참고
Click here to download the full example code
Introduction || Tensors || Autograd || Building Models || TensorBoard Support || Training Models || Model Understanding
Training with PyTorch¶
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)))
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)))

Sandal Sneaker Ankle Boot Coat
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([[0.3557, 0.8645, 0.5204, 0.7741, 0.1139, 0.4145, 0.7348, 0.6162, 0.7117,
0.4355],
[0.1968, 0.2940, 0.9526, 0.3006, 0.2524, 0.0563, 0.8280, 0.6981, 0.1534,
0.7943],
[0.8296, 0.4251, 0.3695, 0.4926, 0.9087, 0.2559, 0.7270, 0.5436, 0.4969,
0.0528],
[0.2731, 0.1459, 0.0121, 0.0764, 0.9622, 0.0705, 0.4019, 0.9218, 0.3255,
0.4819]])
tensor([1, 5, 3, 7])
Total loss for this batch: 2.2294564247131348
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)
# We don't need gradients on to do reporting
model.train(False)
running_vloss = 0.0
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: 1.7202785758674144
batch 2000 loss: 0.8457988122999668
batch 3000 loss: 0.7344358410686255
batch 4000 loss: 0.6348215425089002
batch 5000 loss: 0.6132729426613077
batch 6000 loss: 0.5689538897902239
batch 7000 loss: 0.535509346962208
batch 8000 loss: 0.5065724759609439
batch 9000 loss: 0.49358063461841084
batch 10000 loss: 0.5068647645610618
batch 11000 loss: 0.4623187284328742
batch 12000 loss: 0.4442366441761842
batch 13000 loss: 0.4400626162974513
batch 14000 loss: 0.45063220444787294
batch 15000 loss: 0.41373204464805896
LOSS train 0.41373204464805896 valid 0.4639061391353607
EPOCH 2:
batch 1000 loss: 0.4049252691374859
batch 2000 loss: 0.4027956905339379
batch 3000 loss: 0.3870164828733832
batch 4000 loss: 0.3765921371775912
batch 5000 loss: 0.3810157497630571
batch 6000 loss: 0.38515496355987855
batch 7000 loss: 0.3881014699799125
batch 8000 loss: 0.38654046844306866
batch 9000 loss: 0.35264253223035485
batch 10000 loss: 0.35956855410384014
batch 11000 loss: 0.3746511521441862
batch 12000 loss: 0.35857216321751184
batch 13000 loss: 0.3264845824319054
batch 14000 loss: 0.33541295046557207
batch 15000 loss: 0.3707398298132175
LOSS train 0.3707398298132175 valid 0.3628268837928772
EPOCH 3:
batch 1000 loss: 0.33687457587916286
batch 2000 loss: 0.33213954120439304
batch 3000 loss: 0.3441374519288947
batch 4000 loss: 0.33658539128087434
batch 5000 loss: 0.31185333648807134
batch 6000 loss: 0.3150397274837887
batch 7000 loss: 0.353223590657406
batch 8000 loss: 0.3398171684925983
batch 9000 loss: 0.32288358722522387
batch 10000 loss: 0.3145468523482559
batch 11000 loss: 0.3245599505953287
batch 12000 loss: 0.321184238298873
batch 13000 loss: 0.31208387652691455
batch 14000 loss: 0.30990570707632287
batch 15000 loss: 0.333060434007697
LOSS train 0.333060434007697 valid 0.3348868787288666
EPOCH 4:
batch 1000 loss: 0.302605270275053
batch 2000 loss: 0.30221588003780564
batch 3000 loss: 0.31341029232504664
batch 4000 loss: 0.30225226803854455
batch 5000 loss: 0.2959274389560742
batch 6000 loss: 0.3017923433062388
batch 7000 loss: 0.3103919526273894
batch 8000 loss: 0.3142617614178744
batch 9000 loss: 0.2908107597097842
batch 10000 loss: 0.30662362715498237
batch 11000 loss: 0.2936881104596905
batch 12000 loss: 0.29864081332642306
batch 13000 loss: 0.2882055617513561
batch 14000 loss: 0.2811639510149398
batch 15000 loss: 0.29043961816684893
LOSS train 0.29043961816684893 valid 0.31826725602149963
EPOCH 5:
batch 1000 loss: 0.2909442294574983
batch 2000 loss: 0.27795928239824935
batch 3000 loss: 0.28197013997076553
batch 4000 loss: 0.27602926642545994
batch 5000 loss: 0.27291372970746125
batch 6000 loss: 0.26500764907761276
batch 7000 loss: 0.2763223950124077
batch 8000 loss: 0.28239086670237523
batch 9000 loss: 0.2731176992859073
batch 10000 loss: 0.2828177599106548
batch 11000 loss: 0.28514709650713665
batch 12000 loss: 0.27786594535020415
batch 13000 loss: 0.27871137002642354
batch 14000 loss: 0.29793933952681256
batch 15000 loss: 0.28550324632457885
LOSS train 0.28550324632457885 valid 0.3166631758213043
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
Total running time of the script: ( 5 minutes 4.804 seconds)