Introduction || Tensors || Autograd || Building Models || TensorBoard Support || Training Models || Model Understanding

Training with PyTorch

Created On: Nov 30, 2021 | Last Updated: May 31, 2023 | Last Verified: Nov 05, 2024

Follow along with the video below or on youtube.

Introduction

In past videos, we’ve discussed and demonstrated:

• Building models with the neural network layers and functions of the torch.nn module

• The mechanics of automated gradient computation, which is central to gradient-based model training

• Using TensorBoard to visualize training progress and other activities

In this video, we’ll be adding some new tools to your inventory:

• We’ll get familiar with the dataset and dataloader abstractions, and how they ease the process of feeding data to your model during a training loop

• We’ll discuss specific loss functions and when to use them

• We’ll look at PyTorch optimizers, which implement algorithms to adjust model weights based on the outcome of a loss function

Finally, we’ll pull all of these together and see a full PyTorch training loop in action.

Dataset and DataLoader

The Dataset and DataLoader classes encapsulate the process of pulling your data from storage and exposing it to your training loop in batches.

The Dataset is responsible for accessing and processing single instances of data.

The DataLoader pulls instances of data from the Dataset (either automatically or with a sampler that you define), collects them in batches, and returns them for consumption by your training loop. The DataLoader works with all kinds of datasets, regardless of the type of data they contain.

For this tutorial, we’ll be using the Fashion-MNIST dataset provided by TorchVision. We use torchvision.transforms.Normalize() to zero-center and normalize the distribution of the image tile content, and download both training and validation data splits.

import torch
import torchvision
import torchvision.transforms as transforms

# PyTorch TensorBoard support
from torch.utils.tensorboard import SummaryWriter
from datetime import datetime


transform = transforms.Compose(
    [transforms.ToTensor(),
    transforms.Normalize((0.5,), (0.5,))])

# Create datasets for training & validation, download if necessary
training_set = torchvision.datasets.FashionMNIST('./data', train=True, transform=transform, download=True)
validation_set = torchvision.datasets.FashionMNIST('./data', train=False, transform=transform, download=True)

# Create data loaders for our datasets; shuffle for training, not for validation
training_loader = torch.utils.data.DataLoader(training_set, batch_size=4, shuffle=True)
validation_loader = torch.utils.data.DataLoader(validation_set, batch_size=4, shuffle=False)

# Class labels
classes = ('T-shirt/top', 'Trouser', 'Pullover', 'Dress', 'Coat',
        'Sandal', 'Shirt', 'Sneaker', 'Bag', 'Ankle Boot')

# Report split sizes
print('Training set has {} instances'.format(len(training_set)))
print('Validation set has {} instances'.format(len(validation_set)))

  0%|          | 0.00/26.4M [00:00<?, ?B/s]
  0%|          | 65.5k/26.4M [00:00<01:12, 364kB/s]
  1%|          | 197k/26.4M [00:00<00:35, 732kB/s]
  2%|1         | 492k/26.4M [00:00<00:20, 1.29MB/s]
  6%|6         | 1.64M/26.4M [00:00<00:05, 4.19MB/s]
 15%|#4        | 3.83M/26.4M [00:00<00:02, 8.09MB/s]
 33%|###2      | 8.68M/26.4M [00:00<00:00, 18.0MB/s]
 49%|####8     | 12.9M/26.4M [00:00<00:00, 21.3MB/s]
 66%|######5   | 17.4M/26.4M [00:01<00:00, 26.0MB/s]
 83%|########2 | 21.9M/26.4M [00:01<00:00, 27.2MB/s]
 99%|#########8| 26.1M/26.4M [00:01<00:00, 30.6MB/s]
100%|##########| 26.4M/26.4M [00:01<00:00, 19.3MB/s]

  0%|          | 0.00/29.5k [00:00<?, ?B/s]
100%|##########| 29.5k/29.5k [00:00<00:00, 326kB/s]

  0%|          | 0.00/4.42M [00:00<?, ?B/s]
  1%|1         | 65.5k/4.42M [00:00<00:12, 361kB/s]
  5%|5         | 229k/4.42M [00:00<00:06, 678kB/s]
 19%|#9        | 852k/4.42M [00:00<00:01, 2.39MB/s]
 44%|####3     | 1.93M/4.42M [00:00<00:00, 4.12MB/s]
100%|##########| 4.42M/4.42M [00:00<00:00, 6.03MB/s]

