Visual Document Retrieval on ChartQA
93.6NDCG@5ReAlign
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| ReAlignRetriever Type=Multi-modal, Backbone=Qwen2026.04 | 93.6 | 94 | — | — | |
| ReAlignRetriever Type=Multi-modal, Backbone=Phi3V2026.04 | 87.9 | 88.4 | — | — | |
| VDocRetrieverRetriever Type=Multi-modal2026.04 | 86 | 87.1 | — | — | |
| ColPaliRetriever Type=Multi-modal2026.04 | 83.8 | 84.7 | — | — | |
| NV-EmbedRetriever Type=Text-based2026.04 | 79.4 | 80.1 | — | — | |
| E5-VRetriever Type=Multi-modal2026.04 | 78.6 | 79.9 | — | — | |
| DSERetriever Type=Multi-modal2026.04 | 76.6 | 77.1 | — | — | |
| E5-MistralRetriever Type=Text-based2026.04 | 73.6 | 74.2 | — | — | |
| SigLIP 2Retriever Type=Multi-modal2026.04 | 69.9 | 71.9 | — | — | |
| VLM2VecRetriever Type=Multi-modal2026.04 | 69 | 71.4 | — | — | |
| BGERetriever Type=Text-based2026.04 | 61.4 | 62.7 | — | — | |
| BM25Retriever Type=Text-based2026.04 | 50 | 52.3 | — | — | |
| CLIPRetriever Type=Multi-modal2026.04 | 33 | 35.4 | — | — | |
| BGE-LargeModality=Text-Based2026.05 | — | — | 53.76 | 42.47 | |
| BM25Modality=Text-Based2026.05 | — | — | 56.69 | 43.66 | |
| ColPaliModality=Visual-Based2026.05 | — | — | 77.16 | 62.68 | |
| DSEModality=Visual-Based2026.05 | — | — | 78.13 | 63.42 | |
| DSEModality=Hybrid2026.05 | — | — | 74.51 | 62.68 | |
| GTRModality=Text-Based2026.05 | — | — | 56.55 | 43.5 | |
| MiniCPMModality=Text-Based2026.05 | — | — | 73.96 | 60.91 | |
| NV-EmbedModality=Text-Based2026.05 | — | — | 52.09 | 43.04 | |
| SigLIPModality=Visual-Based2026.05 | — | — | 66.16 | 47.62 | |
| UnveilModality=Visual-Textual2026.05 | — | — | 79.53 | 66.75 | |
| UnveilModality=Visual-Only2026.05 | — | — | 80.36 | 66.4 | |
| VisRAGModality=Visual-Based2026.05 | — | — | 78.83 | 64.54 | |
| VisRAGModality=Hybrid2026.05 | — | — | 77.58 | 64.35 |