Video-ChatGPT
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Video-based generative performance | Video-ChatGPT benchmark | Correctness Score61.6 | 76 | |
| Video Question Answering | Video-ChatGPT | Correctness Score4.09 | 28 | |
| Video Text Generation | Video-ChatGPT benchmark v1 (test) | Correctness of Information (CI)3.4 | 14 | |
| Video Understanding | Video-ChatGPT (VCG) (test) | LLM Score (GPT-3.5)4.06 | 13 | |
| Video Question Answering | Video-ChatGPT benchmark zero-shot (test) | CI3.07 | 12 | |
| Video Conversation | Video-ChatGPT benchmark ActivityNet-200 (test) | Correctness3.08 | 10 |
Showing 6 of 6 rows