PyTorch Tutorial – Implementing Deep Neural Networks Using PyTorch

PyTorch Tutorial:

Let’s start this PyTorch Tutorial blog by establishing a fact that Deep Learning is something that is being used by everyone today, ranging from Virtual Assistance to getting recommendations while shopping! With newer tools emerging to make better use of Deep Learning, programming and implementation have become easier.

This PyTorch Tutorial will give you a complete insight into PyTorch in the following sequence:

• What is PyTorch
• Features of PyTorch
• Installing PyTorch
• The NumPy Bridge
• PyTorch: AutoGrad Module
• Use Case: Image Classifier

Deep Learning Frameworks In Python

Python is preferred for coding and working with Deep Learning and hence has a wide spectrum of frameworks to choose from. Such as:

• • TensorFlow
  • PyTorch
  • Keras
  • Theano
  • Lasagne

What Is PyTorch?

It’s a Python based scientific computing package targeted at two sets of audiences:

• A replacement for NumPy to make use of the power of GPUs.
• Deep Learning research platform that provides maximum flexibility and speed.
  

Features of PyTorch – Highlights

1. 1. Native support for Python and use of its libraries
      
   2. Actively used in the development of Facebook for all of it’s Deep Learning requirements in the platform.
   3. PyTorch ensures an easy to use API which helps with easier usability and better understanding when making use of the API.
   4. Dynamic Computation Graphs are a major highlight here as they ensure the graph build-up dynamically – at every point of code execution, the graph is built along and can be manipulated at run-time.
   5. PyTorch is fast and feels native, hence ensuring easy coding and fast processing.
   6. The support for CUDA ensures that the code can run on the GPU, thereby decreasing the time needed to run the code and increasing the overall performance of the system.

Installing PyTorch

Moving ahead in this PyTorch Tutorial, let’s see how simple it is to actually install PyTorch on your machine.

It’s pretty straight-forward based on the system properties such as the Operating System or the package managers. It can be installed from the Command Prompt or within an IDE such as PyCharm etc.

Next up on this PyTorch Tutorial blog, let us check out how NumPy is integrated into PyTorch.

Tensors

Tensors are similar to NumPy’s n dimensional arrays, with the addition being that Tensors can also be used on a GPU to accelerate computing.

Let’s construct a simple tensor and check the output. First let’s check out on how we can construct a 5×3 matrix which is uninitiated:

x = torch.empty(5, 3)
print(x)

Output: 

tensor([[8.3665e+22, 4.5580e-41, 1.6025e-03],
        [3.0763e-41, 0.0000e+00, 0.0000e+00],
        [0.0000e+00, 0.0000e+00, 3.4438e-41],
        [0.0000e+00, 4.8901e-36, 2.8026e-45],
        [6.6121e+31, 0.0000e+00, 9.1084e-44]])

Now let’s construct a randomly initialized matrix:

    x = torch.rand(5, 3)
    print(x)

Output:

tensor([[0.1607, 0.0298, 0.7555],
        [0.8887, 0.1625, 0.6643],
        [0.7328, 0.5419, 0.6686],
        [0.0793, 0.1133, 0.5956],
        [0.3149, 0.9995, 0.6372]])

Construct a tensor directly from data:

x = torch.tensor([5.5, 3])
print(x)

Output:

tensor([5.5000, 3.0000])

y = torch.rand(5, 3)
print(x + y)

    tensor([[ 0.2349, -0.0427, -0.5053],
            [ 0.6455,  0.1199,  0.4239],
            [ 0.1279,  0.1105,  1.4637],
            [ 0.4259, -0.0763, -0.9671],
            [ 0.6856,  0.5047,  0.4250]])

x = torch.randn(4, 4)
y = x.view(16)
z = x.view(-1, 8) # the size -1 is inferred from other dimensions
print(x.size(), y.size(), z.size())

torch.Size([4, 4]) torch.Size([16]) torch.Size([2, 8])

NumPy For PyTorch

NumPy is a library for the Python programming language, adding support for large, multi-dimensional arrays and matrices, along with a large collection of high-level mathematical functions to operate on these arrays.

It is also used as:

• Library providing tools for integrating C/C++ and FORTRAN code.
• Operations with linear algebra, Fourier transforms and random number capabilities.

Besides its obvious scientific uses, NumPy can also be used as an efficient multi-dimensional container of generic data and arbitrary data-types can be defined as well.

This allows NumPy to seamlessly and speedily integrate with a wide variety of databases!

The NumPy Bridge – Arrays And Tensors

Converting a Torch Tensor to a NumPy array and vice versa is a breeze!

The Torch Tensor and NumPy array will share their underlying memory locations and changing one will change the other.

Converting A Torch Tensor To A NumPy Array:

a = torch.ones(5)
print(a)

Output: tensor([1., 1., 1., 1., 1.])

b = a.numpy()
print(b)

Output: [1. 1. 1. 1. 1.]

