Skip to content

Latest commit

 

History

History
 
 

mask_rcnn

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Mask R-CNN

Introduction

@article{He_2017,
   title={Mask R-CNN},
   journal={2017 IEEE International Conference on Computer Vision (ICCV)},
   publisher={IEEE},
   author={He, Kaiming and Gkioxari, Georgia and Dollar, Piotr and Girshick, Ross},
   year={2017},
   month={Oct}
}

Results and models

Backbone Style Lr schd Mem (GB) Inf time (fps) box AP mask AP Download
R-50-FPN caffe 1x 4.3 38.0 34.4 model | log
R-50-FPN pytorch 1x 4.4 16.1 38.2 34.7 model | log
R-50-FPN pytorch 2x - - 39.2 35.4 model | log
R-101-FPN caffe 1x 40.4 36.4 model | log
R-101-FPN pytorch 1x 6.4 13.5 40.0 36.1 model | log
R-101-FPN pytorch 2x - - 40.8 36.6 model | log
X-101-32x4d-FPN pytorch 1x 7.6 11.3 41.9 37.5 model | log
X-101-32x4d-FPN pytorch 2x - - 42.2 37.8 model | log
X-101-64x4d-FPN pytorch 1x 10.7 8.0 42.8 38.4 model | log
X-101-64x4d-FPN pytorch 2x - - 42.7 38.1 model | log
X-101-32x8d-FPN pytorch 1x - - 42.8 38.3

Pre-trained Models

We also train some models with longer schedules and multi-scale training. The users could finetune them for downstream tasks.

Backbone Style Lr schd Mem (GB) Inf time (fps) box AP mask AP Download
R-50-FPN caffe 2x 4.3 40.3 36.5 model | log
R-50-FPN caffe 3x 4.3 40.8 37.0 model | log
X-101-32x8d-FPN pytorch 1x - 43.6 39.0
X-101-32x8d-FPN pytorch 3x - 44.0 39.3