PyTorch device (MPS, CUDA, CPU)
import torch
# Get cpu, gpu or mps device for training.
device = (
"cuda"
if torch.cuda.is_available()
else "mps"
if torch.backends.mps.is_available()
else "cpu"
)
print(f"Using {device=}")Matplotlib multiple plot
import numpy as np
import matplotlib.pyplot as plt
x = np.arange(1,11)
y = 3*x+2
z = x**2+5
# --------------------------------------------
fig, axs = plt.subplots(nrows=2, ncols=2, figsize=(12, 4))
axs[0][0].plot(x, y)
axs[0][1].plot(x, z)
axs[1][0].plot(x, z)
axs[1][1].plot(x, y)
plt.show()Argument Parser in python
import os
from argparse import ArgumentParser
if __name__ == "__main__":
# enable CLI commands
parser = ArgumentParser()
parser.add_argument('--data', type=str,
default=os.getcwd() + '/example_dataset')
parser.add_argument('--max_steps', type=int, default=100)
parser.add_argument('--lr', type=float, default=1e-3)
parser.add_argument('--batch_size', type=int, default=2)
args = parser.parse_args()
print(f"{args=}")
print(f"batch size: {args.batch_size}")