fix FlUX.1_dev guidance_batches bug for pad case in _split_inputs_int… #1607
+6
−4
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Outdated suggestions cannot be applied.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.
…o_batches
What does this PR do?
Fixes # (issue)
https://habana.atlassian.net/browse/CS-864
Before submitting
runtime error when num_images_per_prompt is not divisible by batch_size .
test cmd:
python text_to_image_generation.py --model_name_or_path /nfs_home/data/FLUX.1-dev --prompts "A cat holding a sign that says hello world" --num_images_per_prompt 10 --batch_size 4 --num_inference_steps 4 --image_save_dir /tmp/flux_1_images --scheduler flow_match_euler_discrete --use_habana --use_hpu_graphs --gaudi_config Habana/stable-diffusion --bf16
Traceback (most recent call last):
File "/host/mnt/ctrl/disk1/personal/huijuanz/ws/optimum-habana/examples/stable-diffusion/text_to_image_generation.py", line 699, in
main()
File "/host/mnt/ctrl/disk1/personal/huijuanz/ws/optimum-habana/examples/stable-diffusion/text_to_image_generation.py", line 662, in main
outputs = pipeline(prompt=args.prompts, **kwargs_call)
File "/usr/local/lib/python3.10/dist-packages/torch/utils/_contextlib.py", line 116, in decorate_context
return func(*args, **kwargs)
File "/host/mnt/ctrl/disk1/personal/huijuanz/ws/optimum-habana/optimum/habana/diffusers/pipelines/flux/pipeline_flux.py", line 638, in call
) = self._split_inputs_into_batches(batch_size, latents, prompt_embeds, pooled_prompt_embeds, guidance)
File "/host/mnt/ctrl/disk1/personal/huijuanz/ws/optimum-habana/optimum/habana/diffusers/pipelines/flux/pipeline_flux.py", line 371, in _split_inputs_into_batches
sequence_to_stack = (guidance_batches[-1],) + tuple(
File "/host/mnt/ctrl/disk1/personal/huijuanz/ws/optimum-habana/optimum/habana/diffusers/pipelines/flux/pipeline_flux.py", line 372, in
torch.zeros_like(guidance_batches[-1][0][None, :]) for _ in range(num_dummy_samples)
IndexError: too many indices for tensor of dimension 0