pytorch
https://github.com/pytorch/pytorch
Python
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Python not yet supported919 Subscribers
View all SubscribersAdd a CodeTriage badge to pytorch
Help out
- Issues
- 1.10.0 failed to build due to missing TensorBody.h
- Deterministic implementation for grid_sampler_2d_backward_cuda
- NCCL Network is unreachable / Connection refused when initializing DDP
- Inconsistent list indexing behavior
- Can not guarantee reproducible for `nn.ReflectionPad2d`
- Generalized matmul
- `torch.flip`, `torch.roll`, `torch.tile` has arguments `dims=` for both dim/multidim while other functions have argument `dim=` for the same usage
- `(svd|pca)_lowrank`: backward is unstable when for a matrix `A`, the parameter `q` is set to a value `q > rank(A)`.
- CPU only mode with cudatoolkit 11.3
- Torch Profiler does not count FLOPs for backward pass
- Docs
- Python not yet supported