Traceable wrapper subclass support for deferred runtime asserts #126198
Conversation
[ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/126198
Note: Links to docs will display an error until the docs builds have been completed. ✅ You can merge normally! (1 Unrelated Failure)As of commit 471a2bc with merge base 8c38d0c ( BROKEN TRUNK - The following job failed but were present on the merge base:👉 Rebase onto the `viable/strict` branch to avoid these failures
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
…serts" The padded dense -> jagged conversion op has the signature: ``` _fbgemm_dense_to_jagged_forward(Tensor dense, Tensor[] offsets, SymInt? total_L=None) -> Tensor ``` when `total_L` is not specified, the meta registration has a data-dependent output shape (based on `offsets[0][-1]`). Returning an unbacked SymInt here should work in theory, but traceable wrapper subclass support is missing in later code to handle deferred runtime asserts. This PR fixes this. [ghstack-poisoned]
…serts" The padded dense -> jagged conversion op has the signature: ``` _fbgemm_dense_to_jagged_forward(Tensor dense, Tensor[] offsets, SymInt? total_L=None) -> Tensor ``` when `total_L` is not specified, the meta registration has a data-dependent output shape (based on `offsets[0][-1]`). Returning an unbacked SymInt here should work in theory, but traceable wrapper subclass support is missing in later code to handle deferred runtime asserts. This PR fixes this. [ghstack-poisoned]
|
@pytorchbot merge |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Merge failedReason: 1 mandatory check(s) failed. The first few are: Dig deeper by viewing the failures on hud |
…serts" The padded dense -> jagged conversion op has the signature: ``` _fbgemm_dense_to_jagged_forward(Tensor dense, Tensor[] offsets, SymInt? total_L=None) -> Tensor ``` when `total_L` is not specified, the meta registration has a data-dependent output shape (based on `offsets[0][-1]`). Returning an unbacked SymInt here should work in theory, but traceable wrapper subclass support is missing in later code to handle deferred runtime asserts. This PR fixes this. [ghstack-poisoned]
|
@pytorchbot merge |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Merge failedReason: 1 mandatory check(s) failed. The first few are: Dig deeper by viewing the failures on hud |
|
@pytorchbot merge -i |
Merge failedReason: 20 jobs have failed, first few of them are: pull / linux-focal-py3.12-clang10 / test (default, 2, 3, linux.2xlarge), pull / linux-focal-py3.8-clang10 / test (default, 2, 3, linux.2xlarge), pull / linux-focal-py3.8-clang10 / test (crossref, 2, 2, linux.2xlarge), pull / linux-jammy-py3.10-clang15-asan / test (default, 2, 6, linux.4xlarge), pull / linux-jammy-py3.8-gcc11 / test (default, 2, 3, linux.2xlarge) Details for Dev Infra teamRaised by workflow job |
…serts" The padded dense -> jagged conversion op has the signature: ``` _fbgemm_dense_to_jagged_forward(Tensor dense, Tensor[] offsets, SymInt? total_L=None) -> Tensor ``` when `total_L` is not specified, the meta registration has a data-dependent output shape (based on `offsets[0][-1]`). Returning an unbacked SymInt here should work in theory, but traceable wrapper subclass support is missing in later code to handle deferred runtime asserts. This PR fixes this. [ghstack-poisoned]
|
@pytorchbot merge |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Merge failedReason: 1 mandatory check(s) failed. The first few are: Dig deeper by viewing the failures on hud |
…serts" The padded dense -> jagged conversion op has the signature: ``` _fbgemm_dense_to_jagged_forward(Tensor dense, Tensor[] offsets, SymInt? total_L=None) -> Tensor ``` when `total_L` is not specified, the meta registration has a data-dependent output shape (based on `offsets[0][-1]`). Returning an unbacked SymInt here should work in theory, but traceable wrapper subclass support is missing in later code to handle deferred runtime asserts. This PR fixes this. [ghstack-poisoned]
|
@pytorchbot merge |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Stack from ghstack (oldest at bottom):
The padded dense -> jagged conversion op has the signature:
when
total_Lis not specified, the meta registration has a data-dependent output shape (based onoffsets[0][-1]). Returning an unbacked SymInt here should work in theory, but traceable wrapper subclass support is missing in later code to handle deferred runtime asserts. This PR fixes this.