Generation on DVisRAG (Real)
69.03AccuracyRobustVisRAG
Evaluation Results
| Method | Links | |
|---|---|---|
| RobustVisRAGRetrieval Setting=Oracle2026.02 | 69.03 | |
| VisRAG-Gen-FM (FARE)Fine-tuning Strategy=FARE, Retrieval Setting=Oracle2026.02 | 64.46 | |
| VisRAG-Gen-FM (PEFT)Fine-tuning Strategy=PEFT, Retrieval Setting=Oracle2026.02 | 63.52 | |
| VisRAG-GenInput Modality=Vision, Retrieval Setting=Oracle2026.02 | 62.68 | |
| VisRAG-Gen-FM (FFT)Fine-tuning Strategy=FFT, Retrieval Setting=Oracle2026.02 | 62.54 | |
| RobustVisRAGRetrieval Setting=top-32026.02 | 59.44 | |
| GPT-4oInput Modality=Vision, Retrieval Setting=Oracle2026.02 | 58.61 | |
| RobustVisRAGRetrieval Setting=top-22026.02 | 57.91 | |
| RobustVisRAGRetrieval Setting=top-12026.02 | 55.39 | |
| VisRAG-Gen-FM (FARE)Fine-tuning Strategy=FARE, Retrieval Setting=top-32026.02 | 53.46 | |
| VisRAG-Gen-FM (FARE)Fine-tuning Strategy=FARE, Retrieval Setting=top-22026.02 | 53.14 | |
| VisRAG-Gen-FM (FARE)Fine-tuning Strategy=FARE, Retrieval Setting=top-12026.02 | 50.95 | |
| VisRAG-Gen-FM (FFT)Fine-tuning Strategy=FFT, Retrieval Setting=top-22026.02 | 49.59 | |
| VisRAG-Gen-FM (PEFT)Fine-tuning Strategy=PEFT, Retrieval Setting=top-32026.02 | 49.35 | |
| VisRAG-Gen-FM (FFT)Fine-tuning Strategy=FFT, Retrieval Setting=top-32026.02 | 49.03 | |
| VisRAG-Gen-FM (PEFT)Fine-tuning Strategy=PEFT, Retrieval Setting=top-22026.02 | 48.82 | |
| VisRAG-Gen-FM (FFT)Fine-tuning Strategy=FFT, Retrieval Setting=top-12026.02 | 47.29 | |
| VisRAG-Gen-FM (PEFT)Fine-tuning Strategy=PEFT, Retrieval Setting=top-12026.02 | 47.22 | |
| VisRAG-GenInput Modality=Vision, Retrieval Setting=top-22026.02 | 47.15 | |
| VisRAG-GenInput Modality=Vision, Retrieval Setting=top-32026.02 | 47.14 | |
| GPT-4oInput Modality=Vision, Retrieval Setting=top-32026.02 | 44.81 | |
| VisRAG-GenInput Modality=Vision, Retrieval Setting=top-12026.02 | 44.7 | |
| GPT-4oInput Modality=Vision, Retrieval Setting=top-22026.02 | 44.08 | |
| GPT-4o (T)Input Modality=Text, Retrieval Setting=Oracle2026.02 | 41.54 | |
| GPT-4oInput Modality=Vision, Retrieval Setting=top-12026.02 | 41.49 | |
| GPT-4o (T)Input Modality=Text, Retrieval Setting=top-32026.02 | 29.44 | |
| GPT-4o (T)Input Modality=Text, Retrieval Setting=top-22026.02 | 28.73 | |
| GPT-4o (T)Input Modality=Text, Retrieval Setting=top-12026.02 | 27.61 | |
| MiniCPM (T)Input Modality=Text, Retrieval Setting=Oracle2026.02 | 25.22 | |
| MiniCPM (T)Input Modality=Text, Retrieval Setting=top-22026.02 | 20.07 | |
| MiniCPM (T)Input Modality=Text, Retrieval Setting=top-32026.02 | 18.8 | |
| MiniCPM (T)Input Modality=Text, Retrieval Setting=top-12026.02 | 18.21 |