Multimodal Visual Question Assessment on EvalQABench (test)
79.58AccuracyLOVA³-7B
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| LOVA³-7BLLM=Vicuna-7B2024.05 | 79.58 | 79.15 | 79.72 | 49.26 | |
| InternLM-XCLLM=InternLM-7B2024.05 | 69.58 | 70.66 | 68.76 | 52.62 | |
| LLaVA-1.5LLM=Vicuna-7B2024.05 | 64.92 | 61.28 | 69.8 | 33.84 | |
| Qwen-VL-ChatLLM=Qwen-7B2024.05 | 63.66 | 63.48 | 63.9 | 49.34 | |
| InstructBLIPLLM=Vicuna-13B2024.05 | 61.42 | 57.6 | 69.18 | 24.82 | |
| CogVLMLLM=Vicuna-7B2024.05 | 60.64 | 56.59 | 69.88 | 19.32 | |
| BLIP2LLM=Flan-T5-XXL-11B2024.05 | 58 | 82.79 | 32.47 | 87.8 | |
| InstructBLIPLLM=Vicuna-7B2024.05 | 38.04 | 41.49 | 48.47 | 29.76 |