularization_loss': 0.05668213,\n",
      " 'Loss/total_loss': 0.10324408,\n",
      " 'learning_rate': 2.9592513e-05}\n",
      "INFO:tensorflow:Step 49500 per-step time 0.815s\n",
      "I1020 07:01:51.370055 4682925568 model_lib_v2.py:705] Step 49500 per-step time 0.815s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.030145627,\n",
      " 'Loss/localization_loss': 0.015620335,\n",
      " 'Loss/regularization_loss': 0.056682013,\n",
      " 'Loss/total_loss': 0.10244797,\n",
      " 'learning_rate': 2.055168e-05}\n",
      "I1020 07:01:51.372300 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.030145627,\n",
      " 'Loss/localization_loss': 0.015620335,\n",
      " 'Loss/regularization_loss': 0.056682013,\n",
      " 'Loss/total_loss': 0.10244797,\n",
      " 'learning_rate': 2.055168e-05}\n",
      "INFO:tensorflow:Step 49600 per-step time 0.801s\n",
      "I1020 07:03:11.434710 4682925568 model_lib_v2.py:705] Step 49600 per-step time 0.801s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.036089007,\n",
      " 'Loss/localization_loss': 0.026011191,\n",
      " 'Loss/regularization_loss': 0.056681942,\n",
      " 'Loss/total_loss': 0.11878214,\n",
      " 'learning_rate': 1.3153553e-05}\n",
      "I1020 07:03:11.434967 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.036089007,\n",
      " 'Loss/localization_loss': 0.026011191,\n",
      " 'Loss/regularization_loss': 0.056681942,\n",
      " 'Loss/total_loss': 0.11878214,\n",
      " 'learning_rate': 1.3153553e-05}\n",
      "INFO:tensorflow:Step 49700 per-step time 0.800s\n",
      "I1020 07:04:31.448045 4682925568 model_lib_v2.py:705] Step 49700 per-step time 0.800s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05439926,\n",
      " 'Loss/localization_loss': 0.004281751,\n",
      " 'Loss/regularization_loss': 0.056681897,\n",
      " 'Loss/total_loss': 0.11536291,\n",
      " 'learning_rate': 7.398128e-06}\n",
      "I1020 07:04:31.448385 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05439926,\n",
      " 'Loss/localization_loss': 0.004281751,\n",
      " 'Loss/regularization_loss': 0.056681897,\n",
      " 'Loss/total_loss': 0.11536291,\n",
      " 'learning_rate': 7.398128e-06}\n",
      "INFO:tensorflow:Step 49800 per-step time 0.874s\n",
      "I1020 07:05:58.899969 4682925568 model_lib_v2.py:705] Step 49800 per-step time 0.874s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.041151665,\n",
      " 'Loss/localization_loss': 0.0047094068,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.102542944,\n",
      " 'learning_rate': 3.2877922e-06}\n",
      "I1020 07:05:58.902972 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.041151665,\n",
      " 'Loss/localization_loss': 0.0047094068,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.102542944,\n",
      " 'learning_rate': 3.2877922e-06}\n",
      "INFO:tensorflow:Step 49900 per-step time 0.803s\n",
      "I1020 07:07:19.180706 4682925568 model_lib_v2.py:705] Step 49900 per-step time 0.803s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.060971707,\n",
      " 'Loss/localization_loss': 0.019691385,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13734496,\n",
      " 'learning_rate': 8.2254405e-07}\n",
      "I1020 07:07:19.180938 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.060971707,\n",
      " 'Loss/localization_loss': 0.019691385,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13734496,\n",
      " 'learning_rate': 8.2254405e-07}\n",
      "INFO:tensorflow:Step 50000 per-step time 0.795s\n",
      "I1020 07:08:38.691440 4682925568 model_lib_v2.py:705] Step 50000 per-step time 0.795s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.057653375,\n",
      " 'Loss/localization_loss': 0.021229373,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13556463,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:08:38.692378 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.057653375,\n",
      " 'Loss/localization_loss': 0.021229373,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13556463,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 50100 per-step time 0.827s\n",
      "I1020 07:10:01.423395 4682925568 model_lib_v2.py:705] Step 50100 per-step time 0.827s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04511562,\n",
      " 'Loss/localization_loss': 0.007383116,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10918061,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:10:01.428614 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04511562,\n",
      " 'Loss/localization_loss': 0.007383116,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10918061,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 50200 per-step time 0.863s\n",
      "I1020 07:11:27.672234 4682925568 model_lib_v2.py:705] Step 50200 per-step time 0.863s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07777595,\n",
      " 'Loss/localization_loss': 0.040861472,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.17531928,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:11:27.672538 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07777595,\n",
      " 'Loss/localization_loss': 0.040861472,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.17531928,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 50300 per-step time 0.918s\n",
      "I1020 07:12:59.557960 4682925568 model_lib_v2.py:705] Step 50300 per-step time 0.918s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.059986323,\n",
      " 'Loss/localization_loss': 0.015125636,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13179383,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:12:59.564450 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.059986323,\n",
      " 'Loss/localization_loss': 0.015125636,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13179383,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 50400 per-step time 0.983s\n",
      "I1020 07:14:37.819369 4682925568 model_lib_v2.py:705] Step 50400 per-step time 0.983s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.039382238,\n",
      " 'Loss/localization_loss': 0.0062595643,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10232367,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:14:37.821863 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.039382238,\n",
      " 'Loss/localization_loss': 0.0062595643,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10232367,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 50500 per-step time 0.852s\n",
      "I1020 07:16:03.031465 4682925568 model_lib_v2.py:705] Step 50500 per-step time 0.852s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05409721,\n",
      " 'Loss/localization_loss': 0.015883168,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12666225,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:16:03.031715 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05409721,\n",
      " 'Loss/localization_loss': 0.015883168,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12666225,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 50600 per-step time 0.818s\n",
      "I1020 07:17:24.883023 4682925568 model_lib_v2.py:705] Step 50600 per-step time 0.818s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.049708966,\n",
      " 'Loss/localization_loss': 0.029697929,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13608876,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:17:24.885779 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.049708966,\n",
      " 'Loss/localization_loss': 0.029697929,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13608876,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 50700 per-step time 0.860s\n",
      "I1020 07:18:50.905296 4682925568 model_lib_v2.py:705] Step 50700 per-step time 0.860s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06780029,\n",
      " 'Loss/localization_loss': 0.036780577,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.16126274,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:18:50.908247 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06780029,\n",
      " 'Loss/localization_loss': 0.036780577,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.16126274,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 50800 per-step time 0.816s\n",
      "I1020 07:20:12.530085 4682925568 model_lib_v2.py:705] Step 50800 per-step time 0.816s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04233968,\n",
      " 'Loss/localization_loss': 0.0066190152,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10564056,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:20:12.530326 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04233968,\n",
      " 'Loss/localization_loss': 0.0066190152,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10564056,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 50900 per-step time 0.831s\n",
      "I1020 07:21:35.690759 4682925568 model_lib_v2.py:705] Step 50900 per-step time 0.831s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04743109,\n",
      " 'Loss/localization_loss': 0.016612077,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.120725036,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:21:35.693110 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04743109,\n",
      " 'Loss/localization_loss': 0.016612077,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.120725036,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 51000 per-step time 0.811s\n",
      "I1020 07:22:56.776791 4682925568 model_lib_v2.py:705] Step 51000 per-step time 0.811s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.054561175,\n",
      " 'Loss/localization_loss': 0.0104362825,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12167933,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:22:56.777024 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.054561175,\n",
      " 'Loss/localization_loss': 0.0104362825,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12167933,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 51100 per-step time 0.887s\n",
      "I1020 07:24:25.481752 4682925568 model_lib_v2.py:705] Step 51100 per-step time 0.887s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0439245,\n",
      " 'Loss/localization_loss': 0.019400371,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12000674,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:24:25.482080 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0439245,\n",
      " 'Loss/localization_loss': 0.019400371,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12000674,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 51200 per-step time 0.810s\n",
      "I1020 07:25:46.523152 4682925568 model_lib_v2.py:705] Step 51200 per-step time 0.810s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.058440484,\n",
      " 'Loss/localization_loss': 0.021224448,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.1363468,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:25:46.523390 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.058440484,\n",
