facebookresearch / facebookresearch/sam2
When my ann_frame_idx is not 0, running the model throws an error
Open
- Dominant language
- Jupyter Notebook
- Stars
- 19.9k
- Forks
- 2.5k
- PR merge metrics
- No merged PRs in 30d
Description
ann_frame_idx = 24 # the frame index we interact with
When my ann_frame_idx is not 0, running the model throws an error: "for out_obj_id, out_mask in video_segments[out_frame_idx].items(): KeyError: 0". This indicates that when I execute predictor.propagate_in_video(inference_state), the model only propagates backward. How can I make the model propagate across all images?
Contributor guide
Assessment
This issue has not been assessed yet.