beginner/introyt/trainingyt

Run in Google Colab

Colab

Download Notebook

Notebook

View on GitHub

GitHub

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

Training with PyTorch

Created On: Nov 30, 2021 | Last Updated: May 06, 2026 | 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.v2.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
from torchvision.transforms import v2

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


transform = v2.Compose([
    v2.ToImage(),
    v2.ToDtype(torch.float32, scale=True),
    v2.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(f'Training set has {len(training_set)} instances')
print(f'Validation set has {len(validation_set)} instances')

  0%|          | 0.00/26.4M [00:00<?, ?B/s]
  0%|          | 65.5k/26.4M [00:00<01:09, 377kB/s]
  1%|          | 197k/26.4M [00:00<00:44, 595kB/s]
  3%|▎         | 754k/26.4M [00:00<00:14, 1.78MB/s]
 11%|█         | 2.95M/26.4M [00:00<00:03, 6.03MB/s]
 30%|███       | 8.03M/26.4M [00:00<00:01, 14.3MB/s]
 52%|█████▏    | 13.9M/26.4M [00:01<00:00, 20.7MB/s]
 75%|███████▌  | 19.9M/26.4M [00:01<00:00, 29.4MB/s]
 88%|████████▊ | 23.3M/26.4M [00:01<00:00, 26.4MB/s]
100%|██████████| 26.4M/26.4M [00:01<00:00, 18.8MB/s]

  0%|          | 0.00/29.5k [00:00<?, ?B/s]
100%|██████████| 29.5k/29.5k [00:00<00:00, 342kB/s]

  0%|          | 0.00/4.42M [00:00<?, ?B/s]
  1%|▏         | 65.5k/4.42M [00:00<00:11, 377kB/s]
  5%|▌         | 229k/4.42M [00:00<00:05, 708kB/s]
 21%|██        | 918k/4.42M [00:00<00:01, 2.19MB/s]
 82%|████████▏ | 3.60M/4.42M [00:00<00:00, 7.41MB/s]
100%|██████████| 4.42M/4.42M [00:00<00:00, 6.33MB/s]

  0%|          | 0.00/5.15k [00:00<?, ?B/s]
100%|██████████| 5.15k/5.15k [00:00<00:00, 30.7MB/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)))

Coat  Coat  Sneaker  Pullover

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().__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(f'Total loss for this batch: {loss.item()}')

tensor([[0.1818, 0.7835, 0.0766, 0.8312, 0.8500, 0.9796, 0.4417, 0.0902, 0.0348,
         0.0058],
        [0.4393, 0.7969, 0.9903, 0.8091, 0.6351, 0.9124, 0.9064, 0.9872, 0.4067,
         0.9831],
        [0.3131, 0.2167, 0.2575, 0.9653, 0.3791, 0.3318, 0.4625, 0.5323, 0.8916,
         0.5218],
        [0.4826, 0.9142, 0.8760, 0.1572, 0.9682, 0.8461, 0.5260, 0.7244, 0.0855,
         0.7132]])
tensor([1, 5, 3, 7])
Total loss for this batch: 2.0796797275543213

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(f'  batch {i + 1} loss: {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(f'runs/fashion_trainer_{timestamp}')
epoch_number = 0

EPOCHS = 5

best_vloss = 1_000_000.

for epoch in range(EPOCHS):
    print(f'EPOCH {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(f'LOSS train {avg_loss} valid {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 = f'model_{timestamp}_{epoch_number}'
        torch.save(model.state_dict(), model_path)

