generator-based live step status (track_tqdm doesnt forward on zerogpu); fix gen_mask tqdm crash

progress via track_tqdm + SAM3 tqdm; dilation all-frames with base fallback + reset on new video

2-col layout (output beside input/mask); live mask dilation; man-dancing->robot example

mask video beside input; SAM3 generate-mask button; dilation slider; real mask-video examples

SAM3 video segmentation (tracked); video mask preview + mask-video upload

concise description: base model + LoRA links, via diffusers

audio folded into single prompt; inpaint example None->placeholder mask

remove brush (dict crash) -> mask upload; type-guard duration; single prompt

SAM3 per-frame tracking; brush=static region

fixes: natural-speed/aspect, audio, progress, prompts, examples

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub