Multimodal Conversation on LLaVA-Bench In-the-Wild v1 (test)
0.726General ScoreShareGPT4V-7B
Evaluation Results
| Method | Links | |
|---|---|---|
| ShareGPT4V-7BLanguage Model=Vicuna-7B2023.11 | 0.726 | |
| LLaVA-1.5Language Model=Vicuna-13B2023.11 | 0.707 | |
| LLaVA-1.5Language Model=Vicuna-7B2023.11 | 0.634 | |
| LLaVALanguage Model=Vicuna-7B2023.11 | 0.63 | |
| InstructBLIPLanguage Model=Vicuna-7B2023.11 | 0.609 | |
| InstructBLIPLanguage Model=FLAN-T52023.11 | 0.582 | |
| BLIP-2Language Model=FLAN-T52023.11 | 0.381 |