Fix dtype error in MHA layer/change dtype checking mechanism for manual cast #14791
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
torch.nn.MultiheadAttention have so much forks inside their forward call and is impossible to patch their inside call of computation operationg. I decide to give it "always fp32" when manual_cast is used.
And since the dtype error I got is weird, it claim the tensor it got is fp32 and fp8, which should never happen.(Should at least be fp32 and fp16)
So it is possible that torch_utils.get_param is not working for my case.
Based on some test, I finally decide to just go through all the parameters to check if there are any different dtype parameters.
Checklist: