Multimodal Evaluation on Vision-Flan Multimodal Suite
115.8Relative ScoreVisNec
Evaluation Results
| Method | Links | |
|---|---|---|
| VisNecBackbone=LLaVA-v1.5-7B, Training Dataset=Vision-Flan-186K, Sampling Ratio=15% (28K)2026.03 | 115.8 | |
| CoIDOBackbone=LLaVA-v1.5-7B, Training Dataset=Vision-Flan-186K, Sampling Ratio=15% (28K)2026.03 | 103.6 | |
| ICONSBackbone=LLaVA-v1.5-7B, Training Dataset=Vision-Flan-186K, Sampling Ratio=15% (28K)2026.03 | 103.2 | |
| PreSelBackbone=LLaVA-v1.5-7B, Training Dataset=Vision-Flan-186K, Sampling Ratio=15% (28K)2026.03 | 100.6 | |
| PreSelReq. Inst.=15%, Sel. Inst.=28K2025.03 | 100.1 | |
| LLaVA-1.5-7BReq. Inst.=100%, Sel. Inst.=186K2025.03 | 100 | |
| Full-Data(186K)Backbone=LLaVA-v1.5-7B, Training Dataset=Vision-Flan-186K, Sampling Ratio=100%2026.03 | 100 | |
| COINCIDEReq. Inst.=100%, Sel. Inst.=28K2025.03 | 97.1 | |
| COINCIDEBackbone=LLaVA-v1.5-7B, Training Dataset=Vision-Flan-186K, Sampling Ratio=15% (28K)2026.03 | 97.1 | |
| XMASBackbone=LLaVA-v1.5-7B, Training Dataset=Vision-Flan-186K, Sampling Ratio=15% (28K)2026.03 | 96.9 | |
| RandomBackbone=LLaVA-v1.5-7B, Training Dataset=Vision-Flan-186K, Sampling Ratio=15% (28K)2026.03 | 96.7 | |
| RandomReq. Inst.=15%, Sel. Inst.=28K2025.03 | 96.1 | |
| EL2NReq. Inst.=100%, Sel. Inst.=28K2025.03 | 95.2 | |
| EL2NBackbone=LLaVA-v1.5-7B, Training Dataset=Vision-Flan-186K, Sampling Ratio=15% (28K)2026.03 | 95.2 | |
| Self-FilterBackbone=LLaVA-v1.5-7B, Training Dataset=Vision-Flan-186K, Sampling Ratio=15% (28K)2026.03 | 95 | |
| Self-FilterReq. Inst.=100%, Sel. Inst.=28K2025.03 | 94.2 | |
| CLIP-ScoreReq. Inst.=100%, Sel. Inst.=28K2025.03 | 92.2 | |
| CLIP-ScoreBackbone=LLaVA-v1.5-7B, Training Dataset=Vision-Flan-186K, Sampling Ratio=15% (28K)2026.03 | 92.2 | |
| TypiClustReq. Inst.=15%, Sel. Inst.=28K2025.03 | 92 | |
| TypiClustBackbone=LLaVA-v1.5-7B, Training Dataset=Vision-Flan-186K, Sampling Ratio=15% (28K)2026.03 | 92 | |
| IFDBackbone=LLaVA-v1.5-7B, Training Dataset=Vision-Flan-186K, Sampling Ratio=15% (28K)2026.03 | 91.6 | |
| IFDReq. Inst.=100%, Sel. Inst.=28K2025.03 | 90 | |
| PerplexityReq. Inst.=100%, Sel. Inst.=28K2025.03 | 88.2 |