--- datasets: - Xiaodong/open-r1-video-4k base_model: - Qwen/Qwen2-VL-7B-Instruct ---