PyTorch Example

This is a complete example of PyTorch code that trains a CNN and saves to W&B.

You can find this example on GitHub and see the results on W&B.

from __future__ import print_function
import argparse
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from torchvision import datasets, transforms
import wandb
class Net(nn.Module):
def __init__(self):
super(Net, self).__init__()
self.conv1 = nn.Conv2d(1, 10, kernel_size=5)
self.conv2 = nn.Conv2d(10, 20, kernel_size=5)
self.conv2_drop = nn.Dropout2d()
self.fc1 = nn.Linear(320, 50)
self.fc2 = nn.Linear(50, 10)
def forward(self, x):
x = F.relu(F.max_pool2d(self.conv1(x), 2))
x = F.relu(F.max_pool2d(self.conv2_drop(self.conv2(x)), 2))
x = x.view(-1, 320)
x = F.relu(self.fc1(x))
x = F.dropout(x,
x = self.fc2(x)
return F.log_softmax(x, dim=1)
def train(config, model, device, train_loader, optimizer, epoch):
for batch_idx, (data, target) in enumerate(train_loader):
data, target =,
output = model(data)
loss = F.nll_loss(output, target)
if batch_idx % config.log_interval == 0:
print('Train Epoch: {} [{}/{} ({:.0f}%)]\tLoss: {:.6f}'.format(
epoch, batch_idx * len(data), len(train_loader.dataset),
100. * batch_idx / len(train_loader), loss.item()))
def test(config, model, device, test_loader):
test_loss = 0
correct = 0
example_images = []
with torch.no_grad():
for data, target in test_loader:
data, target =,
output = model(data)
test_loss += F.nll_loss(output, target, reduction='sum').item() # sum up batch loss
pred = output.max(1, keepdim=True)[1] # get the index of the max log-probability
correct += pred.eq(target.view_as(pred)).sum().item()
# Save the first input tensor in each test batch as an example image
example_images.append(wandb.Image(data[0], caption="Pred: {} Truth: {}".format(pred[0].item(), target[0])))
test_loss /= len(test_loader.dataset)
print('\nTest set: Average loss: {:.4f}, Accuracy: {}/{} ({:.0f}%)\n'.format(
test_loss, correct, len(test_loader.dataset),
100. * correct / len(test_loader.dataset)))
# Log the images and metrics
"Examples": example_images,
"Test Accuracy": 100. * correct / len(test_loader.dataset),
"Test Loss": test_loss})
def main():
# Training settings
parser = argparse.ArgumentParser(description='PyTorch MNIST Example')
parser.add_argument('--batch-size', type=int, default=64, metavar='N',
help='input batch size for training (default: 64)')
parser.add_argument('--test-batch-size', type=int, default=1000, metavar='N',
help='input batch size for testing (default: 1000)')
parser.add_argument('--epochs', type=int, default=10, metavar='N',
help='number of epochs to train (default: 10)')
parser.add_argument('--lr', type=float, default=0.01, metavar='LR',
help='learning rate (default: 0.01)')
parser.add_argument('--momentum', type=float, default=0.5, metavar='M',
help='SGD momentum (default: 0.5)')
parser.add_argument('--no-cuda', action='store_true', default=False,
help='disables CUDA training')
parser.add_argument('--seed', type=int, default=1, metavar='S',
help='random seed (default: 1)')
parser.add_argument('--log-interval', type=int, default=10, metavar='N',
help='how many batches to wait before logging training status')
args = parser.parse_args()
use_cuda = not args.no_cuda and torch.cuda.is_available()
# We load all of the arguments into config to save as hyperparameters
# Config is a variable that holds and saves hyperparameters and inputs
config = wandb.config
device = torch.device("cuda" if use_cuda else "cpu")
kwargs = {'num_workers': 1, 'pin_memory': True} if use_cuda else {}
train_loader =
datasets.MNIST('../data', train=True, download=True,
transforms.Normalize((0.1307,), (0.3081,))
batch_size=config.batch_size, shuffle=True, **kwargs)
test_loader =
datasets.MNIST('../data', train=False, transform=transforms.Compose([
transforms.Normalize((0.1307,), (0.3081,))
batch_size=config.test_batch_size, shuffle=True, **kwargs)
model = Net().to(device)
optimizer = optim.SGD(model.parameters(),, momentum=config.momentum)
# This magic line lets us save the pytorch model and track all of the gradients and optionally parameters
for epoch in range(1, config.epochs + 1):
train(config, model, device, train_loader, optimizer, epoch)
test(config, model, device, test_loader)
if __name__ == '__main__':