Multimodal Visual Perception on MMVP 1.0 (test)
60.73AccuracyFull Attention
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Full AttentionBackbone=LLaVA-1.5-13B2026.04 | 60.73 | 60.69 | |
| SinkRouterBackbone=LLaVA-1.5-13B2026.04 | 60.73 | 60.54 | |
| Full AttentionBackbone=LLaVA-1.5-7B2026.04 | 53.44 | 54.05 | |
| SinkRouterBackbone=LLaVA-1.5-7B2026.04 | 52.63 | 53.82 |