  0%|          | 0.00/5.15k [00:00<?, ?B/s]
100%|##########| 5.15k/5.15k [00:00<00:00, 59.8MB/s]
Training set has 60000 instances
Validation set has 10000 instances

As always, let’s visualize the data as a sanity check:

import matplotlib.pyplot as plt
import numpy as np

# Helper function for inline image display
def matplotlib_imshow(img, one_channel=False):
    if one_channel:
        img = img.mean(dim=0)
    img = img / 2 + 0.5     # unnormalize
    npimg = img.numpy()
    if one_channel:
        plt.imshow(npimg, cmap="Greys")
    else:
        plt.imshow(np.transpose(npimg, (1, 2, 0)))

dataiter = iter(training_loader)
images, labels = next(dataiter)

# Create a grid from the images and show them
img_grid = torchvision.utils.make_grid(images)
matplotlib_imshow(img_grid, one_channel=True)
print('  '.join(classes[labels[j]] for j in range(4)))

Shirt  Dress  Ankle Boot  Shirt

The Model

The model we’ll use in this example is a variant of LeNet-5 - it should be familiar if you’ve watched the previous videos in this series.

import torch.nn as nn
import torch.nn.functional as F

# PyTorch models inherit from torch.nn.Module
class GarmentClassifier(nn.Module):
    def __init__(self):
        super(GarmentClassifier, self).__init__()
        self.conv1 = nn.Conv2d(1, 6, 5)
        self.pool = nn.MaxPool2d(2, 2)
        self.conv2 = nn.Conv2d(6, 16, 5)
        self.fc1 = nn.Linear(16 * 4 * 4, 120)
        self.fc2 = nn.Linear(120, 84)
        self.fc3 = nn.Linear(84, 10)

    def forward(self, x):
        x = self.pool(F.relu(self.conv1(x)))
        x = self.pool(F.relu(self.conv2(x)))
        x = x.view(-1, 16 * 4 * 4)
        x = F.relu(self.fc1(x))
        x = F.relu(self.fc2(x))
        x = self.fc3(x)
        return x


model = GarmentClassifier()

Loss Function

For this example, we’ll be using a cross-entropy loss. For demonstration purposes, we’ll create batches of dummy output and label values, run them through the loss function, and examine the result.

loss_fn = torch.nn.CrossEntropyLoss()

# NB: Loss functions expect data in batches, so we're creating batches of 4
# Represents the model's confidence in each of the 10 classes for a given input
dummy_outputs = torch.rand(4, 10)
# Represents the correct class among the 10 being tested
dummy_labels = torch.tensor([1, 5, 3, 7])

print(dummy_outputs)
print(dummy_labels)

loss = loss_fn(dummy_outputs, dummy_labels)
print('Total loss for this batch: {}'.format(loss.item()))

tensor([[0.5710, 0.1400, 0.4561, 0.4383, 0.7203, 0.9709, 0.1245, 0.2202, 0.1940,
         0.8591],
        [0.1837, 0.2509, 0.4660, 0.1502, 0.5045, 0.7752, 0.3802, 0.8293, 0.8036,
         0.3418],
        [0.3109, 0.9613, 0.2721, 0.9507, 0.7536, 0.2035, 0.6705, 0.3518, 0.1788,
         0.2229],
        [0.8422, 0.2492, 0.3488, 0.6375, 0.1783, 0.4648, 0.4307, 0.0165, 0.8220,
         0.7794]])
tensor([1, 5, 3, 7])
Total loss for this batch: 2.3466196060180664

Optimizer

For this example, we’ll be using simple stochastic gradient descent with momentum.

It can be instructive to try some variations on this optimization scheme:

• Learning rate determines the size of the steps the optimizer takes. What does a different learning rate do to the your training results, in terms of accuracy and convergence time?

• Momentum nudges the optimizer in the direction of strongest gradient over multiple steps. What does changing this value do to your results?

• Try some different optimization algorithms, such as averaged SGD, Adagrad, or Adam. How do your results differ?