      " 'Loss/localization_loss': 0.021224448,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.1363468,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 51300 per-step time 0.819s\n",
      "I1020 07:27:08.449318 4682925568 model_lib_v2.py:705] Step 51300 per-step time 0.819s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05871721,\n",
      " 'Loss/localization_loss': 0.010663131,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12606221,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:27:08.449561 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05871721,\n",
      " 'Loss/localization_loss': 0.010663131,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12606221,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 51400 per-step time 0.844s\n",
      "I1020 07:28:32.847765 4682925568 model_lib_v2.py:705] Step 51400 per-step time 0.844s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.023975743,\n",
      " 'Loss/localization_loss': 0.0065309443,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.08718856,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:28:32.850082 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.023975743,\n",
      " 'Loss/localization_loss': 0.0065309443,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.08718856,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 51500 per-step time 0.855s\n",
      "I1020 07:29:58.373138 4682925568 model_lib_v2.py:705] Step 51500 per-step time 0.855s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.02695406,\n",
      " 'Loss/localization_loss': 0.0046905996,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.08832653,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:29:58.373375 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.02695406,\n",
      " 'Loss/localization_loss': 0.0046905996,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.08832653,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 51600 per-step time 0.853s\n",
      "I1020 07:31:23.703021 4682925568 model_lib_v2.py:705] Step 51600 per-step time 0.853s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.054561086,\n",
      " 'Loss/localization_loss': 0.008910297,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.120153256,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:31:23.706918 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.054561086,\n",
      " 'Loss/localization_loss': 0.008910297,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.120153256,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 51700 per-step time 0.813s\n",
      "I1020 07:32:45.001321 4682925568 model_lib_v2.py:705] Step 51700 per-step time 0.813s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.038732372,\n",
      " 'Loss/localization_loss': 0.019941188,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.11535543,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:32:45.001550 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.038732372,\n",
      " 'Loss/localization_loss': 0.019941188,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.11535543,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 51800 per-step time 0.864s\n",
      "I1020 07:34:11.435918 4682925568 model_lib_v2.py:705] Step 51800 per-step time 0.864s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03673529,\n",
      " 'Loss/localization_loss': 0.028353129,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12177029,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:34:11.436154 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03673529,\n",
      " 'Loss/localization_loss': 0.028353129,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12177029,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 51900 per-step time 0.845s\n",
      "I1020 07:35:35.923737 4682925568 model_lib_v2.py:705] Step 51900 per-step time 0.845s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.047631588,\n",
      " 'Loss/localization_loss': 0.007827407,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.112140864,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:35:35.927138 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.047631588,\n",
      " 'Loss/localization_loss': 0.007827407,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.112140864,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 52000 per-step time 0.803s\n",
      "I1020 07:36:56.231391 4682925568 model_lib_v2.py:705] Step 52000 per-step time 0.803s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.10400255,\n",
      " 'Loss/localization_loss': 0.06651413,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.22719854,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:36:56.231621 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.10400255,\n",
      " 'Loss/localization_loss': 0.06651413,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.22719854,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 52100 per-step time 0.817s\n",
      "I1020 07:38:17.985502 4682925568 model_lib_v2.py:705] Step 52100 per-step time 0.817s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.07303359,\n",
      " 'Loss/localization_loss': 0.01663448,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.14634994,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:38:17.985791 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.07303359,\n",
      " 'Loss/localization_loss': 0.01663448,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.14634994,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 52200 per-step time 0.823s\n",
      "I1020 07:39:40.283522 4682925568 model_lib_v2.py:705] Step 52200 per-step time 0.823s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03343585,\n",
      " 'Loss/localization_loss': 0.0077735684,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.097891286,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:39:40.283759 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03343585,\n",
      " 'Loss/localization_loss': 0.0077735684,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.097891286,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 52300 per-step time 0.811s\n",
      "I1020 07:41:01.417806 4682925568 model_lib_v2.py:705] Step 52300 per-step time 0.811s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.090953395,\n",
      " 'Loss/localization_loss': 0.034455683,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.18209095,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:41:01.420101 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.090953395,\n",
      " 'Loss/localization_loss': 0.034455683,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.18209095,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 52400 per-step time 0.822s\n",
      "I1020 07:42:23.593176 4682925568 model_lib_v2.py:705] Step 52400 per-step time 0.822s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.048762597,\n",
      " 'Loss/localization_loss': 0.016688442,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12213291,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:42:23.593583 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.048762597,\n",
      " 'Loss/localization_loss': 0.016688442,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12213291,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 52500 per-step time 0.817s\n",
      "I1020 07:43:45.245363 4682925568 model_lib_v2.py:705] Step 52500 per-step time 0.817s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08661093,\n",
      " 'Loss/localization_loss': 0.030596478,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.17388928,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:43:45.245590 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08661093,\n",
      " 'Loss/localization_loss': 0.030596478,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.17388928,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 52600 per-step time 0.820s\n",
      "I1020 07:45:07.283855 4682925568 model_lib_v2.py:705] Step 52600 per-step time 0.820s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.061537907,\n",
      " 'Loss/localization_loss': 0.02217562,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.1403954,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:45:07.286217 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.061537907,\n",
      " 'Loss/localization_loss': 0.02217562,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.1403954,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 52700 per-step time 0.818s\n",
      "I1020 07:46:29.105890 4682925568 model_lib_v2.py:705] Step 52700 per-step time 0.818s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.054037377,\n",
      " 'Loss/localization_loss': 0.034456883,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.14517613,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:46:29.106130 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.054037377,\n",
      " 'Loss/localization_loss': 0.034456883,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.14517613,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 52800 per-step time 0.841s\n",
      "I1020 07:47:53.182301 4682925568 model_lib_v2.py:705] Step 52800 per-step time 0.841s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04331818,\n",
      " 'Loss/localization_loss': 0.022202656,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12220271,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:47:53.184629 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04331818,\n",
      " 'Loss/localization_loss': 0.022202656,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12220271,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 52900 per-step time 0.829s\n",
      "I1020 07:49:16.072633 4682925568 model_lib_v2.py:705] Step 52900 per-step time 0.829s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.038365237,\n",
      " 'Loss/localization_loss': 0.009602719,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10464983,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:49:16.075072 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.038365237,\n",
      " 'Loss/localization_loss': 0.009602719,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10464983,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 53000 per-step time 0.811s\n",
      "I1020 07:50:37.207583 4682925568 model_lib_v2.py:705] Step 53000 per-step time 0.811s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06946428,\n",
      " 'Loss/localization_loss': 0.008482057,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.1346282,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:50:37.207821 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06946428,\n",
      " 'Loss/localization_loss': 0.008482057,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.1346282,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 53100 per-step time 0.812s\n",
