PyTorch Tutorial – Setup PyTorch

This is the first in the series of tutorials on Building Deep Learning Models with Pytorch. In this part we would cover the following:

1. Introduction to Pytorch
2. Setup Pytorch Locally
3. Import the Required Modules

1. Introduction to Pytorch

Pytorch is a Python library (some call it framework) developed by Meta AI used for building deep learning models.  And you know that deep learning is a subset of machine learning. Therefore we can say that Pytorch is a tool for building advanced machine learning models.

In conventional machine learning, we performed tasks like building classification models (classifiers), regression models etc. With Pytorch we can do more. We could build more complex models for tasks like image recognition and language processing,

2. Setup Pytorch Locally

In this tutorial, we would be using Jupyter Notebook.

You can setup PyTorch on Windows, Mac or Linux by running the command below:

conda install pytorch torchvision -c pytorch

This command may take several minutes to complete if you are installing PyTorch for the first time

3. Import the Required Modules

We would need a number of modules including:

• nn – provided by PyTorch and helps you create a NeuralNetwork
• DataLoader – helps you efficiently load data for training
• datasets – from torchvision and provides us with datasets for image analysis
• ToTensor – converts an image or ndarray to PyTorch tensor and scales the values as needed.

So add the following:

import torch
from torch import nn
from torch.utils.data import DataLoader
from torchvision import datasets
from torchvision.transforms import ToTensor

In the next tutorial, we would download our training and test datasets.