%matplotlib inlineAn adaptation of training a classifier tutorial using Habana Gaudi AI processors.
In this tutorial, we will train a Convolutional Neural Network (CNN) to classify images. We will start with CPU training, and then continue to make the needed changes for training the model on Gaudi HPU.
Handling data
Generally, when you have to deal with image, text, audio or video data, you can use standard python packages that load data into a numpy array. Then you can convert this array into a torch.*Tensor.
- For images, packages such as Pillow, OpenCV are useful
- For audio, packages such as scipy and librosa
- For text, either raw Python or Cython based loading, or NLTK and SpaCy are useful
Specifically for vision, we have created a package called torchvision, that has data loaders for common datasets such as Imagenet, CIFAR10, MNIST, etc. and data transformers for images, viz., torchvision.datasets and torch.utils.data.DataLoader.
This provides a huge convenience and avoids writing boilerplate code.
For this tutorial, we will use the CIFAR10 dataset. It has the classes: ‘airplane’, ‘automobile’, ‘bird’, ‘cat’, ‘deer’, ‘dog’, ‘frog’, ‘horse’, ‘ship’, ‘truck’. The images in CIFAR-10 are of size 3x32x32, i.e. 3-channel color images of 32×32 pixels in size.
cifar10
Training an image classifier
We will do the following steps in order:
- Load and normalize the CIFAR10 training and test datasets using
torchvision - Define a Convolutional Neural Network
- Define a loss function
- Train the network on the training data
- Test the network on the test data
The first step:
- Load and normalize CIFAR10
Using torchvision, it’s extremely easy to load CIFAR10.
import torch
import torchvision
import torchvision.transforms as transformsThe output of torchvision datasets are PILImage images of range [0, 1]. We transform them to Tensors of normalized range [-1, 1].
transform = transforms.Compose(
[transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])
batch_size = 64
trainset = torchvision.datasets.CIFAR10(root='./data', train=True,
download=True, transform=transform)
trainloader = torch.utils.data.DataLoader(trainset, batch_size=batch_size,
shuffle=True, num_workers=2)
testset = torchvision.datasets.CIFAR10(root='./data', train=False,
download=True, transform=transform)
testloader = torch.utils.data.DataLoader(testset, batch_size=batch_size,
shuffle=False, num_workers=2)
classes = ('plane', 'car', 'bird', 'cat',
'deer', 'dog', 'frog', 'horse', 'ship', 'truck')Files already downloaded and verified
Files already downloaded and verifiedLet us show some of the training images, for fun.
import matplotlib.pyplot as plt
import numpy as np
# functions to show an image
def imshow(img):
img = img / 2 + 0.5 # unnormalize
npimg = img.numpy()
plt.imshow(np.transpose(npimg, (1, 2, 0)))
plt.show()
# get some random training images
dataiter = iter(trainloader)
images, labels = dataiter.next()
# show images
imshow(torchvision.utils.make_grid(images))
# print labels
print(' '.join('%5s' % classes[labels[j]] for j in range(batch_size)))
plane plane dog plane cat deer bird deer bird cat ship plane bird dog car dog frog bird dog ship deer plane cat dog truck bird bird truck ship horse deer frog deer car plane frog truck car frog truck deer dog cat truck car plane dog truck car plane ship deer dog bird deer horse plane horse bird truck deer plane ship truck- Define a Convolutional Neural Network
Copy the neural network from the Neural Networks section before and modify it to take 3-channel images (instead of 1-channel images as it was defined).
import torch.nn as nn
import torch.nn.functional as F
class Net(nn.Module):
def __init__(self):
super().__init__()
self.conv1 = nn.Conv2d(3, 6, 5)
self.pool = nn.MaxPool2d(2, 2)
self.conv2 = nn.Conv2d(6, 16, 5)
self.fc1 = nn.Linear(16 * 5 * 5, 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 = torch.flatten(x, 1) # flatten all dimensions except batch
x = F.relu(self.fc1(x))
x = F.relu(self.fc2(x))
x = self.fc3(x)
return x
net = Net()- Define a Loss function and optimizer
Let’s use a Classification Cross-Entropy loss and SGD with momentum.
