Visual Document Retrieval on ViDoRe Avg. across 4 datasets v2
0.58Full NDCGEOS-Adaptive
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| EOS-AdaptiveModel=Jina Embeddings v4, gamma=0.22026.01 | 0.58 | 0.89 | 0.53 | 90.91 | — | |
| EOS-AdaptiveModel=Jina Embeddings v4, gamma=0.12026.01 | 0.58 | 0.82 | 0.47 | 81.47 | — | |
| EOS-AdaptiveModel=Jina Embeddings v4, gamma=0.052026.01 | 0.58 | 0.73 | 0.38 | 66.41 | — | |
| RandomModel=Jina Embeddings v4, gamma=0.22026.01 | 0.58 | 0.9 | 0.51 | 88.46 | — | |
| RandomModel=Jina Embeddings v4, gamma=0.12026.01 | 0.58 | 0.83 | 0.46 | 78.49 | — | |
| RandomModel=Jina Embeddings v4, gamma=0.052026.01 | 0.58 | 0.75 | 0.39 | 67.83 | — | |
| ClusterModel=Jina Embeddings v4, gamma=0.22026.01 | 0.58 | 0.83 | 0.47 | 80.81 | — | |
| ClusterModel=Jina Embeddings v4, gamma=0.12026.01 | 0.58 | 0.73 | 0.4 | 67.98 | — | |
| ClusterModel=Jina Embeddings v4, gamma=0.052026.01 | 0.58 | 0.63 | 0.33 | 57.07 | — | |
| SAP-MeanModel=Jina Embeddings v4, gamma=0.22026.01 | 0.58 | 0.91 | 0.55 | 95.6 | — | |
| SAP-MeanModel=Jina Embeddings v4, gamma=0.12026.01 | 0.58 | 0.85 | 0.51 | 88.49 | — | |
| SAP-MeanModel=Jina Embeddings v4, gamma=0.052026.01 | 0.58 | 0.78 | 0.45 | 77.14 | — | |
| SAP-MaxModel=Jina Embeddings v4, gamma=0.22026.01 | 0.58 | 0.91 | 0.56 | 97.18 | — | |
| SAP-MaxModel=Jina Embeddings v4, gamma=0.12026.01 | 0.58 | 0.85 | 0.52 | 89.24 | — | |
| SAP-MaxModel=Jina Embeddings v4, gamma=0.052026.01 | 0.58 | 0.78 | 0.43 | 73.33 | — | |
| EOS-AdaptiveModel=ColPali, gamma=0.22026.01 | 0.56 | 0.86 | 0.44 | 78.6 | — | |
| EOS-AdaptiveModel=ColPali, gamma=0.12026.01 | 0.56 | 0.79 | 0.38 | 68.96 | — | |
| EOS-AdaptiveModel=ColPali, gamma=0.052026.01 | 0.56 | 0.7 | 0.32 | 56.94 | — | |
| RandomModel=ColPali, gamma=0.22026.01 | 0.56 | 0.9 | 0.49 | 86.89 | — | |
| RandomModel=ColPali, gamma=0.12026.01 | 0.56 | 0.84 | 0.44 | 78.04 | — | |
| RandomModel=ColPali, gamma=0.052026.01 | 0.56 | 0.77 | 0.39 | 69.28 | — | |
| ClusterModel=ColPali, gamma=0.22026.01 | 0.56 | 0.88 | 0.47 | 84 | — | |
| ClusterModel=ColPali, gamma=0.12026.01 | 0.56 | 0.78 | 0.39 | 69.41 | — | |
| ClusterModel=ColPali, gamma=0.052026.01 | 0.56 | 0.67 | 0.29 | 52.04 | — | |
| SAP-MeanModel=ColPali, gamma=0.22026.01 | 0.56 | 0.93 | 0.52 | 92.57 | — | |
| SAP-MeanModel=ColPali, gamma=0.12026.01 | 0.56 | 0.88 | 0.48 | 86.15 | — | |
| SAP-MeanModel=ColPali, gamma=0.052026.01 | 0.56 | 0.81 | 0.44 | 78.32 | — | |
| SAP-MaxModel=ColPali, gamma=0.22026.01 | 0.56 | 0.94 | 0.51 | 91.09 | — | |
| SAP-MaxModel=ColPali, gamma=0.12026.01 | 0.56 | 0.88 | 0.48 | 86.13 | — | |
