Knowledge-Based Visual Question Answering on E-VQA (Accuracy)
43.6AccuracyIBA-Qwen
Evaluation results
| Method | Links | |
|---|---|---|
| IBA-QwenBackbone=Qwen-2.5-VL, Category=Our proposed IBA2026.06 | 43.6 | |
| IBA-LLaVABackbone=Llama-3.1-8B, Category=Our proposed IBA2026.06 | 43.2 | |
| EchoSightBackbone=Llama-3.1-8B, Category=Retrieve Augmented Models requiring Fine-tuning2026.06 | 41.9 | |
| ReflectiVABackbone=Llama-3.1-8B, Category=Retrieve Augmented Models requiring Fine-tuning2026.06 | 38.6 | |
| Article-qwenBackbone=Qwen-2.5-VL, Category=Zero-shot with Retrieval2026.06 | 35.6 | |
| 1Stage-QwenBackbone=Qwen-2.5-VL, Category=Zero-shot with Re-rank2026.06 | 34.3 | |
| 2Stage-QwenBackbone=Qwen-2.5-VL, Category=Zero-shot with Re-rank2026.06 | 34.1 | |
| 2Stage-llavaBackbone=Llama-3.1-8B, Category=Zero-shot with Re-rank2026.06 | 23.1 | |
| Article-llavaBackbone=Llama-3.1-8B, Category=Zero-shot with Retrieval2026.06 | 23 | |
| Para-QwenBackbone=Qwen-2.5-VL, Category=Zero-shot base models2026.06 | 21.2 | |
| Para-LlavaBackbone=Llama-3.1-8B, Category=Zero-shot base models2026.06 | 13.3 | |
| 1Stage-llavaBackbone=Llama-3.1-8B, Category=Zero-shot with Re-rank2026.06 | 4 |