import torch.optim as optim
criterion = nn.CrossEntropyLoss()
optimizer = optim.SGD(net.parameters(), lr=0.001, momentum=0.9)- Train the network
This is when things start to get interesting. We simply have to loop over our data iterator, and feed the inputs to the network and optimize.
for epoch in range(2): # loop over the dataset multiple times
running_loss = 0.0
for i, data in enumerate(trainloader, 0):
# get the inputs; data is a list of [inputs, labels]
inputs, labels = data
# zero the parameter gradients
optimizer.zero_grad()
# forward + backward + optimize
outputs = net(inputs)
loss = criterion(outputs, labels)
loss.backward()
optimizer.step()
# print statistics
running_loss += loss.item()
if i % 200 == 199: # print every 200 mini-batches
print('[%d, %5d] loss: %.3f' %
(epoch + 1, i + 1, running_loss / 200))
running_loss = 0.0
print('Finished Training')[1, 200] loss: 2.303
[1, 400] loss: 2.300
[1, 600] loss: 2.297
[2, 200] loss: 2.282
[2, 400] loss: 2.259
[2, 600] loss: 2.177
Finished TrainingCode language: CSS (css)Let’s quickly save our trained model:
PATH = './cifar_net.pth'
torch.save(net.state_dict(), PATH)See here _ for more details on saving PyTorch models.
- Test the network on the test data
We have trained the network for 2 passes over the training dataset. But we need to check if the network has learnt anything at all.
We will check this by predicting the class label that the neural network outputs, and checking it against the ground-truth. If the prediction is correct, we add the sample to the list of correct predictions.
Okay, first step. Let us display an image from the test set to get familiar.
dataiter = iter(testloader)
images, labels = dataiter.next()
# print images
imshow(torchvision.utils.make_grid(images))
print('GroundTruth: ', ' '.join('%5s' % classes[labels[j]] for j in range(4)))
GroundTruth: cat ship ship planeCode language: HTTP (http)Next, let’s load back in our saved model (note: saving and re-loading the model wasn’t necessary here, we only did it to illustrate how to do so):
net = Net()
net.load_state_dict(torch.load(PATH))Okay, now let us see what the neural network thinks these examples above are:
outputs = net(images)The outputs are energies for the 10 classes. The higher the energy for a class, the more the network thinks that the image is of the particular class. So, let’s get the index of the highest energy:
_, predicted = torch.max(outputs, 1)
print('Predicted: ', ' '.join('%5s' % classes[predicted[j]]
for j in range(4)))Predicted: cat ship ship shipCode language: HTTP (http)The results seem pretty good.
Let us look at how the network performs on the whole dataset.
correct = 0
total = 0
# since we're not training, we don't need to calculate the gradients for our outputs
with torch.no_grad():
for data in testloader:
images, labels = data
# calculate outputs by running images through the network
outputs = net(images)
# the class with the highest energy is what we choose as prediction
_, predicted = torch.max(outputs.data, 1)
total += labels.size(0)
correct += (predicted == labels).sum().item()
print('Accuracy of the network on the 10000 test images: %d %%' % (
100 * correct / total))Accuracy of the network on the 10000 test images: 31 %That looks way better than chance, which is 10% accuracy (randomly picking a class out of 10 classes). Seems like the network learnt something.
Hmmm, what are the classes that performed well, and the classes that did not perform well:
# prepare to count predictions for each class
correct_pred = {classname: 0 for classname in classes}
total_pred = {classname: 0 for classname in classes}
# again no gradients needed
with torch.no_grad():
for data in testloader:
images, labels = data
outputs = net(images)
_, predictions = torch.max(outputs, 1)
# collect the correct predictions for each class
for label, prediction in zip(labels, predictions):
if label == prediction:
correct_pred[classes[label]] += 1
total_pred[classes[label]] += 1
# print accuracy for each class
for classname, correct_count in correct_pred.items():
accuracy = 100 * float(correct_count) / total_pred[classname]
print("Accuracy for class {:5s} is: {:.1f} %".format(classname,
accuracy))Accuracy for class plane is: 52.6 %
Accuracy for class car is: 27.6 %
Accuracy for class bird is: 1.8 %
Accuracy for class cat is: 18.3 %
Accuracy for class deer is: 2.4 %
Accuracy for class dog is: 10.2 %
Accuracy for class frog is: 50.2 %
Accuracy for class horse is: 40.5 %
Accuracy for class ship is: 38.0 %
Accuracy for class truck is: 40.6 %Code language: CSS (css)Okay, so what next?