| SAP-MaxModel=ColPali, gamma=0.052026.01 | 0.56 | 0.82 | 0.44 | 78.26 | — | |
| EOS-AdaptiveModel=ColQwen2, gamma=0.22026.01 | 0.54 | 0.84 | 0.47 | 87.27 | — | |
| EOS-AdaptiveModel=ColQwen2, gamma=0.12026.01 | 0.54 | 0.75 | 0.41 | 76.5 | — | |
| EOS-AdaptiveModel=ColQwen2, gamma=0.052026.01 | 0.54 | 0.66 | 0.33 | 62.35 | — | |
| RandomModel=ColQwen2, gamma=0.22026.01 | 0.54 | 0.87 | 0.46 | 85.61 | — | |
| RandomModel=ColQwen2, gamma=0.12026.01 | 0.54 | 0.78 | 0.41 | 75.96 | — | |
| RandomModel=ColQwen2, gamma=0.052026.01 | 0.54 | 0.69 | 0.36 | 66.16 | — | |
| ClusterModel=ColQwen2, gamma=0.22026.01 | 0.54 | 0.8 | 0.45 | 84.05 | — | |
| ClusterModel=ColQwen2, gamma=0.12026.01 | 0.54 | 0.7 | 0.39 | 72.16 | — | |
| ClusterModel=ColQwen2, gamma=0.052026.01 | 0.54 | 0.58 | 0.32 | 58.59 | — | |
| SAP-MeanModel=ColQwen2, gamma=0.22026.01 | 0.54 | 0.89 | 0.48 | 88.61 | — | |
| SAP-MeanModel=ColQwen2, gamma=0.12026.01 | 0.54 | 0.82 | 0.44 | 81.44 | — | |
| SAP-MeanModel=ColQwen2, gamma=0.052026.01 | 0.54 | 0.74 | 0.4 | 75.18 | — | |
| SAP-MaxModel=ColQwen2, gamma=0.22026.01 | 0.54 | 0.89 | 0.49 | 90.79 | — | |
| SAP-MaxModel=ColQwen2, gamma=0.12026.01 | 0.54 | 0.82 | 0.44 | 81.13 | — | |
| SAP-MaxModel=ColQwen2, gamma=0.052026.01 | 0.54 | 0.74 | 0.4 | 74.82 | — | |
| BiModernVBertBackbone=ModernBERT-base, Params=250M, Scoring=Cosine2026.03 | — | — | — | — | 10.9 | |
| ColModernVBertBackbone=ModernBERT-base, Params=250M, Scoring=MaxSim2026.03 | — | — | — | — | 33.4 | |
| ColNomic-7BBackbone=ColQwen2.5-7B, Params=7.0B, Scoring=MaxSim2026.03 | — | — | — | — | 60.4 | |
| ColPaliBackbone=PaliGemma-3B, Params=3.0B, Scoring=MaxSim2026.03 | — | — | — | — | 54.7 | |
| DSE-Qwen2Backbone=Qwen2-VL-2B, Params=2.0B, Scoring=Cosine2026.03 | — | — | — | — | 55.7 | |
| JinaCLIPBackbone=CLIP-ViT-L, Params=400M, Scoring=Cosine2026.03 | — | — | — | — | 26.7 | |
| NanoVDR-LBackbone=ModernBERT-base, Params=151M, Scoring=Cosine2026.03 | — | — | — | — | 61.5 | |
| NanoVDR-MBackbone=BERT-base, Params=112M, Scoring=Cosine2026.03 | — | — | — | — | 62.2 | |
| NanoVDR-SBackbone=DistilBERT, Params=69M, Scoring=Cosine2026.03 | — | — | — | — | 60.5 | |
| NanoVDR-S-MultiBackbone=DistilBERT, Params=69M, Scoring=Cosine2026.03 | — | — | — | — | 61.9 | |
| Qwen3-VL-EmbBackbone=Qwen3-VL-2B, Params=2.0B, Scoring=Cosine2026.03 | — | — | — | — | 65.3 | |
| SigLIP2Backbone=So400m, Params=400M, Scoring=Cosine2026.03 | — | — | — | — | 20.1 | |
| Tomoro-4BBackbone=ColQwen3-4B, Params=4.0B, Scoring=MaxSim2026.03 | — | — | — | — | 65.2 | |
| Tomoro-8BBackbone=ColQwen3-8B, Params=8.0B, Scoring=MaxSim2026.03 | — | — | — | — | 65 |