Pytorch Lessons - Basics imports

I will start the Pytorch Series with interview-like formatted questions for us. Here, we will see questions on tutorials. Tutorial link

Usually, common libraries used are:

import torch 
import torchvision
import torch.nn as nn
import numpy as np
import torchvision.transforms as transforms

What are these and why use?

1. torch – PyTorch Core Library

It's widely used for deep learning and tensor computation, similar to NumPy, but with GPU acceleration.

Common Uses:

• Tensor creation and operations (like NumPy arrays but faster with GPU)

• Defining and training neural networks

• Autograd (automatic differentiation)

• Optimizers for training models

2. torchvision – Vision-Specific Utilities for PyTorch

orchvision is a library that provides computer vision utilities like:

• Pretrained models (ResNet, AlexNet, etc.)

• Image datasets (CIFAR10, MNIST, etc.)

• Image transforms and data loaders

Common Uses:

• Load and preprocess image data

• Use pre-trained vision models for transfer learning

3. torch.nn – Neural Network Module

torch.nn provides classes to help build neural networks easily. You define models as subclasses of nn.Module.

🔹 Common Uses:

• Define layers like nn.Linear, nn.Conv2d, nn.ReLU, etc.

• Create custom neural networks

4. torchvision.transforms – Data Transformations

transforms are used to modify and preprocess image data before feeding it to the model.

🔹 Common Uses:

• Convert images to tensors

• Normalize image pixels

• Resize, crop, rotate, and augment images

5. numpy – Numerical Python

Common Uses:

• Matrix operations

• Data manipulation and numerical analysis

• Integration with PyTorch (converting between NumPy arrays and tensors)