fix video diffusion black output if --bf16 is set. #1685
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
add hpu graph support Signed-off-by: Wang, Yi A <[email protected]>
|
find it's caused by sdpa. so use hpu AttnProcessor2_0 to replace the original one |
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
Signed-off-by: Wang, Yi A <[email protected]>
Signed-off-by: Wang, Yi A <[email protected]>
Signed-off-by: Wang, Yi A <[email protected]>
|
The code quality check failed, please run |
|
(Need to run |
libinta
reviewed
Jan 22, 2025
| @@ -372,7 +383,7 @@ def __call__( | |||
| # 4. Encode input image using VAE | |||
| image = self.video_processor.preprocess(image, height=height, width=width) | |||
| # torch.randn is broken on HPU so running it on CPU | |||
| rand_device = "cpu" if device.type == "hpu" else device | |||
| rand_device = torch.device("cpu") if device.type == "hpu" else device | |||
There was a problem hiding this comment.
@sywangyi is it still requiring rand to be on cpu to fix the black output issue?
There was a problem hiding this comment.
no, black issue is fixed by using fused sdpa, and sdp_on_bf16.
this "cpu“ is fix the ci failure.
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Outdated suggestions cannot be applied.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.
add hpu graph support
follow the readme guide
PT_HPU_MAX_COMPOUND_OP_SIZE=1 python image_to_video_generation.
py --model_name_or_path "stabilityai/stable-video-diffusion-img2vid-xt" --image_path "https://huggingface.co/datasets/huggingface/document
ation-images/resolve/main/diffusers/svd/rocket.png" --num_videos_per_prompt 1 --video_save_dir /tmp/stable_video_diffusion_xt --save_frames_as_images --use_habana --use_hpu_graphs --gaudi_config Habana/stable-diffusion --sdp_on_bf16 --bf16
the output video is black in 1.19