      "I1020 07:51:58.440216 4682925568 model_lib_v2.py:705] Step 53100 per-step time 0.812s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.027922885,\n",
      " 'Loss/localization_loss': 0.010038053,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.09464281,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:51:58.440521 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.027922885,\n",
      " 'Loss/localization_loss': 0.010038053,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.09464281,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 53200 per-step time 0.810s\n",
      "I1020 07:53:19.480509 4682925568 model_lib_v2.py:705] Step 53200 per-step time 0.810s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.059707507,\n",
      " 'Loss/localization_loss': 0.0111883795,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12757775,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:53:19.480799 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.059707507,\n",
      " 'Loss/localization_loss': 0.0111883795,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12757775,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 53300 per-step time 0.815s\n",
      "I1020 07:54:41.036506 4682925568 model_lib_v2.py:705] Step 53300 per-step time 0.815s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03966365,\n",
      " 'Loss/localization_loss': 0.0053564883,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.101702005,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:54:41.036770 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03966365,\n",
      " 'Loss/localization_loss': 0.0053564883,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.101702005,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 53400 per-step time 0.831s\n",
      "I1020 07:56:04.140244 4682925568 model_lib_v2.py:705] Step 53400 per-step time 0.831s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.058408953,\n",
      " 'Loss/localization_loss': 0.012445043,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12753586,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:56:04.140570 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.058408953,\n",
      " 'Loss/localization_loss': 0.012445043,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12753586,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 53500 per-step time 0.830s\n",
      "I1020 07:57:27.101248 4682925568 model_lib_v2.py:705] Step 53500 per-step time 0.830s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.038437657,\n",
      " 'Loss/localization_loss': 0.035191078,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13031061,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:57:27.101733 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.038437657,\n",
      " 'Loss/localization_loss': 0.035191078,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13031061,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 53600 per-step time 0.806s\n",
      "I1020 07:58:47.683145 4682925568 model_lib_v2.py:705] Step 53600 per-step time 0.806s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.059237566,\n",
      " 'Loss/localization_loss': 0.09668663,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.21260607,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 07:58:47.683466 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.059237566,\n",
      " 'Loss/localization_loss': 0.09668663,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.21260607,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 53700 per-step time 0.803s\n",
      "I1020 08:00:07.949784 4682925568 model_lib_v2.py:705] Step 53700 per-step time 0.803s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.051903687,\n",
      " 'Loss/localization_loss': 0.009731872,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.11831743,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:00:07.950021 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.051903687,\n",
      " 'Loss/localization_loss': 0.009731872,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.11831743,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 53800 per-step time 0.831s\n",
      "I1020 08:01:31.024554 4682925568 model_lib_v2.py:705] Step 53800 per-step time 0.831s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.088599086,\n",
      " 'Loss/localization_loss': 0.048251208,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.19353217,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:01:31.026841 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.088599086,\n",
      " 'Loss/localization_loss': 0.048251208,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.19353217,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 53900 per-step time 0.814s\n",
      "I1020 08:02:52.429702 4682925568 model_lib_v2.py:705] Step 53900 per-step time 0.814s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.053409256,\n",
      " 'Loss/localization_loss': 0.01758481,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12767594,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:02:52.429932 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.053409256,\n",
      " 'Loss/localization_loss': 0.01758481,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12767594,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 54000 per-step time 0.812s\n",
      "I1020 08:04:13.614332 4682925568 model_lib_v2.py:705] Step 54000 per-step time 0.812s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.063053995,\n",
      " 'Loss/localization_loss': 0.04596608,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.16570196,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:04:13.614559 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.063053995,\n",
      " 'Loss/localization_loss': 0.04596608,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.16570196,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 54100 per-step time 0.831s\n",
      "I1020 08:05:36.705513 4682925568 model_lib_v2.py:705] Step 54100 per-step time 0.831s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.046691395,\n",
      " 'Loss/localization_loss': 0.013202006,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.11657527,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:05:36.707776 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.046691395,\n",
      " 'Loss/localization_loss': 0.013202006,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.11657527,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 54200 per-step time 0.891s\n",
      "I1020 08:07:05.790672 4682925568 model_lib_v2.py:705] Step 54200 per-step time 0.891s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.058876183,\n",
      " 'Loss/localization_loss': 0.016050098,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13160816,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:07:05.793420 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.058876183,\n",
      " 'Loss/localization_loss': 0.016050098,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13160816,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 54300 per-step time 0.811s\n",
      "I1020 08:08:26.901296 4682925568 model_lib_v2.py:705] Step 54300 per-step time 0.811s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04789476,\n",
      " 'Loss/localization_loss': 0.026714643,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13129127,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:08:26.901533 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04789476,\n",
      " 'Loss/localization_loss': 0.026714643,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13129127,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 54400 per-step time 0.818s\n",
      "I1020 08:09:48.660362 4682925568 model_lib_v2.py:705] Step 54400 per-step time 0.818s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04519768,\n",
      " 'Loss/localization_loss': 0.0064145755,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10829413,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:09:48.660597 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04519768,\n",
      " 'Loss/localization_loss': 0.0064145755,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10829413,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 54500 per-step time 0.832s\n",
      "I1020 08:11:11.865539 4682925568 model_lib_v2.py:705] Step 54500 per-step time 0.832s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.016653063,\n",
      " 'Loss/localization_loss': 0.0104465755,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.08378151,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:11:11.868308 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.016653063,\n",
      " 'Loss/localization_loss': 0.0104465755,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.08378151,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 54600 per-step time 0.822s\n",
      "I1020 08:12:34.067301 4682925568 model_lib_v2.py:705] Step 54600 per-step time 0.822s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.028531883,\n",
      " 'Loss/localization_loss': 0.005548114,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.09076187,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:12:34.067546 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.028531883,\n",
      " 'Loss/localization_loss': 0.005548114,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.09076187,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 54700 per-step time 0.817s\n",
      "I1020 08:13:55.734945 4682925568 model_lib_v2.py:705] Step 54700 per-step time 0.817s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.076287,\n",
      " 'Loss/localization_loss': 0.012813166,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.14578204,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:13:55.735173 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.076287,\n",
      " 'Loss/localization_loss': 0.012813166,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.14578204,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 54800 per-step time 0.857s\n",
      "I1020 08:15:21.430611 4682925568 model_lib_v2.py:705] Step 54800 per-step time 0.857s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.05242732,\n",
      " 'Loss/localization_loss': 0.016949447,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12605864,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:15:21.432867 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.05242732,\n",
      " 'Loss/localization_loss': 0.016949447,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12605864,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 54900 per-step time 0.850s\n",
      "I1020 08:16:46.398733 4682925568 model_lib_v2.py:705] Step 54900 per-step time 0.850s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06263504,\n",
      " 'Loss/localization_loss': 0.023940504,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.14325741,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:16:46.400911 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06263504,\n",