# Optimizers specified in the torch.optim package
optimizer = torch.optim.SGD(model.parameters(), lr=0.001, momentum=0.9)

The Training Loop

Below, we have a function that performs one training epoch. It enumerates data from the DataLoader, and on each pass of the loop does the following:

• Gets a batch of training data from the DataLoader

• Zeros the optimizer’s gradients

• Performs an inference - that is, gets predictions from the model for an input batch

• Calculates the loss for that set of predictions vs. the labels on the dataset

• Calculates the backward gradients over the learning weights

• Tells the optimizer to perform one learning step - that is, adjust the model’s learning weights based on the observed gradients for this batch, according to the optimization algorithm we chose

• It reports on the loss for every 1000 batches.

• Finally, it reports the average per-batch loss for the last 1000 batches, for comparison with a validation run

def train_one_epoch(epoch_index, tb_writer):
    running_loss = 0.
    last_loss = 0.

    # Here, we use enumerate(training_loader) instead of
    # iter(training_loader) so that we can track the batch
    # index and do some intra-epoch reporting
    for i, data in enumerate(training_loader):
        # Every data instance is an input + label pair
        inputs, labels = data

        # Zero your gradients for every batch!
        optimizer.zero_grad()

        # Make predictions for this batch
        outputs = model(inputs)

        # Compute the loss and its gradients
        loss = loss_fn(outputs, labels)
        loss.backward()

        # Adjust learning weights
        optimizer.step()

        # Gather data and report
        running_loss += loss.item()
        if i % 1000 == 999:
            last_loss = running_loss / 1000 # loss per batch
            print('  batch {} loss: {}'.format(i + 1, last_loss))
            tb_x = epoch_index * len(training_loader) + i + 1
            tb_writer.add_scalar('Loss/train', last_loss, tb_x)
            running_loss = 0.

    return last_loss

Per-Epoch Activity

There are a couple of things we’ll want to do once per epoch:

• Perform validation by checking our relative loss on a set of data that was not used for training, and report this

• Save a copy of the model

Here, we’ll do our reporting in TensorBoard. This will require going to the command line to start TensorBoard, and opening it in another browser tab.

# Initializing in a separate cell so we can easily add more epochs to the same run
timestamp = datetime.now().strftime('%Y%m%d_%H%M%S')
writer = SummaryWriter('runs/fashion_trainer_{}'.format(timestamp))
epoch_number = 0

EPOCHS = 5

best_vloss = 1_000_000.

for epoch in range(EPOCHS):
    print('EPOCH {}:'.format(epoch_number + 1))

    # Make sure gradient tracking is on, and do a pass over the data
    model.train(True)
    avg_loss = train_one_epoch(epoch_number, writer)


    running_vloss = 0.0
    # Set the model to evaluation mode, disabling dropout and using population
    # statistics for batch normalization.
    model.eval()

    # Disable gradient computation and reduce memory consumption.
    with torch.no_grad():
        for i, vdata in enumerate(validation_loader):
            vinputs, vlabels = vdata
            voutputs = model(vinputs)
            vloss = loss_fn(voutputs, vlabels)
            running_vloss += vloss

    avg_vloss = running_vloss / (i + 1)
    print('LOSS train {} valid {}'.format(avg_loss, avg_vloss))

    # Log the running loss averaged per batch
    # for both training and validation
    writer.add_scalars('Training vs. Validation Loss',
                    { 'Training' : avg_loss, 'Validation' : avg_vloss },
                    epoch_number + 1)
    writer.flush()

    # Track best performance, and save the model's state
    if avg_vloss < best_vloss:
        best_vloss = avg_vloss
        model_path = 'model_{}_{}'.format(timestamp, epoch_number)
        torch.save(model.state_dict(), model_path)

