main
Update document for autocast on CPU (#135299)
[doc] Add documentation for division by zero behavior in autograd (#155987)
Fix comment on broadcasting example to clarify dimension mismatch (#162177)
[3/n] Remove references to TorchScript in PyTorch docs (#158315)
Fix typo in FP16 accumulation section (#167703)
Redirect the custom ops landing page :D (#139634)
Update DDP dynamo debug docs (#118295)
Fix the example in the extending.func.rst (#109279)
[autograd][docs] Add more details on why save_for_backward is important in extending autograd note (#153005)
Update faq.rst so OOM section mentions checkpoint (#62709)
update supported OS for Intel client GPU (#161699)
[BE] fix typos in docs/ (#156080)
[ROCm] Ck backend UX refactor (#152951)
Add guidance on how to migrate kernels to the libtorch stable ABI (#167112)
Enable TF32 as fp32 internal precision for matmul/linear/conv (#157520)
Fix to modules.rst: indent line with activation functions (#139667)
update mps note with more details (#78669)
Update warning of TF32 (#158209)
add Out Notes (#151306)
[cuBLAS] update cuBLAS determinism docs, remove workspace requirement checks (#161749)
Document limitations of weights_only in SECURITY.md and torch.load doc (#165645)
Removing conda references from PyTorch Docs (#152702)