      " 'Loss/localization_loss': 0.023940504,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.14325741,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 55000 per-step time 0.809s\n",
      "I1020 08:18:07.321924 4682925568 model_lib_v2.py:705] Step 55000 per-step time 0.809s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.089692995,\n",
      " 'Loss/localization_loss': 0.03316585,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.17954072,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:18:07.322149 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.089692995,\n",
      " 'Loss/localization_loss': 0.03316585,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.17954072,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 55100 per-step time 0.816s\n",
      "I1020 08:19:28.963298 4682925568 model_lib_v2.py:705] Step 55100 per-step time 0.816s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.039793618,\n",
      " 'Loss/localization_loss': 0.012242932,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.108718425,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:19:28.963568 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.039793618,\n",
      " 'Loss/localization_loss': 0.012242932,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.108718425,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 55200 per-step time 0.971s\n",
      "I1020 08:21:06.077626 4682925568 model_lib_v2.py:705] Step 55200 per-step time 0.971s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.044973195,\n",
      " 'Loss/localization_loss': 0.013226933,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.114882,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:21:06.079935 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.044973195,\n",
      " 'Loss/localization_loss': 0.013226933,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.114882,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 55300 per-step time 0.932s\n",
      "I1020 08:22:39.250993 4682925568 model_lib_v2.py:705] Step 55300 per-step time 0.932s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04886718,\n",
      " 'Loss/localization_loss': 0.015820477,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.121369526,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:22:39.251243 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04886718,\n",
      " 'Loss/localization_loss': 0.015820477,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.121369526,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 55400 per-step time 0.895s\n",
      "I1020 08:24:08.770181 4682925568 model_lib_v2.py:705] Step 55400 per-step time 0.895s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08498106,\n",
      " 'Loss/localization_loss': 0.03739295,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.17905588,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:24:08.772421 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08498106,\n",
      " 'Loss/localization_loss': 0.03739295,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.17905588,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 55500 per-step time 0.900s\n",
      "I1020 08:25:38.738903 4682925568 model_lib_v2.py:705] Step 55500 per-step time 0.900s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.058656067,\n",
      " 'Loss/localization_loss': 0.024853442,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.14019138,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:25:38.742742 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.058656067,\n",
      " 'Loss/localization_loss': 0.024853442,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.14019138,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 55600 per-step time 0.874s\n",
      "I1020 08:27:06.080435 4682925568 model_lib_v2.py:705] Step 55600 per-step time 0.874s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.055344258,\n",
      " 'Loss/localization_loss': 0.035652354,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.14767848,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:27:06.080671 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.055344258,\n",
      " 'Loss/localization_loss': 0.035652354,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.14767848,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 55700 per-step time 0.883s\n",
      "I1020 08:28:34.418465 4682925568 model_lib_v2.py:705] Step 55700 per-step time 0.883s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06694124,\n",
      " 'Loss/localization_loss': 0.053421035,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.17704415,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:28:34.420843 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06694124,\n",
      " 'Loss/localization_loss': 0.053421035,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.17704415,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 55800 per-step time 0.860s\n",
      "I1020 08:30:00.365983 4682925568 model_lib_v2.py:705] Step 55800 per-step time 0.860s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04466917,\n",
      " 'Loss/localization_loss': 0.005200136,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10655118,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:30:00.366299 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04466917,\n",
      " 'Loss/localization_loss': 0.005200136,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10655118,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 55900 per-step time 0.860s\n",
      "I1020 08:31:26.358988 4682925568 model_lib_v2.py:705] Step 55900 per-step time 0.860s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.073596776,\n",
      " 'Loss/localization_loss': 0.019307395,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.14958605,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:31:26.359240 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.073596776,\n",
      " 'Loss/localization_loss': 0.019307395,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.14958605,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 56000 per-step time 0.850s\n",
      "I1020 08:32:51.325963 4682925568 model_lib_v2.py:705] Step 56000 per-step time 0.850s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0727885,\n",
      " 'Loss/localization_loss': 0.009880091,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13935046,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:32:51.326842 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0727885,\n",
      " 'Loss/localization_loss': 0.009880091,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13935046,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 56100 per-step time 0.920s\n",
      "I1020 08:34:23.287099 4682925568 model_lib_v2.py:705] Step 56100 per-step time 0.920s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.046333387,\n",
      " 'Loss/localization_loss': 0.0061392076,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10915446,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:34:23.290274 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.046333387,\n",
      " 'Loss/localization_loss': 0.0061392076,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10915446,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 56200 per-step time 0.847s\n",
      "I1020 08:35:47.955629 4682925568 model_lib_v2.py:705] Step 56200 per-step time 0.847s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03485755,\n",
      " 'Loss/localization_loss': 0.008767716,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10030714,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:35:47.955865 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03485755,\n",
      " 'Loss/localization_loss': 0.008767716,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10030714,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 56300 per-step time 0.846s\n",
      "I1020 08:37:12.595755 4682925568 model_lib_v2.py:705] Step 56300 per-step time 0.846s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.031198395,\n",
      " 'Loss/localization_loss': 0.007221738,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.095102005,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:37:12.596013 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.031198395,\n",
      " 'Loss/localization_loss': 0.007221738,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.095102005,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 56400 per-step time 0.849s\n",
      "I1020 08:38:37.464761 4682925568 model_lib_v2.py:705] Step 56400 per-step time 0.849s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.044469193,\n",
      " 'Loss/localization_loss': 0.017073255,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.11822432,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:38:37.464997 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.044469193,\n",
      " 'Loss/localization_loss': 0.017073255,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.11822432,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 56500 per-step time 0.847s\n",
      "I1020 08:40:02.181192 4682925568 model_lib_v2.py:705] Step 56500 per-step time 0.847s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.041311942,\n",
      " 'Loss/localization_loss': 0.010410015,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10840383,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:40:02.181430 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.041311942,\n",
      " 'Loss/localization_loss': 0.010410015,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10840383,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 56600 per-step time 0.845s\n",
      "I1020 08:41:26.649977 4682925568 model_lib_v2.py:705] Step 56600 per-step time 0.845s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03201359,\n",
      " 'Loss/localization_loss': 0.008500269,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.09719573,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:41:26.650213 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03201359,\n",
      " 'Loss/localization_loss': 0.008500269,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.09719573,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 56700 per-step time 0.846s\n",
      "I1020 08:42:51.274600 4682925568 model_lib_v2.py:705] Step 56700 per-step time 0.846s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.057736516,\n",
      " 'Loss/localization_loss': 0.0237478,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13816619,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:42:51.275089 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.057736516,\n",
      " 'Loss/localization_loss': 0.0237478,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13816619,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 56800 per-step time 0.896s\n",
      "I1020 08:44:20.861929 4682925568 model_lib_v2.py:705] Step 56800 per-step time 0.896s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.059923567,\n",
