--- license: bsd-3-clause base_model: Qwen/Qwen3-VL-2B-Instruct library_name: peft pipeline_tag: video-text-to-text tags: - caption-evaluation - video-captioning - image-captioning - lora - qwen3-vl --- # Rigel Rigel is a caption quality scorer for videos and images. Given a video or image, a candidate caption, and optionally reference captions, it outputs a continuous score in [0, 1]. ## Usage ```bash git clone https://github.com/SHUonkei/Rigel cd Rigel uv sync uv run rigel-score --media path/to/video.mp4 \ --cand "a man is playing guitar on stage" \ --ref "a musician performs a guitar solo" ``` ```python from rigel import from_pretrained model = from_pretrained() # downloads this repository scores = model( media_path=["path/to/video.mp4"], cand=["a man is playing guitar on stage"], refs=[["a musician performs a guitar solo"]], # [[]] for reference-free scoring ) ```