Multimodal Chinese Cultural Knowledge Understanding on CCBench (test)
47.6Average ScoreInternLM-XComposer
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| InternLM-XComposerLanguage Model=InternLM-7B2023.09 | 47.6 | 50.9 | 53.5 | 0.42 | 10.3 | 55 | 73.3 | 50 | |
| Qwen-VL-ChatLanguage Model=Qwen-7B2023.09 | 39.3 | 40.4 | 33.3 | 0.319 | 3.4 | 67 | 51.1 | 42.4 | |
| LLaVA-1.5Language Model=LLaMA2-7B2023.09 | 16.4 | 15.8 | 19.2 | 0.109 | 3.4 | 21 | 37.8 | 12.1 | |
| mPLUG-Owl2023.09 | 12.9 | 22.8 | 17.2 | 0.067 | 0 | 25 | 4.4 | 7.6 | |
| InstructBLIPLanguage Model=Vicuna-7B2023.09 | 12.1 | 8.8 | 9.1 | 0.21 | 0 | 12 | 6.7 | 18.2 | |
| VisualGLM2023.09 | 9.2 | 14 | 11.1 | 0.084 | 0 | 14 | 4.4 | 7.6 | |
| LLaVA2023.09 | 8.3 | 10.5 | 8.1 | 0.076 | 1.7 | 8 | 11.1 | 10.6 | |
| MiniGPT-4Language Model=Vicuna-7B2023.09 | 1.7 | 7 | 4 | 0 | 0 | 1 | 0 | 0 | |
| OpenFlamingoLanguage Model=MPT-7B2023.09 | 0.7 | 1.8 | 0 | 0.008 | 0 | 0 | 2.2 | 1.5 |