      " 'Loss/localization_loss': 0.016652659,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.1332581,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:44:20.864201 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.059923567,\n",
      " 'Loss/localization_loss': 0.016652659,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.1332581,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 56900 per-step time 0.827s\n",
      "I1020 08:45:43.545957 4682925568 model_lib_v2.py:705] Step 56900 per-step time 0.827s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.09050726,\n",
      " 'Loss/localization_loss': 0.04548163,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.19267076,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:45:43.546190 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.09050726,\n",
      " 'Loss/localization_loss': 0.04548163,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.19267076,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 57000 per-step time 0.811s\n",
      "I1020 08:47:04.623588 4682925568 model_lib_v2.py:705] Step 57000 per-step time 0.811s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03741489,\n",
      " 'Loss/localization_loss': 0.03266419,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12676096,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:47:04.623823 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03741489,\n",
      " 'Loss/localization_loss': 0.03266419,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12676096,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 57100 per-step time 0.821s\n",
      "I1020 08:48:26.768051 4682925568 model_lib_v2.py:705] Step 57100 per-step time 0.821s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.11894521,\n",
      " 'Loss/localization_loss': 0.083628945,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.25925604,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:48:26.770577 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.11894521,\n",
      " 'Loss/localization_loss': 0.083628945,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.25925604,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 57200 per-step time 0.795s\n",
      "I1020 08:49:46.259981 4682925568 model_lib_v2.py:705] Step 57200 per-step time 0.795s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06073678,\n",
      " 'Loss/localization_loss': 0.01315893,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13057758,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:49:46.260205 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06073678,\n",
      " 'Loss/localization_loss': 0.01315893,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13057758,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 57300 per-step time 0.881s\n",
      "I1020 08:51:14.351889 4682925568 model_lib_v2.py:705] Step 57300 per-step time 0.881s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08927514,\n",
      " 'Loss/localization_loss': 0.042372625,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.18832964,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:51:14.354434 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08927514,\n",
      " 'Loss/localization_loss': 0.042372625,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.18832964,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 57400 per-step time 1.145s\n",
      "I1020 08:53:08.904217 4682925568 model_lib_v2.py:705] Step 57400 per-step time 1.145s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.057267576,\n",
      " 'Loss/localization_loss': 0.02196151,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13591096,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:53:08.909676 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.057267576,\n",
      " 'Loss/localization_loss': 0.02196151,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13591096,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 57500 per-step time 0.842s\n",
      "I1020 08:54:33.105524 4682925568 model_lib_v2.py:705] Step 57500 per-step time 0.842s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.052660592,\n",
      " 'Loss/localization_loss': 0.0129923215,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.122334786,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:54:33.105755 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.052660592,\n",
      " 'Loss/localization_loss': 0.0129923215,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.122334786,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 57600 per-step time 0.822s\n",
      "I1020 08:55:55.321886 4682925568 model_lib_v2.py:705] Step 57600 per-step time 0.822s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.029650338,\n",
      " 'Loss/localization_loss': 0.003994059,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.090326265,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:55:55.322211 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.029650338,\n",
      " 'Loss/localization_loss': 0.003994059,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.090326265,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 57700 per-step time 0.831s\n",
      "I1020 08:57:18.440666 4682925568 model_lib_v2.py:705] Step 57700 per-step time 0.831s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.036469635,\n",
      " 'Loss/localization_loss': 0.011908147,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10505965,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:57:18.442900 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.036469635,\n",
      " 'Loss/localization_loss': 0.011908147,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10505965,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 57800 per-step time 0.815s\n",
      "I1020 08:58:39.924360 4682925568 model_lib_v2.py:705] Step 57800 per-step time 0.815s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.052384287,\n",
      " 'Loss/localization_loss': 0.015655471,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12472163,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 08:58:39.924623 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.052384287,\n",
      " 'Loss/localization_loss': 0.015655471,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12472163,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 57900 per-step time 0.823s\n",
      "I1020 09:00:02.179802 4682925568 model_lib_v2.py:705] Step 57900 per-step time 0.823s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.017822124,\n",
      " 'Loss/localization_loss': 0.009989543,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.08449354,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:00:02.180038 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.017822124,\n",
      " 'Loss/localization_loss': 0.009989543,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.08449354,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 58000 per-step time 0.820s\n",
      "I1020 09:01:24.192914 4682925568 model_lib_v2.py:705] Step 58000 per-step time 0.820s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06553379,\n",
      " 'Loss/localization_loss': 0.01895768,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.14117333,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:01:24.193176 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06553379,\n",
      " 'Loss/localization_loss': 0.01895768,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.14117333,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 58100 per-step time 0.871s\n",
      "I1020 09:02:51.334131 4682925568 model_lib_v2.py:705] Step 58100 per-step time 0.871s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.033516042,\n",
      " 'Loss/localization_loss': 0.012794026,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10299194,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:02:51.336766 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.033516042,\n",
      " 'Loss/localization_loss': 0.012794026,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10299194,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 58200 per-step time 0.815s\n",
      "I1020 09:04:12.851427 4682925568 model_lib_v2.py:705] Step 58200 per-step time 0.815s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.039074514,\n",
      " 'Loss/localization_loss': 0.011674815,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.1074312,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:04:12.851714 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.039074514,\n",
      " 'Loss/localization_loss': 0.011674815,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.1074312,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 58300 per-step time 0.837s\n",
      "I1020 09:05:36.558242 4682925568 model_lib_v2.py:705] Step 58300 per-step time 0.837s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.050383784,\n",
      " 'Loss/localization_loss': 0.02772122,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13478687,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:05:36.560828 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.050383784,\n",
      " 'Loss/localization_loss': 0.02772122,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.13478687,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 58400 per-step time 0.820s\n",
      "I1020 09:06:58.566368 4682925568 model_lib_v2.py:705] Step 58400 per-step time 0.820s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.08172344,\n",
      " 'Loss/localization_loss': 0.02845561,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.16686091,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:06:58.566602 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.08172344,\n",
      " 'Loss/localization_loss': 0.02845561,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.16686091,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 58500 per-step time 0.816s\n",
      "I1020 09:08:20.179280 4682925568 model_lib_v2.py:705] Step 58500 per-step time 0.816s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.085904494,\n",
      " 'Loss/localization_loss': 0.030910544,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.1734969,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:08:20.179685 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.085904494,\n",
      " 'Loss/localization_loss': 0.030910544,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.1734969,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 58600 per-step time 0.815s\n",
      "I1020 09:09:41.706800 4682925568 model_lib_v2.py:705] Step 58600 per-step time 0.815s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.034088727,\n",
      " 'Loss/localization_loss': 0.0139458915,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.104716495,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:09:41.707033 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.034088727,\n",
