Visual Perception on MMBench (P)
80.1AccuracyMG-LLaVA
Evaluation Results
| Method | Links | |
|---|---|---|
| MG-LLaVALLM=Yi1.5-34B, Params.=34.4B, Res.=336 (768)2024.06 | 80.1 | |
| MG-LLaVALLM=LLaMA3-8B, Params.=8.4B, Res.=336 (768)2024.06 | 76.5 | |
| GeminiPro Vision2024.06 | 75.2 | |
| GPT-4V2024.06 | 75.1 | |
| MG-LLaVALLM=Phi3-3.8B, Params.=4.2B, Res.=336 (768)2024.06 | 73.4 | |
| MG-LLaVALLM=Vicuna-13B, Params.=13.6B, Res.=336 (768)2024.06 | 72.2 | |
| MG-LLaVALLM=Vicuna-7B, Params.=7.4B, Res.=336 (768)2024.06 | 72.1 | |
| MOVALLM=Vicuna-7B, Params.=10B, Res.=5762024.06 | 70.4 | |
| MiniCPM-V2LLM=MiniCPM-2.4B, Params.=2.8B, Res.=4482024.06 | 69.6 | |
| Mini-GeminiLLM=Vicuna-7B, Params.=7.4B, Res.=336 (768)2024.06 | 69.3 | |
| LLaVA-1.5LLM=Vicuna-13B, Params.=13.4B, Res.=3362024.06 | 69.2 | |
| LLaVA-UHDLLM=Vicuna-13B, Params.=13.4B, Res.=672×10082024.06 | 68 | |
| SPHINX-1kLLM=Vicuna-7B, Params.=10B, Res.=4482024.06 | 67.1 | |
| SPHINXLLM=Vicuna-7B, Params.=10B, Res.=2242024.06 | 66.9 | |
| Qwen-VL-Plus2024.06 | 66.2 | |
| LLaVA-1.5LLM=Vicuna-7B, Params.=7.2B, Res.=3362024.06 | 65.2 | |
| Qwen-VL-ChatLLM=Qwen-7B, Params.=9.6B, Res.=4482024.06 | 60.6 | |
| ShikraLLM=Vicuna-13B, Params.=7.3B, Res.=2242024.06 | 58.8 | |
| IDEFICS-9BLLM=LLaMA-7B, Res.=2242024.06 | 48.2 | |
| Qwen-VLLLM=Qwen-7B, Params.=9.6B, Res.=4482024.06 | 38.2 |