Document and OCR Understanding on Doc&OCR Suite
57.1ChartQAGenLIP
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| GenLIPArch=g/16, Data=8.0B, Evaluation Protocol=Frozen visual representation, LLM Backbone=LLaVA-NeXT-Qwen2.5-7B2026.05 | 57.1 | 65.9 | 69 | 66.8 | 81 | 43.6 | |
| GenLIPArch=So/16, Data=8.0B, Evaluation Protocol=Frozen visual representation, LLM Backbone=LLaVA-NeXT-Qwen2.5-7B2026.05 | 55.3 | 63.5 | 66.3 | 65.7 | 81 | 41.4 | |
| GenLIPArch=L/16, Data=8.0B, Evaluation Protocol=Frozen visual representation, LLM Backbone=LLaVA-NeXT-Qwen2.5-7B2026.05 | 52.7 | 59.2 | 61.7 | 62.9 | 80.4 | 38.8 | |
| SigLIP2Arch=g/16, Data=40.0B, Evaluation Protocol=Frozen visual representation, LLM Backbone=LLaVA-NeXT-Qwen2.5-7B2026.05 | 47.2 | 55.6 | 56.3 | 63.5 | 81 | 36.4 | |
| SigLIP2Arch=So/16, Data=40.0B, Evaluation Protocol=Frozen visual representation, LLM Backbone=LLaVA-NeXT-Qwen2.5-7B2026.05 | 46.6 | 55.6 | 56.3 | 63.5 | 81.3 | 37.2 | |
| OVision2Arch=L/16, Data=12.8B, Evaluation Protocol=Frozen visual representation, LLM Backbone=LLaVA-NeXT-Qwen2.5-7B2026.05 | 42.5 | 49.9 | 49.5 | 58.8 | 78.4 | 33.8 | |
| SigLIPArch=L/16, Data=40.0B, Evaluation Protocol=Frozen visual representation, LLM Backbone=LLaVA-NeXT-Qwen2.5-7B2026.05 | 41.7 | 45.7 | 50.5 | 56 | 79.3 | 34.8 | |
| AIMv2Arch=L/14, Data=12.0B, Evaluation Protocol=Frozen visual representation, LLM Backbone=LLaVA-NeXT-Qwen2.5-7B2026.05 | 36.8 | 30.9 | 46.6 | 54.5 | 76.9 | 37.5 | |
| CLIPArch=L/14, Data=12.8B, Evaluation Protocol=Frozen visual representation, LLM Backbone=LLaVA-NeXT-Qwen2.5-7B2026.05 | 36.6 | 29.6 | 48.4 | 52.6 | 76.3 | 39 |