mlflow_logger.py

import mlflow
import mlflow.pytorch


class MlFlowLogger(object):
    def __init__(self,
                 uri=None,
                 experiment_name=None,
                 model_path='models',
                 **params):
        self.params = params
        self.experiment_name = experiment_name
        self.run = None
        self.uri = uri
        self.model_path = model_path
        self.start()

    def get_or_set_experiment(self):
        print(mlflow.get_tracking_uri())
        if self.experiment_name is None:
            return
        try:
            mlflow.create_experiment(self.experiment_name)
        except Exception:
            print('Experiment {} already exists'
                  .format(self.experiment_name))
        mlflow.set_experiment(self.experiment_name)

    @staticmethod
    def log_param(k, v):
        mlflow.log_param(k, v)

    def log_params(self, params=None):
        if params is None:
            params = self.params

        for k, v in params.items():
            self.log_param(k, v)

    @staticmethod
    def log_metric(k, v):
        mlflow.log_metric(k, v)

    def log_metrics(self, metrics):
        for k, v in metrics.items():
            self.log_metric(k, v)

    def log_model(self, model):
        """ for local saving of models """
        mlflow.pytorch.save_model(model, self.model_path)

    def start(self):
        mlflow.set_tracking_uri(self.uri)
        self.get_or_set_experiment()
        self.run = mlflow.start_run()
        self.log_params()

    def end(self):
        mlflow.end_run()

mnist_example.py

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 time
from mlflow_logger import MlFlowLogger

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, training=self.training)
        x = self.fc2(x)
        return F.log_softmax(x, dim=1)

def train(args, model, device, train_loader, optimizer, epoch):
    model.train()
    running_loss = 0
    total = 0
    for batch_idx, (data, target) in enumerate(train_loader):
        data, target = data.to(device), target.to(device)
        optimizer.zero_grad()
        output = model(data)
        loss = F.nll_loss(output, target)
        running_loss += loss.item()
        total += 1
        loss.backward()
        optimizer.step()
        if batch_idx % args.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()))
    avg_loss = float(running_loss) / total
    return avg_loss

def test(args, model, device, test_loader):
    model.eval()
    test_loss = 0
    correct = 0
    with torch.no_grad():
        for data, target in test_loader:
            data, target = data.to(device), target.to(device)
            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()

    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)))
    return {'val_loss': test_loss, 'val_accuracy': 100. * correct / len(test_loader.dataset)}

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()

    torch.manual_seed(args.seed)

    device = torch.device("cuda" if use_cuda else "cpu")

    config = vars(args)
    import pprint

    print('Running with configuration:')
    pprint.pprint(config)

    logger = MlFlowLogger(experiment_name='mnist', **config)

    kwargs = {'num_workers': 1, 'pin_memory': True} if use_cuda else {}
    train_loader = torch.utils.data.DataLoader(
        datasets.MNIST('../data', train=True, download=True,
                       transform=transforms.Compose([
                           transforms.ToTensor(),
                           transforms.Normalize((0.1307,), (0.3081,))
                       ])),
        batch_size=args.batch_size, shuffle=True, **kwargs)
    test_loader = torch.utils.data.DataLoader(
        datasets.MNIST('../data', train=False, transform=transforms.Compose([
                           transforms.ToTensor(),
                           transforms.Normalize((0.1307,), (0.3081,))
                       ])),
        batch_size=args.test_batch_size, shuffle=True, **kwargs)


    model = Net().to(device)
    optimizer = optim.SGD(model.parameters(), lr=args.lr, momentum=args.momentum)

    st1 = time.time()
    final_accuracy = 0
    for epoch in range(1, args.epochs + 1):
        st = time.time()
        train_loss = train(args, model, device, train_loader, optimizer, epoch)
        et = time.time()
        train_seconds = et - st
        logger.log_metric('train_loss', train_loss)
        logger.log_metric('train_seconds', train_seconds)
        st = time.time()
        val_metrics = test(args, model, device, test_loader)
        logger.log_metrics(val_metrics)
        et = time.time()
        val_seconds = et - st
        logger.log_metric('val_seconds', val_seconds)
        final_accuracy = val_metrics['val_accuracy']

    total_time = time.time() - st1
    logger.log_metric('total_time', total_time)
    logger.log_metric('final_accuracy', final_accuracy)


if __name__ == '__main__':
    main()