Backend 3

PyTorch

PyTorch is a tensor library commonly used for machine learning and GPU-backed research code.

When to use PyTorch

Use the official PyTorch site when you need installation notes, documentation, or ecosystem information.

Official PyTorch site

Use PyTorch when your workflow involves neural networks, tensor operations on a GPU, or model training loops.

It can also be useful for numerical experiments that benefit from tensors and automatic differentiation.

Minimal example

import torch

x = torch.tensor([0.0, 1.0, 2.0])
y = x ** 2

print(y)

Working habit

Check where tensors live before mixing CPU and GPU code. Device mismatches are a common source of early errors.

print(x.device)