Removing .float() (autocast in fp16 will discard this (I think)).
#495
Conversation
This file contains hidden or 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
|
The documentation is not available anymore as the PR was closed or merged. |
|
Just checked. You're 100% right @Narsil . Feel free to merge whenever :-) |
patrickvonplaten
approved these changes
Sep 13, 2022
|
Sorry to re-open this, but another way to solve this is by using # make sure hidden states is in float32
# when running in half-precision
hidden_states = self.norm2.float()(hidden_states.float()).type(hidden_states.dtype)which respects the comment above saying that Normalization must be run in fp16 (idk if the comment is true though). My PR does use this fix for example and it works fine #371 |
PhaneeshB
pushed a commit
to nod-ai/diffusers
that referenced
this pull request
Mar 1, 2023
Co-authored-by: vivian <vivian@nod-labs.com>
yoonseokjin
pushed a commit
to yoonseokjin/diffusers
that referenced
this pull request
Dec 25, 2023
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Outdated suggestions cannot be applied.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.
While investigating performance improvements, the main culprit I found was
autocastwhich actually adds a LOT of copies of tensors a bit everywhereto cast to
float16. I added a lot of custom made casts for now in an offbranch.This PR tries to check 1 of the changes which is this
.float()cast which breakin
fp16mode (because the model is in fp16, while the tensor is now infloat.