AI
Learn PyTorch
PyTorch is the Python library most research and production training code uses for tensors, automatic differentiation, and neural networks.
For programmers who can write Python and want to train or modify models instead of only calling an API.
Start here
Finish the official basics intro, then train the 60 Minute Blitz network and change one hyperparameter on purpose.
Recommended resources
Start here · BeginnerLearn the Basics
Official introduction to tensors, datasets, and the training loop.
pytorch.orgBuild · BeginnerDeep Learning with PyTorch: A 60 Minute Blitz
A short path from tensors to training a small image classifier.
pytorch.orgReference · IntermediatePyTorch documentation
The API reference to consult once the tutorials use a function you do not recognize.
pytorch.orgNext · IntermediateHugging Face Transformers
Shows how PyTorch models are packaged and loaded in current language-model workflows.
huggingface.coRelated topics
Understand transformer models · Fine-tune a language model · Learn linear algebra for machine learning