Let’s perform a sum operation and check the changes in the values:

a.add_(1)
print(a)
print(b)

Output: tensor([2., 2., 2., 2., 2.]) [2. 2. 2. 2. 2.]

Converting A NumPy Array To A Torch Tensor:

import numpy as no
a = np.ones(5)
b = torch.from_numpy(a)
np.add(a, 1, out=a)
print(a)
print(b)

Output:

[2. 2. 2. 2. 2.]
tensor([2., 2., 2., 2., 2.], dtype=torch.float64)

So, as you can see, it is as simple as that!

Next up on this PyTorch Tutorial blog, let’s check out the AutoGrad module of PyTorch.

PyTorch: AutoGrad Module

The autograd package provides automatic differentiation for all operations on Tensors.

It is a define-by-run framework, which means that your backprop is defined by how your code is run, and that every single iteration can be different.

Next up on this PyTorch Tutorial Blog, let’s look an interesting and a simple use case.

PyTorch Use Case: Training an Image Classifier

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 and OpenCV are useful.
• For audio, packages such as Scipy and Librosa.
• For text, either raw Python, Cython based loading or NLTK and SpaCy are useful.

Specifically for vision, there is a package called torchvision, that has data loaders for common datasets such as Imagenet, CIFAR10, MNIST, etc. and data transformers for images.

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 as shown below:

PyTorch: Training The CIFAR10 Classifier

We will do the following steps in order:

1. 1. Loading And Normalizing CIFAR10
      
   2. Define a Convolution Neural Network
      
   3. Define a loss function
   4. Train the network on the training data
   5. Test the network on the test data

1. Loading And Normalizing CIFAR10

Using torchvision, it is very easy to load CIFAR10!

It is as simple as follows:

import torch
import torchvision
import torchvision.transforms as transforms

The 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))])
trainset = torchvision.datasets.CIFAR10(root='./data', train=True,
download=True, transform=transform)
trainloader = torch.utils.data.DataLoader(trainset, batch_size=4,
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=4,
shuffle=False, num_workers=2)
classes = ('plane', 'car', 'bird', 'cat',
'deer', 'dog', 'frog', 'horse', 'ship', 'truck')

Output:

Downloading https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz to ./data/cifar-10-python.tar.gz Files already downloaded and verified 

Next, let us print some training images from the dataset!

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)))


# 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(4)))

Output:

dog  bird horse horse

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


class Net(nn.Module):
def __init__(self):
super(Net, self).__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 = x.view(-1, 16 * 5 * 5)
x = F.relu(self.fc1(x))
x = F.relu(self.fc2(x))
x = self.fc3(x)
return x


net = Net()

import torch.optim as optim

criterion = nn.CrossEntropyLoss()
optimizer = optim.SGD(net.parameters(), lr=0.001, momentum=0.9)

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
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 % 2000 == 1999: # print every 2000 mini-batches
print('[%d, %5d] loss: %.3f' %
(epoch + 1, i + 1, running_loss / 2000))
running_loss = 0.0

print('Finished Training')

[1,  2000] loss: 2.236
[1,  4000] loss: 1.880
[1,  6000] loss: 1.676
[1,  8000] loss: 1.586
[1, 10000] loss: 1.515
[1, 12000] loss: 1.464
[2,  2000] loss: 1.410
[2,  4000] loss: 1.360
[2,  6000] loss: 1.360
[2,  8000] loss: 1.325
[2, 10000] loss: 1.312
[2, 12000] loss: 1.302
Finished Training

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 plane

outputs = net(images)

predicted = torch.max(outputs, 1)

print('Predicted: ', ' '.join('%5s' % classes[predicted[j]]
for j in range(4)))

Predicted:    cat   car   car plane

correct = 0
total = 0
with torch.no_grad():
for data in testloader:
images, labels = data
outputs = net(images)
_, 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: 54 %

class_correct = list(0. for i in range(10))
class_total = list(0. for i in range(10))
with torch.no_grad():
for data in testloader:
images, labels = data
outputs = net(images)
_, predicted = torch.max(outputs, 1)
c = (predicted == labels).squeeze()
for i in range(4):
label = labels[i]
class_correct[label] += c[i].item()
class_total[label] += 1


for i in range(10):
print('Accuracy of %5s : %2d %%' % (
classes[i], 100 * class_correct[i] / class_total[i]))

Accuracy of plane : 61 %
Accuracy of   car : 85 %
Accuracy of  bird : 46 %
Accuracy of   cat : 23 %
Accuracy of  deer : 40 %
Accuracy of   dog : 36 %
Accuracy of  frog : 80 %
Accuracy of horse : 59 %
Accuracy of  ship : 65 %
Accuracy of truck : 46 %