PyTorch documentation

PyTorch is an optimized tensor library for deep learning using GPUs and CPUs.

Notes

• Automatic Mixed Precision examples
• Autograd mechanics
• Broadcasting semantics
• CPU threading and TorchScript inference
• CUDA semantics
• Distributed Data Parallel
• Extending PyTorch
• Frequently Asked Questions
• Features for large-scale deployments
• Multiprocessing best practices
• Reproducibility
• Serialization semantics
• Windows FAQ

Language Bindings

• C++
• Javadoc

Python API

• torch
• torch.nn
• torch.nn.functional
• torch.Tensor
• Tensor Attributes
• Tensor Views
• torch.autograd
• torch.cuda
• torch.cuda.amp
• torch.distributed
• torch.distributions
• torch.futures
• torch.hub
• torch.jit
• torch.nn.init
• torch.onnx
• torch.optim
• Quantization
• Distributed RPC Framework
• torch.random
• torch.sparse
• torch.Storage
• torch.utils.bottleneck
• torch.utils.checkpoint
• torch.utils.cpp_extension
• torch.utils.data
• torch.utils.dlpack
• torch.utils.model_zoo
• torch.utils.tensorboard
• Type Info
• Named Tensors
• Named Tensors operator coverage
• torch.__config__

Libraries

• torchaudio
• torchtext
• TorchElastic
• TorchServe
• PyTorch on XLA Devices

Community

• PyTorch Contribution Guide
• PyTorch Governance
• PyTorch Governance | Persons of Interest

Indices and tables

• Index

• Module Index