- Notifications
You must be signed in to change notification settings - Fork 25.9k
Pull requests: pytorch/pytorch
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Update Kineto Submodule ci-no-td Do not run TD on this PR ciflow/rocm Trigger "default" config CI on ROCm ciflow/rocm-mi300 Trigger "default" config CI on ROCm MI300 ciflow/trunk Trigger trunk jobs on your pull request fb-exported Merged meta-exported Reverted topic: not user facing topic category
#167343 opened Nov 7, 2025 by sraikund16 Loading…
[BE][Typing][Dynamo] Type torch/_dynamo/variables/nn_module.py ciflow/inductor module: dynamo
#167342 opened Nov 7, 2025 by Lucaskabela • Draft
[precompile] Integrate AOTI as a backend. ci-no-td Do not run TD on this PR ciflow/inductor ciflow/trunk Trigger trunk jobs on your pull request Merged module: dynamo module: inductor release notes: inductor (aoti) Reverted topic: not user facing topic category
#167338 opened Nov 7, 2025 by zhxchen17 Loading…
Validate Related to torch.optim open source release notes: optim triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
Adam, AdamW's betas type module: optimizer #167336 opened Nov 7, 2025 by PommesPeter Loading…
Build cpuinfo into c10 shared library ciflow/linux-aarch64 linux aarch64 CI workflow ciflow/trunk Trigger trunk jobs on your pull request open source release notes: build release notes category topic: improvements topic category topic: not user facing topic category triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
#167328 opened Nov 7, 2025 by murste01 Loading…
[TESTING] Bumped triton commit to try some disabled tests ciflow/inductor-rocm Trigger "inductor" config CI on ROCm ciflow/rocm Trigger "default" config CI on ROCm ciflow/rocm-mi300 Trigger "default" config CI on ROCm MI300 open source topic: not user facing topic category
#167325 opened Nov 7, 2025 by iupaikov-amd • Draft
cuda max_pool2d: switch to NHWC when output > INT_MAX to avoid overflow module: cuda Related to torch.cuda, and CUDA support in general module: memory format Memory format/layout related issues/changes (channels_last, nhwc) module: pooling open source release notes: cuda release notes category triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
#167322 opened Nov 7, 2025 by lokashrinav Loading…
[4/N] Add return types of Python functions ciflow/inductor oncall: distributed Add this issue/PR to distributed oncall triage queue oncall: jit Add this issue/PR to JIT oncall triage queue open source release notes: distributed (checkpoint) release notes: distributed (sharded) release notes category triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
#167320 opened Nov 7, 2025 by cyyever Loading…
[decomp] Add max_pool2d_with_indices_backward decomposition using gather+reduction module: inductor open source release notes: inductor triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
#167318 opened Nov 7, 2025 by stmcgovern Loading…
feat: Add Memory/CUDA Debugging Utilities and Core Stability Fixes open source triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
#167316 opened Nov 7, 2025 by chandang207 Loading…
[10/N] Fix unused loop variables in tests ciflow/inductor ciflow/mps Run MPS tests (subset of trunk) fx module: inductor oncall: distributed Add this issue/PR to distributed oncall triage queue open source release notes: quantization release notes category triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
#167314 opened Nov 7, 2025 by cyyever Loading…
Fix integer overflow in (fake) quantization module: cpu CPU specific problem (e.g., perf, algorithm) open source release notes: quantization release notes category triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
#167306 opened Nov 7, 2025 by Flamefire Loading…
Fix distributed plugin backend _set_default_backend fail bug oncall: distributed Add this issue/PR to distributed oncall triage queue open source release notes: distributed (c10d) release notes category triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
#167303 opened Nov 7, 2025 by Elvin-Ma Loading…
[TorchElastic] fix #135712, clear MASTER_ADDR in store on agent shutdown oncall: distributed Add this issue/PR to distributed oncall triage queue open source release notes: distributed (torchelastic) triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
#167302 opened Nov 7, 2025 by linxxx3 Loading…
Use _StridedShard to replace shard_order field in DTensorSpec ciflow/inductor oncall: distributed Add this issue/PR to distributed oncall triage queue topic: not user facing topic category
#167300 opened Nov 7, 2025 by zpcore Loading…
Functionality documentation issue #9886 oncall: distributed Add this issue/PR to distributed oncall triage queue open source triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
#167289 opened Nov 7, 2025 by J-Ruiz16 Loading…
docs: PyTorch Compute Platform Quality Levels module: docs Related to our documentation, both in docs/ and docblocks open source topic: not user facing topic category triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
#167282 opened Nov 7, 2025 by dvrogozh Loading…
[DTensor] Fix convolution ops with bias=None in torch.compile module: dtensor distributed tensor tag oncall: distributed Add this issue/PR to distributed oncall triage queue open source topic: not user facing topic category triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
#167258 opened Nov 6, 2025 by stmcgovern Loading…
Add missing return type annotation for named_modules() module: typing Related to mypy type annotations oncall: distributed Add this issue/PR to distributed oncall triage queue open source topic: not user facing topic category triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
#167256 opened Nov 6, 2025 by yashwantbezawada Loading…
[do not review] Use GraphExecGroup in PP oncall: distributed Add this issue/PR to distributed oncall triage queue
#167251 opened Nov 6, 2025 by soulitzer Loading…
[dynamo, BE] Remove IncorrectUsage exception module: dynamo open source topic: not user facing topic category triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
#167249 opened Nov 6, 2025 by yashwantbezawada Loading…
Fix thread safety in getCurrentCUDABlasHandle and getCUDABlasLtWorkspace ci-no-td Do not run TD on this PR ciflow/h100 ciflow/trunk Trigger trunk jobs on your pull request fb-exported Merged meta-exported Reverted topic: not user facing topic category
Run dynamo tests on Python 3.14 open source topic: not user facing topic category
#167246 opened Nov 6, 2025 by guilhermeleobas • Draft
ProTip! Find all pull requests that aren't related to any open issues with -linked:issue.