Skip to content

OOM issue #27

@Yaqing2023

Description

@Yaqing2023

The current implementation in inference_basic.py has
video_frames = pipeline(
image=validation_image,
image_pose=validation_control_images,
...
which basically handles all pose images in one pipeline, this requires significant amount of GPU memory and limits the length of the video. I tried to split the validation_control_images into smaller chunks such as 40 images in a sublist and run the pipeline in a for loop. This works well for me and the final video quality looks good to me. Not sure whether this will affect the ID consistency but at least from what i tried results are good.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions