Reduce WanAnimate TorchAO test input sizes to prevent OOM - #13541
Merged
sayakpaul merged 7 commits intoMay 8, 2026
Conversation
Shrink dummy inputs to avoid OOM on devices without FlashAttention. Reduce hidden_states spatial from 64x64 to 16x16 and frames from 21 to 5, bringing self-attention sequence length from 21,504 to 320.
github-actions
Bot
added
tests
size/S
github-actions
Bot
added
size/S
Contributor
Author
|
The failed test seems no related to my changes. |
github-actions
Bot
added
size/S
github-actions
Bot
added
size/S
sayakpaul
approved these changes
May 1, 2026
Contributor
Author
|
The previous failed CI are not related to my changes. |
Contributor
Author
|
Hi @sayakpaul . The 2 failed CI jobs (LoRA tests with PEFT main, PyTorch Example CPU tests) are unrelated to this PR. This PR only modifies test input sizes in test_models_transformer_wan_animate.py, while the failures are in LoRA/PEFT integration tests and example scripts respectively. These appear to be pre-existing flaky tests on main. |
ADJB1212
pushed a commit
to ADJB1212/diffusers
that referenced
this pull request
May 12, 2026
…e#13541) Shrink dummy inputs to avoid OOM on devices without FlashAttention. Reduce hidden_states spatial from 64x64 to 16x16 and frames from 21 to 5, bringing self-attention sequence length from 21,504 to 320. Co-authored-by: dg845 <58458699+dg845@users.noreply.github.com> Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What does this PR do?
Shrinks dummy inputs in
TestWanAnimateTransformer3DTorchAoto avoid OOM on devices without FlashAttention (e.g. XPU, which falls back to math SDPA and materializes the full O(S²) attention matrix).Reduced
hidden_statesspatial from 64×64 to 16×16 and frames from 21 to 5, bringing self-attention sequence length from 21,504 to 320. Face frame count adjusted from 77 to 13 to satisfy the temporal alignment constraint with the face encoder's stride-2 convolutions.