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, 365kB/s]
  1%|          | 164k/26.4M [00:00<00:44, 588kB/s]
  1%|1         | 393k/26.4M [00:00<00:25, 1.02MB/s]
  5%|5         | 1.38M/26.4M [00:00<00:07, 3.51MB/s]
 12%|#1        | 3.15M/26.4M [00:00<00:03, 6.66MB/s]
 33%|###2      | 8.62M/26.4M [00:00<00:00, 18.4MB/s]
 46%|####6     | 12.2M/26.4M [00:00<00:00, 20.3MB/s]
 67%|######7   | 17.8M/26.4M [00:01<00:00, 27.7MB/s]
 82%|########1 | 21.6M/26.4M [00:01<00:00, 26.9MB/s]
100%|##########| 26.4M/26.4M [00:01<00:00, 19.4MB/s]

  0%|          | 0.00/29.5k [00:00<?, ?B/s]
100%|##########| 29.5k/29.5k [00:00<00:00, 328kB/s]

  0%|          | 0.00/4.42M [00:00<?, ?B/s]
  1%|1         | 65.5k/4.42M [00:00<00:11, 364kB/s]
  5%|5         | 229k/4.42M [00:00<00:06, 683kB/s]
 21%|##1       | 950k/4.42M [00:00<00:01, 2.19MB/s]
 87%|########6 | 3.83M/4.42M [00:00<00:00, 7.63MB/s]
100%|##########| 4.42M/4.42M [00:00<00:00, 6.11MB/s]

  0%|          | 0.00/5.15k [00:00<?, ?B/s]
100%|##########| 5.15k/5.15k [00:00<00:00, 62.2MB/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)))

Pullover  Sneaker  Bag  Bag

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([[0.8937, 0.2081, 0.5468, 0.6872, 0.1564, 0.7713, 0.9198, 0.8174, 0.9875,
         0.7793],
        [0.8907, 0.2391, 0.0996, 0.3334, 0.7295, 0.1826, 0.6372, 0.3656, 0.3169,
         0.7942],
        [0.3086, 0.4954, 0.9744, 0.7601, 0.1210, 0.1600, 0.7295, 0.6759, 0.7711,
         0.6051],
        [0.6827, 0.7837, 0.8372, 0.6836, 0.0472, 0.5039, 0.7699, 0.7211, 0.6981,
         0.7794]])
tensor([1, 5, 3, 7])
Total loss for this batch: 2.452549934387207

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: 1.6508730027526617
  batch 2000 loss: 0.8324885909864679
  batch 3000 loss: 0.7022715212288313
  batch 4000 loss: 0.638390345744323
  batch 5000 loss: 0.5814984522305895
  batch 6000 loss: 0.5703758230829844
  batch 7000 loss: 0.5475103505142033
  batch 8000 loss: 0.5157037968488876
  batch 9000 loss: 0.5172829415169544
  batch 10000 loss: 0.4963044887408614
  batch 11000 loss: 0.4512106528194854
  batch 12000 loss: 0.4672637660060427
  batch 13000 loss: 0.44132990965945645
  batch 14000 loss: 0.4318602813569596
  batch 15000 loss: 0.4501161056239507
LOSS train 0.4501161056239507 valid 0.4366302192211151
EPOCH 2:
  batch 1000 loss: 0.4189814815496211
  batch 2000 loss: 0.40109413378823955
  batch 3000 loss: 0.39180993908757955
  batch 4000 loss: 0.39893185182035085
  batch 5000 loss: 0.38355205544509224
  batch 6000 loss: 0.37717836963915036
  batch 7000 loss: 0.36153815490772834
  batch 8000 loss: 0.3748704566319357
  batch 9000 loss: 0.39526283496156245
  batch 10000 loss: 0.38645872144922033
  batch 11000 loss: 0.3535750466412355
  batch 12000 loss: 0.36304767112922853
  batch 13000 loss: 0.3957259635035589
  batch 14000 loss: 0.3652188332959195
  batch 15000 loss: 0.3550606707877596
LOSS train 0.3550606707877596 valid 0.36149492859840393
EPOCH 3:
  batch 1000 loss: 0.334679770553601
  batch 2000 loss: 0.34545868256222456
  batch 3000 loss: 0.34140541403074165
  batch 4000 loss: 0.3089789447458097
  batch 5000 loss: 0.298582465716092
  batch 6000 loss: 0.34347604986627994
  batch 7000 loss: 0.35283629154079243
  batch 8000 loss: 0.33562464751614607
  batch 9000 loss: 0.3123645003751008
  batch 10000 loss: 0.3371345452560345
  batch 11000 loss: 0.3230560367609287
  batch 12000 loss: 0.3348205538664188
  batch 13000 loss: 0.326315151397459
  batch 14000 loss: 0.3342794756471412
  batch 15000 loss: 0.339980882383752
LOSS train 0.339980882383752 valid 0.3580446243286133
EPOCH 4:
  batch 1000 loss: 0.30579127714385684
  batch 2000 loss: 0.31383266968592943
  batch 3000 loss: 0.28681159449659754
  batch 4000 loss: 0.31314202655950796
  batch 5000 loss: 0.3180921297867753
  batch 6000 loss: 0.3043288108412526
  batch 7000 loss: 0.276163292941972
  batch 8000 loss: 0.30510994749089876
  batch 9000 loss: 0.30453672522089253
  batch 10000 loss: 0.31879224362390235
  batch 11000 loss: 0.31586221162309086
  batch 12000 loss: 0.30726216625855884
  batch 13000 loss: 0.30239222428163703
  batch 14000 loss: 0.2848292864732066
  batch 15000 loss: 0.30757958309168315
LOSS train 0.30757958309168315 valid 0.3482609689235687
EPOCH 5:
  batch 1000 loss: 0.28576952889071255
  batch 2000 loss: 0.27546719777878026
  batch 3000 loss: 0.2735240871296846
  batch 4000 loss: 0.2787255609527556
  batch 5000 loss: 0.2691001614659908
  batch 6000 loss: 0.2964523515957699
  batch 7000 loss: 0.2786440310634134
  batch 8000 loss: 0.30058453428160103
  batch 9000 loss: 0.2833032690953114
  batch 10000 loss: 0.3011711950523386
  batch 11000 loss: 0.28784837749820874
  batch 12000 loss: 0.2778319435380881
  batch 13000 loss: 0.2797238845917891
  batch 14000 loss: 0.2931309028868436
  batch 15000 loss: 0.29013465171492137
LOSS train 0.29013465171492137 valid 0.33448147773742676

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