    epoch_number += 1

EPOCH 1:
  batch 1000 loss: 1.5757870875522495
  batch 2000 loss: 0.8396922070756555
  batch 3000 loss: 0.7189141086805612
  batch 4000 loss: 0.6479787551248446
  batch 5000 loss: 0.6114108842322603
  batch 6000 loss: 0.5684315731613897
  batch 7000 loss: 0.5725126664398704
  batch 8000 loss: 0.5466425174115574
  batch 9000 loss: 0.4993855416374281
  batch 10000 loss: 0.5120424808999523
  batch 11000 loss: 0.4802118537473725
  batch 12000 loss: 0.46094463047379397
  batch 13000 loss: 0.4579705506290775
  batch 14000 loss: 0.449209763904335
  batch 15000 loss: 0.4235263306122506
LOSS train 0.4235263306122506 valid 0.4486749470233917
EPOCH 2:
  batch 1000 loss: 0.4404510329887853
  batch 2000 loss: 0.413129741552053
  batch 3000 loss: 0.3768721194146783
  batch 4000 loss: 0.41575194504414686
  batch 5000 loss: 0.42212956145347563
  batch 6000 loss: 0.3848675963131245
  batch 7000 loss: 0.38230541721559713
  batch 8000 loss: 0.4011253063652548
  batch 9000 loss: 0.377365067936822
  batch 10000 loss: 0.3837543381729629
  batch 11000 loss: 0.3645839117608266
  batch 12000 loss: 0.36969418091571427
  batch 13000 loss: 0.33724303613166556
  batch 14000 loss: 0.36747662454267266
  batch 15000 loss: 0.35758256196024013
LOSS train 0.35758256196024013 valid 0.38773074746131897
EPOCH 3:
  batch 1000 loss: 0.3530504161407007
  batch 2000 loss: 0.3346566373779788
  batch 3000 loss: 0.33705213524679584
  batch 4000 loss: 0.326610338755534
  batch 5000 loss: 0.352336874223176
  batch 6000 loss: 0.335981599662744
  batch 7000 loss: 0.3614574867120391
  batch 8000 loss: 0.3193593590984237
  batch 9000 loss: 0.3327820297169965
  batch 10000 loss: 0.3399248921108956
  batch 11000 loss: 0.34410360914558985
  batch 12000 loss: 0.3260524760524859
  batch 13000 loss: 0.3367623088526598
  batch 14000 loss: 0.3351869106817048
  batch 15000 loss: 0.33320999304026916
LOSS train 0.33320999304026916 valid 0.3541768491268158
EPOCH 4:
  batch 1000 loss: 0.3191974891253049
  batch 2000 loss: 0.29087068734399507
  batch 3000 loss: 0.3225116919652701
  batch 4000 loss: 0.2973894258595683
  batch 5000 loss: 0.31286955276798106
  batch 6000 loss: 0.31318351908670594
  batch 7000 loss: 0.3065412062449468
  batch 8000 loss: 0.3186663959919242
  batch 9000 loss: 0.2922561830548293
  batch 10000 loss: 0.3187969190884833
  batch 11000 loss: 0.31806606633053164
  batch 12000 loss: 0.3081327496774611
  batch 13000 loss: 0.29683933951523794
  batch 14000 loss: 0.31757266068260653
  batch 15000 loss: 0.30277000801215764
LOSS train 0.30277000801215764 valid 0.3563423454761505
EPOCH 5:
  batch 1000 loss: 0.30627669369742216
  batch 2000 loss: 0.28426050725395996
  batch 3000 loss: 0.28992509504104963
  batch 4000 loss: 0.29514700796599935
  batch 5000 loss: 0.2796659715258538
  batch 6000 loss: 0.2972511586781584
  batch 7000 loss: 0.2808332733271491
  batch 8000 loss: 0.27828007936348514
  batch 9000 loss: 0.29243757135427584
  batch 10000 loss: 0.2916102179817426
  batch 11000 loss: 0.2931736901284894
  batch 12000 loss: 0.27542873114589017
  batch 13000 loss: 0.2842219123771174
  batch 14000 loss: 0.28533560235812183
  batch 15000 loss: 0.2806315785603383
LOSS train 0.2806315785603383 valid 0.3183196187019348

To load a saved version of the model:

saved_model = GarmentClassifier()
saved_model.load_state_dict(torch.load(PATH))

Once you’ve loaded the model, it’s ready for whatever you need it for - more training, inference, or analysis.

Note that if your model has constructor parameters that affect model structure, you’ll need to provide them and configure the model identically to the state in which it was saved.

Other Resources

• Docs on the data utilities, including Dataset and DataLoader, at pytorch.org

• A note on the use of pinned memory for GPU training

• Documentation on the datasets available in TorchVision, TorchText, and TorchAudio

• Documentation on the loss functions available in PyTorch

• Documentation on the torch.optim package, which includes optimizers and related tools, such as learning rate scheduling

• A detailed tutorial on saving and loading models

• The Tutorials section of pytorch.org contains tutorials on a broad variety of training tasks, including classification in different domains, generative adversarial networks, reinforcement learning, and more