Don't use source_partitions for quantizing cat #12405
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/12405
Note: Links to docs will display an error until the docs builds have been completed. ✅ You can merge normally! (4 Unrelated Failures)As of commit 0cbe197 with merge base 9591978 ( BROKEN TRUNK - The following jobs 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. |
facebook-github-bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D78138409 |
This PR needs a
|
Summary: Source partitions aren't always reliable. There was model that I recently saw where cat nodes weren't being quantized(likely because they were derived from a different source partition). Moving towards checking all cat nodes should be more reliable Reviewed By: digantdesai Differential Revision: D78138409
Summary: Pull Request resolved: pytorch#12405 Source partitions aren't always reliable. There was model that I recently saw where cat nodes weren't being quantized(likely because they were derived from a different source partition). Moving towards checking all cat nodes should be more reliable Reviewed By: digantdesai Differential Revision: D78138409
|
This pull request was exported from Phabricator. Differential Revision: D78138409 |
3 participants
Summary: Source partitions aren't always reliable. There was model that I recently saw where cat nodes weren't being quantized(likely because they were derived from a different source partition). Moving towards checking all cat nodes should be more reliable
Differential Revision: D78138409