[ONNX] Use TracedONNXFunction op signature to promote inputs to tensors#138770
[ONNX] Use TracedONNXFunction op signature to promote inputs to tensors#138770titaiwangms wants to merge 4 commits intopytorch:mainfrom
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/138770
Note: Links to docs will display an error until the docs builds have been completed. ❗ 1 Active SEVsThere are 1 currently active SEVs. If your PR is affected, please view them below: ✅ No FailuresAs of commit 40a424b with merge base 5aadaaf ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
|
@justinchuby I totally forgot the status of this. Are we still doing this? |
|
Sure! Please feel free to merge |
|
One sec |
|
I am not sure if _process_python_sequences is safe to do in the trace level? I suppose we need to preserve some python lists? |
|
Maybe just do the constants for now? What do you think? |
|
I will test more on this. |
Agree. python list related properties are leveraged heavily in trace_only function. Also, |
|
@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 |
…rs (pytorch#138770) Previous to this PR, in torchlib TracedONNXFunction, the inputs could be python constants even if the annotation sets to TensorTypes. Pull Request resolved: pytorch#138770 Approved by: https://github.com/justinchuby
Previous to this PR, in torchlib TracedONNXFunction, the inputs could be python constants even if the annotation sets to TensorTypes.