      " 'Loss/localization_loss': 0.0139458915,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.104716495,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 58700 per-step time 0.813s\n",
      "I1020 09:11:02.977045 4682925568 model_lib_v2.py:705] Step 58700 per-step time 0.813s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.054563794,\n",
      " 'Loss/localization_loss': 0.03450022,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.14574589,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:11:02.977275 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.054563794,\n",
      " 'Loss/localization_loss': 0.03450022,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.14574589,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 58800 per-step time 0.836s\n",
      "I1020 09:12:26.616544 4682925568 model_lib_v2.py:705] Step 58800 per-step time 0.836s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.06985237,\n",
      " 'Loss/localization_loss': 0.038679898,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.16521414,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:12:26.616796 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.06985237,\n",
      " 'Loss/localization_loss': 0.038679898,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.16521414,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 58900 per-step time 0.819s\n",
      "I1020 09:13:48.545295 4682925568 model_lib_v2.py:705] Step 58900 per-step time 0.819s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03451284,\n",
      " 'Loss/localization_loss': 0.0063684382,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.09756315,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:13:48.545521 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03451284,\n",
      " 'Loss/localization_loss': 0.0063684382,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.09756315,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 59000 per-step time 0.810s\n",
      "I1020 09:15:09.607052 4682925568 model_lib_v2.py:705] Step 59000 per-step time 0.810s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.04501917,\n",
      " 'Loss/localization_loss': 0.015953243,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.11765428,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:15:09.609205 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.04501917,\n",
      " 'Loss/localization_loss': 0.015953243,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.11765428,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 59100 per-step time 0.890s\n",
      "I1020 09:16:38.639773 4682925568 model_lib_v2.py:705] Step 59100 per-step time 0.890s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.050744712,\n",
      " 'Loss/localization_loss': 0.011516613,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.1189432,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:16:38.642050 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.050744712,\n",
      " 'Loss/localization_loss': 0.011516613,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.1189432,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 59200 per-step time 0.813s\n",
      "I1020 09:17:59.907016 4682925568 model_lib_v2.py:705] Step 59200 per-step time 0.813s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.057282146,\n",
      " 'Loss/localization_loss': 0.010454413,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12441843,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:17:59.907271 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.057282146,\n",
      " 'Loss/localization_loss': 0.010454413,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12441843,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 59300 per-step time 0.815s\n",
      "I1020 09:19:21.386575 4682925568 model_lib_v2.py:705] Step 59300 per-step time 0.815s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.0965658,\n",
      " 'Loss/localization_loss': 0.026607933,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.1798556,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:19:21.386827 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.0965658,\n",
      " 'Loss/localization_loss': 0.026607933,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.1798556,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 59400 per-step time 0.812s\n",
      "I1020 09:20:42.638081 4682925568 model_lib_v2.py:705] Step 59400 per-step time 0.812s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.036594126,\n",
      " 'Loss/localization_loss': 0.004742693,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.09801869,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:20:42.638453 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.036594126,\n",
      " 'Loss/localization_loss': 0.004742693,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.09801869,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 59500 per-step time 0.816s\n",
      "I1020 09:22:04.237694 4682925568 model_lib_v2.py:705] Step 59500 per-step time 0.816s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03691805,\n",
      " 'Loss/localization_loss': 0.018315552,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.11191548,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:22:04.239961 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03691805,\n",
      " 'Loss/localization_loss': 0.018315552,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.11191548,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 59600 per-step time 0.815s\n",
      "I1020 09:23:25.779011 4682925568 model_lib_v2.py:705] Step 59600 per-step time 0.815s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.034967847,\n",
      " 'Loss/localization_loss': 0.03525395,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12690367,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:23:25.779239 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.034967847,\n",
      " 'Loss/localization_loss': 0.03525395,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12690367,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 59700 per-step time 0.820s\n",
      "I1020 09:24:47.760982 4682925568 model_lib_v2.py:705] Step 59700 per-step time 0.820s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03325241,\n",
      " 'Loss/localization_loss': 0.006863279,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.096797556,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:24:47.761256 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03325241,\n",
      " 'Loss/localization_loss': 0.006863279,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.096797556,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 59800 per-step time 0.808s\n",
      "I1020 09:26:08.567057 4682925568 model_lib_v2.py:705] Step 59800 per-step time 0.808s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.063644685,\n",
      " 'Loss/localization_loss': 0.01941645,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.139743,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:26:08.567303 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.063644685,\n",
      " 'Loss/localization_loss': 0.01941645,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.139743,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 59900 per-step time 0.829s\n",
      "I1020 09:27:31.451473 4682925568 model_lib_v2.py:705] Step 59900 per-step time 0.829s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.03843419,\n",
      " 'Loss/localization_loss': 0.0059642848,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10108034,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:27:31.453745 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.03843419,\n",
      " 'Loss/localization_loss': 0.0059642848,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.10108034,\n",
      " 'learning_rate': 0.0}\n",
      "INFO:tensorflow:Step 60000 per-step time 0.816s\n",
      "I1020 09:28:53.007172 4682925568 model_lib_v2.py:705] Step 60000 per-step time 0.816s\n",
      "INFO:tensorflow:{'Loss/classification_loss': 0.052042633,\n",
      " 'Loss/localization_loss': 0.014261341,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12298585,\n",
      " 'learning_rate': 0.0}\n",
      "I1020 09:28:53.007405 4682925568 model_lib_v2.py:708] {'Loss/classification_loss': 0.052042633,\n",
      " 'Loss/localization_loss': 0.014261341,\n",
      " 'Loss/regularization_loss': 0.05668187,\n",
      " 'Loss/total_loss': 0.12298585,\n",
      " 'learning_rate': 0.0}\n"
     ]
    }
   ],
   "source": [
    "!{command}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Evaluate the Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "command = \"python {} --model_dir={} --pipeline_config_path={} --checkpoint_dir={}\".format(TRAINING_SCRIPT, paths['CHECKPOINT_PATH'],files['PIPELINE_CONFIG'], paths['CHECKPOINT_PATH'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2022-10-20 09:36:38.719293: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "WARNING:tensorflow:Forced number of epochs for all eval validations to be 1.\n",
      "W1020 09:36:46.867610 4685551104 model_lib_v2.py:1089] Forced number of epochs for all eval validations to be 1.\n",
      "INFO:tensorflow:Maybe overwriting sample_1_of_n_eval_examples: None\n",
      "I1020 09:36:46.867824 4685551104 config_util.py:552] Maybe overwriting sample_1_of_n_eval_examples: None\n",
      "INFO:tensorflow:Maybe overwriting use_bfloat16: False\n",
      "I1020 09:36:46.867902 4685551104 config_util.py:552] Maybe overwriting use_bfloat16: False\n",
      "INFO:tensorflow:Maybe overwriting eval_num_epochs: 1\n",
      "I1020 09:36:46.867973 4685551104 config_util.py:552] Maybe overwriting eval_num_epochs: 1\n",
      "WARNING:tensorflow:Expected number of evaluation epochs is 1, but instead encountered `eval_on_train_input_config.num_epochs` = 0. Overwriting `num_epochs` to 1.\n",
      "W1020 09:36:46.868080 4685551104 model_lib_v2.py:1106] Expected number of evaluation epochs is 1, but instead encountered `eval_on_train_input_config.num_epochs` = 0. Overwriting `num_epochs` to 1.\n",
      "2022-10-20 09:36:46.873474: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "INFO:tensorflow:Reading unweighted datasets: ['Tensorflow/workspace/annotations/test.record']\n",
      "I1020 09:36:46.956741 4685551104 dataset_builder.py:162] Reading unweighted datasets: ['Tensorflow/workspace/annotations/test.record']\n",
      "INFO:tensorflow:Reading record datasets for input file: ['Tensorflow/workspace/annotations/test.record']\n",
      "I1020 09:36:46.957647 4685551104 dataset_builder.py:79] Reading record datasets for input file: ['Tensorflow/workspace/annotations/test.record']\n",
      "INFO:tensorflow:Number of filenames to read: 1\n",
      "I1020 09:36:46.957777 4685551104 dataset_builder.py:80] Number of filenames to read: 1\n",
      "WARNING:tensorflow:num_readers has been reduced to 1 to match input file shards.\n",