    epoch_number += 1

EPOCH 1:
  batch 1000 loss: 1.7405721426457166
  batch 2000 loss: 0.8306381810922175
  batch 3000 loss: 0.677536542580463
  batch 4000 loss: 0.6139906443548389
  batch 5000 loss: 0.6232024310473353
  batch 6000 loss: 0.5409166301020887
  batch 7000 loss: 0.5292866276893765
  batch 8000 loss: 0.5087711073538521
  batch 9000 loss: 0.4886866149562411
  batch 10000 loss: 0.46343287967721697
  batch 11000 loss: 0.43794697365624596
  batch 12000 loss: 0.4498035543341248
  batch 13000 loss: 0.4252461740033468
  batch 14000 loss: 0.43774853999976765
  batch 15000 loss: 0.4059081710058381
LOSS train 0.4059081710058381 valid 0.4468461573123932
EPOCH 2:
  batch 1000 loss: 0.3947927801271326
  batch 2000 loss: 0.3855097612418758
  batch 3000 loss: 0.3798059433118324
  batch 4000 loss: 0.37429591737902956
  batch 5000 loss: 0.37816341016746446
  batch 6000 loss: 0.37183002680432403
  batch 7000 loss: 0.3662146015749604
  batch 8000 loss: 0.3596699382332736
  batch 9000 loss: 0.382679241851205
  batch 10000 loss: 0.36675233062927143
  batch 11000 loss: 0.33918615903297905
  batch 12000 loss: 0.34860096452757716
  batch 13000 loss: 0.3292305163000128
  batch 14000 loss: 0.3388547223967616
  batch 15000 loss: 0.3672620100512358
LOSS train 0.3672620100512358 valid 0.3514325022697449
EPOCH 3:
  batch 1000 loss: 0.32357315335000747
  batch 2000 loss: 0.3235344601000397
  batch 3000 loss: 0.31539368888552416
  batch 4000 loss: 0.33136245063957903
  batch 5000 loss: 0.3268227448241232
  batch 6000 loss: 0.33314924244757277
  batch 7000 loss: 0.31053505857023994
  batch 8000 loss: 0.3215692091492965
  batch 9000 loss: 0.322625958440447
  batch 10000 loss: 0.30893365424661895
  batch 11000 loss: 0.3175491015645384
  batch 12000 loss: 0.3114129657415324
  batch 13000 loss: 0.30487305273596577
  batch 14000 loss: 0.3215428635533026
  batch 15000 loss: 0.3177673206521431
LOSS train 0.3177673206521431 valid 0.32869815826416016
EPOCH 4:
  batch 1000 loss: 0.2934176405906328
  batch 2000 loss: 0.3012557532303035
  batch 3000 loss: 0.2989550880537281
  batch 4000 loss: 0.29748138841960464
  batch 5000 loss: 0.3155643715653532
  batch 6000 loss: 0.2795640707507773
  batch 7000 loss: 0.28220126770006754
  batch 8000 loss: 0.30219886600592144
  batch 9000 loss: 0.287171704797307
  batch 10000 loss: 0.305691383659112
  batch 11000 loss: 0.27300324711605933
  batch 12000 loss: 0.28123975936553325
  batch 13000 loss: 0.30210795575525845
  batch 14000 loss: 0.2928324456654955
  batch 15000 loss: 0.2879913049916649
LOSS train 0.2879913049916649 valid 0.32275378704071045
EPOCH 5:
  batch 1000 loss: 0.28040234959562077
  batch 2000 loss: 0.2868007918667281
  batch 3000 loss: 0.26894103905418887
  batch 4000 loss: 0.24409383708676613
  batch 5000 loss: 0.26315568151612206
  batch 6000 loss: 0.26177643729170813
  batch 7000 loss: 0.28952401467310845
  batch 8000 loss: 0.28789225654256734
  batch 9000 loss: 0.2900867059714146
  batch 10000 loss: 0.28560011239088634
  batch 11000 loss: 0.28498311422584083
  batch 12000 loss: 0.2701355312746473
  batch 13000 loss: 0.2682444914646812
  batch 14000 loss: 0.2745767411998022
  batch 15000 loss: 0.27734026560361236
LOSS train 0.27734026560361236 valid 0.30079397559165955

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