How do we run these neural networks on the Habana Processing Unit (HPU, i.e. Gaudi)?
Training on HPU
First let’s we load the necessary libraries for the model and dataloader, just like on CPU:
NOTE: You must restart the Kernel to execute this on Gaudi HPU
import torch
import torchvision
import torchvision.transforms as transforms
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
class Net(nn.Module):
def __init__(self):
super().__init__()
self.conv1 = nn.Conv2d(3, 6, 5)
self.pool = nn.MaxPool2d(2, 2)
self.conv2 = nn.Conv2d(6, 16, 5)
self.fc1 = nn.Linear(16 * 5 * 5, 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 = torch.flatten(x, 1) # flatten all dimensions except batch
x = F.relu(self.fc1(x))
x = F.relu(self.fc2(x))
x = self.fc3(x)
return x
transform = transforms.Compose(
[transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])
batch_size = 64
trainset = torchvision.datasets.CIFAR10(root='./data', train=True,
download=True, transform=transform)
trainloader = torch.utils.data.DataLoader(trainset, batch_size=batch_size,
shuffle=True)
testset = torchvision.datasets.CIFAR10(root='./data', train=False,
download=True, transform=transform)
testloader = torch.utils.data.DataLoader(testset, batch_size=batch_size,
shuffle=False)Files already downloaded and verified
Files already downloaded and verifiedJust like how you transfer a Tensor onto the HPU, you transfer the neural net onto the HPU.
Let’s load the necessary libraries to load the Habana module and define our device as the hpu device:
import habana_frameworks.torch.core as htcore
device = torch.device("hpu")
# Assuming that we are on a HPU machine, this should print a HPU device:
print(device)hpuAnd then transfer the model onto the HPU
model = Net().to(device)# define the optimzer and permute parameters
criterion = nn.CrossEntropyLoss()
optimizer = optim.SGD(model.parameters(), lr=0.001, momentum=0.9)Remember that you will have to send the inputs and targets at every step to the HPU too:
inputs, labels = inputs.to(device), labels.to(device)
Below code shows how to handle the inputs and targets during the training on HPU.
for epoch in range(2): # loop over the dataset multiple times
model.train()
running_loss = 0.0
for i, data in enumerate(trainloader, 0):
# get the inputs; data is a list of [inputs, labels]
inputs, labels = data
# transfer input data to HPU device
inputs, labels = inputs.to(device), labels.to(device)
# zero the parameter gradients
optimizer.zero_grad()
# forward + backward + optimize
outputs = model(inputs)
loss = criterion(outputs, labels)
loss.backward()
# API call to trigger execution
htcore.mark_step()
optimizer.step()
# API call to trigger execution
htcore.mark_step()
# print statistics
running_loss += loss.item()
if i % 200 == 199: # print every 200 steps
print('[%d, %5d] loss: %.3f' %
(epoch + 1, i + 1, running_loss / 200))
running_loss = 0.0
print('Finished Training')[1, 200] loss: 2.303
[1, 400] loss: 2.300
[1, 600] loss: 2.297
[2, 200] loss: 2.276
[2, 400] loss: 2.213
[2, 600] loss: 2.096
Finished TrainingCode language: CSS (css)BSD 3-Clause License
Copyright (c) 2021 Habana Labs, Ltd. an Intel Company.
Copyright (c) 2017, Pytorch contributors All rights reserved.
Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
- Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
- Neither the name of the copyright holder nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS “AS IS” AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