      "W1020 09:36:46.957831 4685551104 dataset_builder.py:86] num_readers has been reduced to 1 to match input file shards.\n",
      "WARNING:tensorflow:From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/object_detection/builders/dataset_builder.py:100: parallel_interleave (from tensorflow.python.data.experimental.ops.interleave_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use `tf.data.Dataset.interleave(map_func, cycle_length, block_length, num_parallel_calls=tf.data.AUTOTUNE)` instead. If sloppy execution is desired, use `tf.data.Options.deterministic`.\n",
      "W1020 09:36:46.963583 4685551104 deprecation.py:350] From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/object_detection/builders/dataset_builder.py:100: parallel_interleave (from tensorflow.python.data.experimental.ops.interleave_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use `tf.data.Dataset.interleave(map_func, cycle_length, block_length, num_parallel_calls=tf.data.AUTOTUNE)` instead. If sloppy execution is desired, use `tf.data.Options.deterministic`.\n",
      "WARNING:tensorflow:From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/object_detection/builders/dataset_builder.py:235: DatasetV1.map_with_legacy_function (from tensorflow.python.data.ops.dataset_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use `tf.data.Dataset.map()\n",
      "W1020 09:36:46.999363 4685551104 deprecation.py:350] From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/object_detection/builders/dataset_builder.py:235: DatasetV1.map_with_legacy_function (from tensorflow.python.data.ops.dataset_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use `tf.data.Dataset.map()\n",
      "WARNING:tensorflow:From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/tensorflow/python/util/dispatch.py:1176: sparse_to_dense (from tensorflow.python.ops.sparse_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Create a `tf.sparse.SparseTensor` and use `tf.sparse.to_dense` instead.\n",
      "W1020 09:36:50.978501 4685551104 deprecation.py:350] From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/tensorflow/python/util/dispatch.py:1176: sparse_to_dense (from tensorflow.python.ops.sparse_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Create a `tf.sparse.SparseTensor` and use `tf.sparse.to_dense` instead.\n",
      "WARNING:tensorflow:From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/tensorflow/python/util/dispatch.py:1176: to_float (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use `tf.cast` instead.\n",
      "W1020 09:36:51.871863 4685551104 deprecation.py:350] From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/tensorflow/python/util/dispatch.py:1176: to_float (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use `tf.cast` instead.\n",
      "INFO:tensorflow:Waiting for new checkpoint at Tensorflow/workspace/models/my_ssd_mobnet\n",
      "I1020 09:36:54.190524 4685551104 checkpoint_utils.py:142] Waiting for new checkpoint at Tensorflow/workspace/models/my_ssd_mobnet\n",
      "INFO:tensorflow:Found new checkpoint at Tensorflow/workspace/models/my_ssd_mobnet/ckpt-61\n",
      "I1020 09:36:54.191749 4685551104 checkpoint_utils.py:151] Found new checkpoint at Tensorflow/workspace/models/my_ssd_mobnet/ckpt-61\n",
      "/Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/keras/backend.py:451: UserWarning: `tf.keras.backend.set_learning_phase` is deprecated and will be removed after 2020-10-11. To update it, simply pass a True/False value to the `training` argument of the `__call__` method of your layer or model.\n",
      "  warnings.warn(\n",
      "WARNING:tensorflow:From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/tensorflow/python/util/dispatch.py:1176: to_int64 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use `tf.cast` instead.\n",
      "W1020 09:37:19.627454 4685551104 deprecation.py:350] From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/tensorflow/python/util/dispatch.py:1176: to_int64 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use `tf.cast` instead.\n",
      "INFO:tensorflow:Finished eval step 0\n",
      "I1020 09:37:19.645826 4685551104 model_lib_v2.py:966] Finished eval step 0\n",
      "WARNING:tensorflow:From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/tensorflow/python/autograph/impl/api.py:459: py_func (from tensorflow.python.ops.script_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "tf.py_func is deprecated in TF V2. Instead, there are two\n",
      "    options available in V2.\n",
      "    - tf.py_function takes a python function which manipulates tf eager\n",
      "    tensors instead of numpy arrays. It's easy to convert a tf eager tensor to\n",
      "    an ndarray (just call tensor.numpy()) but having access to eager tensors\n",
      "    means `tf.py_function`s can use accelerators such as GPUs as well as\n",
      "    being differentiable using a gradient tape.\n",
      "    - tf.numpy_function maintains the semantics of the deprecated tf.py_func\n",
      "    (it is not differentiable, and manipulates numpy arrays). It drops the\n",
      "    stateful argument making all functions stateful.\n",
      "    \n",
      "W1020 09:37:19.847602 4685551104 deprecation.py:350] From /Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/tensorflow/python/autograph/impl/api.py:459: py_func (from tensorflow.python.ops.script_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "tf.py_func is deprecated in TF V2. Instead, there are two\n",
      "    options available in V2.\n",
      "    - tf.py_function takes a python function which manipulates tf eager\n",
      "    tensors instead of numpy arrays. It's easy to convert a tf eager tensor to\n",
      "    an ndarray (just call tensor.numpy()) but having access to eager tensors\n",
      "    means `tf.py_function`s can use accelerators such as GPUs as well as\n",
      "    being differentiable using a gradient tape.\n",
      "    - tf.numpy_function maintains the semantics of the deprecated tf.py_func\n",
      "    (it is not differentiable, and manipulates numpy arrays). It drops the\n",
      "    stateful argument making all functions stateful.\n",
      "    \n",
      "INFO:tensorflow:Finished eval step 100\n",
      "I1020 09:37:30.052146 4685551104 model_lib_v2.py:966] Finished eval step 100\n",
      "INFO:tensorflow:Finished eval step 200\n",
      "I1020 09:37:38.446758 4685551104 model_lib_v2.py:966] Finished eval step 200\n",
      "INFO:tensorflow:Finished eval step 300\n",
      "I1020 09:37:46.338866 4685551104 model_lib_v2.py:966] Finished eval step 300\n",
      "INFO:tensorflow:Finished eval step 400\n",
      "I1020 09:37:53.294330 4685551104 model_lib_v2.py:966] Finished eval step 400\n",
      "INFO:tensorflow:Finished eval step 500\n",
      "I1020 09:38:01.172449 4685551104 model_lib_v2.py:966] Finished eval step 500\n",
      "INFO:tensorflow:Finished eval step 600\n",
      "I1020 09:38:09.351558 4685551104 model_lib_v2.py:966] Finished eval step 600\n",
      "INFO:tensorflow:Finished eval step 700\n",
      "I1020 09:38:17.411180 4685551104 model_lib_v2.py:966] Finished eval step 700\n",
      "INFO:tensorflow:Finished eval step 800\n",
      "I1020 09:38:25.880192 4685551104 model_lib_v2.py:966] Finished eval step 800\n",
      "INFO:tensorflow:Finished eval step 900\n",
      "I1020 09:38:33.392398 4685551104 model_lib_v2.py:966] Finished eval step 900\n",
      "INFO:tensorflow:Finished eval step 1000\n",
      "I1020 09:38:41.581067 4685551104 model_lib_v2.py:966] Finished eval step 1000\n",
      "INFO:tensorflow:Finished eval step 1100\n",
      "I1020 09:38:49.388792 4685551104 model_lib_v2.py:966] Finished eval step 1100\n",
      "INFO:tensorflow:Finished eval step 1200\n",
      "I1020 09:38:57.263008 4685551104 model_lib_v2.py:966] Finished eval step 1200\n",
      "INFO:tensorflow:Finished eval step 1300\n",
      "I1020 09:39:06.115298 4685551104 model_lib_v2.py:966] Finished eval step 1300\n",
      "INFO:tensorflow:Finished eval step 1400\n",
      "I1020 09:39:14.391453 4685551104 model_lib_v2.py:966] Finished eval step 1400\n",
      "INFO:tensorflow:Finished eval step 1500\n",
      "I1020 09:39:21.921330 4685551104 model_lib_v2.py:966] Finished eval step 1500\n",
      "INFO:tensorflow:Finished eval step 1600\n",
      "I1020 09:39:29.161106 4685551104 model_lib_v2.py:966] Finished eval step 1600\n",
      "INFO:tensorflow:Finished eval step 1700\n",
      "I1020 09:39:36.312777 4685551104 model_lib_v2.py:966] Finished eval step 1700\n",
      "INFO:tensorflow:Finished eval step 1800\n",
      "I1020 09:39:44.305160 4685551104 model_lib_v2.py:966] Finished eval step 1800\n",
      "INFO:tensorflow:Finished eval step 1900\n",
      "I1020 09:39:51.287573 4685551104 model_lib_v2.py:966] Finished eval step 1900\n",
      "INFO:tensorflow:Finished eval step 2000\n",
      "I1020 09:39:58.936908 4685551104 model_lib_v2.py:966] Finished eval step 2000\n",
      "INFO:tensorflow:Finished eval step 2100\n",
      "I1020 09:40:06.658300 4685551104 model_lib_v2.py:966] Finished eval step 2100\n",
      "INFO:tensorflow:Finished eval step 2200\n",
      "I1020 09:40:13.575745 4685551104 model_lib_v2.py:966] Finished eval step 2200\n",
      "INFO:tensorflow:Finished eval step 2300\n",
      "I1020 09:40:20.217338 4685551104 model_lib_v2.py:966] Finished eval step 2300\n",
      "INFO:tensorflow:Finished eval step 2400\n",
      "I1020 09:40:26.881622 4685551104 model_lib_v2.py:966] Finished eval step 2400\n",
      "INFO:tensorflow:Finished eval step 2500\n",
      "I1020 09:40:33.503698 4685551104 model_lib_v2.py:966] Finished eval step 2500\n",
      "INFO:tensorflow:Finished eval step 2600\n",
      "I1020 09:40:40.055673 4685551104 model_lib_v2.py:966] Finished eval step 2600\n",
      "INFO:tensorflow:Finished eval step 2700\n",
      "I1020 09:40:46.492365 4685551104 model_lib_v2.py:966] Finished eval step 2700\n",
      "INFO:tensorflow:Performing evaluation on 2726 images.\n",
      "I1020 09:40:48.201750 4685551104 coco_evaluation.py:293] Performing evaluation on 2726 images.\n",
      "creating index...\n",
      "index created!\n",
      "INFO:tensorflow:Loading and preparing annotation results...\n",
      "I1020 09:40:48.209238 4685551104 coco_tools.py:116] Loading and preparing annotation results...\n",
      "INFO:tensorflow:DONE (t=0.16s)\n",
      "I1020 09:40:48.371578 4685551104 coco_tools.py:138] DONE (t=0.16s)\n",
      "creating index...\n",
      "index created!\n",
      "Running per image evaluation...\n",
      "Evaluate annotation type *bbox*\n",
      "DONE (t=11.40s).\n",
      "Accumulating evaluation results...\n",
      "DONE (t=4.27s).\n",
      " Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.876\n",
      " Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 1.000\n",
      " Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.992\n",
      " Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = -1.000\n",
      " Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = -1.000\n",
      " Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.876\n",
      " Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.903\n",
      " Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.904\n",
      " Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.904\n",
      " Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = -1.000\n",
      " Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = -1.000\n",
      " Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.904\n",
      "INFO:tensorflow:Eval metrics at step 60000\n",
      "I1020 09:41:04.797380 4685551104 model_lib_v2.py:1015] Eval metrics at step 60000\n",
      "INFO:tensorflow:\t+ DetectionBoxes_Precision/mAP: 0.876417\n",
      "I1020 09:41:04.819560 4685551104 model_lib_v2.py:1018] \t+ DetectionBoxes_Precision/mAP: 0.876417\n",
      "INFO:tensorflow:\t+ DetectionBoxes_Precision/mAP@.50IOU: 0.999709\n",
      "I1020 09:41:04.820291 4685551104 model_lib_v2.py:1018] \t+ DetectionBoxes_Precision/mAP@.50IOU: 0.999709\n",
      "INFO:tensorflow:\t+ DetectionBoxes_Precision/mAP@.75IOU: 0.991950\n",
      "I1020 09:41:04.821019 4685551104 model_lib_v2.py:1018] \t+ DetectionBoxes_Precision/mAP@.75IOU: 0.991950\n",
      "INFO:tensorflow:\t+ DetectionBoxes_Precision/mAP (small): -1.000000\n",
      "I1020 09:41:04.821733 4685551104 model_lib_v2.py:1018] \t+ DetectionBoxes_Precision/mAP (small): -1.000000\n",
      "INFO:tensorflow:\t+ DetectionBoxes_Precision/mAP (medium): -1.000000\n",
      "I1020 09:41:04.822328 4685551104 model_lib_v2.py:1018] \t+ DetectionBoxes_Precision/mAP (medium): -1.000000\n",
      "INFO:tensorflow:\t+ DetectionBoxes_Precision/mAP (large): 0.876417\n",
      "I1020 09:41:04.822865 4685551104 model_lib_v2.py:1018] \t+ DetectionBoxes_Precision/mAP (large): 0.876417\n",
      "INFO:tensorflow:\t+ DetectionBoxes_Recall/AR@1: 0.903056\n",
      "I1020 09:41:04.823384 4685551104 model_lib_v2.py:1018] \t+ DetectionBoxes_Recall/AR@1: 0.903056\n",
      "INFO:tensorflow:\t+ DetectionBoxes_Recall/AR@10: 0.903802\n",
      "I1020 09:41:04.823933 4685551104 model_lib_v2.py:1018] \t+ DetectionBoxes_Recall/AR@10: 0.903802\n",
      "INFO:tensorflow:\t+ DetectionBoxes_Recall/AR@100: 0.903874\n",
      "I1020 09:41:04.824442 4685551104 model_lib_v2.py:1018] \t+ DetectionBoxes_Recall/AR@100: 0.903874\n",
      "INFO:tensorflow:\t+ DetectionBoxes_Recall/AR@100 (small): -1.000000\n",
      "I1020 09:41:04.824923 4685551104 model_lib_v2.py:1018] \t+ DetectionBoxes_Recall/AR@100 (small): -1.000000\n",
      "INFO:tensorflow:\t+ DetectionBoxes_Recall/AR@100 (medium): -1.000000\n",
      "I1020 09:41:04.825495 4685551104 model_lib_v2.py:1018] \t+ DetectionBoxes_Recall/AR@100 (medium): -1.000000\n",
      "INFO:tensorflow:\t+ DetectionBoxes_Recall/AR@100 (large): 0.903874\n",
      "I1020 09:41:04.826066 4685551104 model_lib_v2.py:1018] \t+ DetectionBoxes_Recall/AR@100 (large): 0.903874\n",
      "INFO:tensorflow:\t+ Loss/localization_loss: 0.021997\n",
      "I1020 09:41:04.826504 4685551104 model_lib_v2.py:1018] \t+ Loss/localization_loss: 0.021997\n",
      "INFO:tensorflow:\t+ Loss/classification_loss: 0.064621\n",
      "I1020 09:41:04.827448 4685551104 model_lib_v2.py:1018] \t+ Loss/classification_loss: 0.064621\n",
      "INFO:tensorflow:\t+ Loss/regularization_loss: 0.056682\n",
      "I1020 09:41:04.827892 4685551104 model_lib_v2.py:1018] \t+ Loss/regularization_loss: 0.056682\n",
      "INFO:tensorflow:\t+ Loss/total_loss: 0.143300\n",
      "I1020 09:41:04.828305 4685551104 model_lib_v2.py:1018] \t+ Loss/total_loss: 0.143300\n",
      "INFO:tensorflow:Waiting for new checkpoint at Tensorflow/workspace/models/my_ssd_mobnet\n",
      "I1020 09:41:54.196712 4685551104 checkpoint_utils.py:142] Waiting for new checkpoint at Tensorflow/workspace/models/my_ssd_mobnet\n",
      "^C\n",
      "Traceback (most recent call last):\n",
      "  File \"Tensorflow/models/research/object_detection/model_main_tf2.py\", line 114, in <module>\n",
      "    tf.compat.v1.app.run()\n",
      "  File \"/Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/tensorflow/python/platform/app.py\", line 36, in run\n",
      "    _run(main=main, argv=argv, flags_parser=_parse_flags_tolerate_undef)\n",
      "  File \"/Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/absl/app.py\", line 308, in run\n",
      "    _run_main(main, args)\n",
      "  File \"/Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/absl/app.py\", line 254, in _run_main\n",
      "    sys.exit(main(argv))\n",
      "  File \"Tensorflow/models/research/object_detection/model_main_tf2.py\", line 81, in main\n",
      "    model_lib_v2.eval_continuously(\n",
      "  File \"/Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/object_detection/model_lib_v2.py\", line 1135, in eval_continuously\n",
      "    for latest_checkpoint in tf.train.checkpoints_iterator(\n",
      "  File \"/Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/tensorflow/python/training/checkpoint_utils.py\", line 200, in checkpoints_iterator\n",
      "    new_checkpoint_path = wait_for_new_checkpoint(\n",
      "  File \"/Users/ritvikseth/opt/anaconda3/envs/asl_new/lib/python3.8/site-packages/tensorflow/python/training/checkpoint_utils.py\", line 149, in wait_for_new_checkpoint\n",
      "    time.sleep(seconds_to_sleep)\n",
      "KeyboardInterrupt\n"
     ]
    }
   ],
   "source": [
    "!{command}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Load model from checkpoint"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import tensorflow as tf\n",
    "from object_detection.utils import label_map_util\n",
    "from object_detection.utils import visualization_utils as viz_utils\n",
    "from object_detection.builders import model_builder\n",
    "from object_detection.utils import config_util"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-10-21 13:58:23.751251: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n"
     ]
    }
   ],
   "source": [
    "# Load pipeline config and build a detection model\n",
    "configs = config_util.get_configs_from_pipeline_file(files['PIPELINE_CONFIG'])\n",
    "detection_model = model_builder.build(model_config=configs['model'], is_training=False)\n",
    "\n",
    "# Restore checkpoint\n",
    "ckpt = tf.compat.v2.train.Checkpoint(model=detection_model)\n",
    "ckpt.restore(os.path.join(paths['CHECKPOINT_PATH'], 'ckpt-61')).expect_partial()\n",
    "\n",
    "@tf.function\n",
    "def detect_fn(image):\n",
    "    image, shapes = detection_model.preprocess(image)\n",
    "    prediction_dict = detection_model.predict(image, shapes)\n",
    "    detections = detection_model.postprocess(prediction_dict, shapes)\n",
    "    return detections\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Detect an image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "import cv2 \n",
    "import numpy as np\n",
    "from matplotlib import pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "category_index = label_map_util.create_category_index_from_labelmap(files['LABELMAP'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "IMAGE_PATH = os.path.join(paths['IMAGE_PATH'], 'test', 'Y.44789998-3957-11ed-bf11-acde48001122.jpg')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from PIL import ImageFont\n",
    "img = cv2.imread(IMAGE_PATH)\n",
    "image_np = np.array(img)\n",
    "\n",
    "input_tensor = tf.convert_to_tensor(np.expand_dims(image_np, 0), dtype=tf.float32)\n",
    "detections = detect_fn(input_tensor)\n",
    "\n",
    "num_detections = int(detections.pop('num_detections'))\n",
    "detections = {key: value[0, :num_detections].numpy()\n",
    "              for key, value in detections.items()}\n",
    "detections['num_detections'] = num_detections\n",
    "\n",
    "# detection_classes should be ints.\n",
    "detections['detection_classes'] = detections['detection_classes'].astype(np.int64)\n",
    "\n",
    "label_id_offset = 1\n",
    "image_np_with_detections = image_np.copy()\n",
    "\n",
    "viz_utils.visualize_boxes_and_labels_on_image_array(\n",
    "            image_np_with_detections,\n",
    "            detections['detection_boxes'],\n",
    "            detections['detection_classes']+label_id_offset,\n",
    "            detections['detection_scores'],\n",
    "            category_index,\n",
    "            use_normalized_coordinates=True,\n",
    "            max_boxes_to_draw=5,\n",
    "            min_score_thresh=.3,\n",
    "            agnostic_mode=False,\n",
    "            line_thickness=10,\n",
    "            # font = ImageFont.truetype(30)\n",
    "        )\n",
    "\n",
    "plt.imshow(cv2.cvtColor(image_np_with_detections, cv2.COLOR_BGR2RGB))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "RealTime Detections"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [],
   "source": [
    "cap = cv2.VideoCapture(0)\n",
    "width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))\n",
    "height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))\n",
    "\n",
    "while cap.isOpened(): \n",
    "    ret, frame = cap.read()\n",
    "    image_np = np.array(frame)\n",
    "    \n",
    "    input_tensor = tf.convert_to_tensor(np.expand_dims(image_np, 0), dtype=tf.float32)\n",
    "    detections = detect_fn(input_tensor)\n",
    "    \n",
    "    num_detections = int(detections.pop('num_detections'))\n",
    "    detections = {key: value[0, :num_detections].numpy()\n",
    "                  for key, value in detections.items()}\n",
    "    detections['num_detections'] = num_detections\n",
    "\n",
    "    # detection_classes should be ints.\n",
    "    detections['detection_classes'] = detections['detection_classes'].astype(np.int64)\n",
    "\n",
    "    label_id_offset = 1\n",
    "    image_np_with_detections = image_np.copy()\n",
    "\n",
    "    viz_utils.visualize_boxes_and_labels_on_image_array(\n",
    "                image_np_with_detections,\n",
    "                detections['detection_boxes'],\n",
    "                detections['detection_classes']+label_id_offset,\n",
    "                detections['detection_scores'],\n",
    "                category_index,\n",
    "                use_normalized_coordinates=True,\n",
    "                max_boxes_to_draw=5,\n",
    "                min_score_thresh=.2,\n",
    "                agnostic_mode=False)\n",
    "\n",
    "    cv2.imshow('object detection', cv2.resize(image_np_with_detections, (800, 600)))\n",
    "    \n",
    "    if cv2.waitKey(10) & 0xFF == ord('q'):\n",
    "        cap.release()\n",
    "        cv2.destroyAllWindows()\n",
    "        break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "name": "3. Training and Detection.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3.10.6 ('.venv': venv)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.8"
  },
  "vscode": {
   "interpreter": {
    "hash": "0815a73c222cb5321ee52471892dda2fd99823873db920dd40274b8a6c97254c"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}