Spatial Reasoning on WhatsUp-VLM COCOQA (test)
98.4Accuracy(IA)3
Evaluation Results
| Method | Links | |
|---|---|---|
| (IA)3Backbone=Qwen2.5-VL-3B, Trainable Parameters=524.K2026.06 | 98.4 | |
| (IA)3Backbone=LLaVA-7B, Trainable Parameters=664.K2026.06 | 96.6 | |
| ScAle (attn)Backbone=Qwen2.5-VL-3B, Trainable Parameters=5762026.06 | 95 | |
| ScAle (attn)Backbone=LLaVA-7B, Trainable Parameters=1.02K2026.06 | 91.5 | |
| FrozenBackbone=Qwen2.5-VL-3B2026.06 | 83.9 | |
| FrozenBackbone=LLaVA-7B2026.06 | 56.3 |