Visual Question Answering on GAP-VQA (val)
0.637GPT ScoreGPT-4V
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GPT-4Vtraining_status=Pre-fine-tuning2024.10 | 0.637 | — | |
| MiniCPM-Llama3-V-2.5-8Btraining_status=Post-fine-tuning, method=GAP-VQA2024.10 | 0.477 | 0.3 | |
| Qwen2-VL-2Btraining_status=Post-fine-tuning, method=GAP-VQA2024.10 | 0.285 | 0.116 | |
| Qwen2-VL-7Btraining_status=Post-fine-tuning, method=GAP-VQA2024.10 | 0.25 | 0.043 | |
| Qwen2-VL-7Btraining_status=Pre-fine-tuning2024.10 | 0.207 | — | |
| Qwen2-VL-2Btraining_status=Pre-fine-tuning2024.10 | 0.169 | — | |
| MiniCPM-Llama3-V-2.5-8Btraining_status=Pre-fine-tuning2024.10 | 0.147 | — |