Datasets:
| name: Video-MME-v2 | |
| description: Video-MME-v2 is a benchmark for evaluating the next stage of video understanding capabilities of multimodal large language models. | |
| evaluation_framework: video-mme-v2 | |
| tasks: | |
| - id: video-mme-v2 | |
| config: default | |
| split: test |