High-resolution Visual Search on V*
86.91Top-1 AccuracyEntropy-Gradient Grounding
Evaluation Results
| Method | Links | |
|---|---|---|
| Entropy-Gradient GroundingTraining Approach=Training-free, VLM=Qwen 2.5 VL2026.04 | 86.91 | |
| TEVA 7BTraining Approach=Training-based, VLM=TEVA2026.04 | 77.1 | |
| SEALTraining Approach=Training-based, VLM=Llava 1.52026.04 | 75.3 | |
| Entropy-Gradient GroundingTraining Approach=Training-free, VLM=LLaVA 1.62026.04 | 73.3 | |
| Base modelTraining Approach=Training-free, VLM=Qwen 2.5 VL2026.04 | 73.3 | |
| Entropy-Gradient GroundingTraining Approach=Training-free, VLM=InternVL 3.52026.04 | 67.53 | |
| ViCropTraining Approach=Training-free, VLM=LLaVA 1.62026.04 | 61.78 | |
| TEVA 3BTraining Approach=Training-based, VLM=TEVA2026.04 | 61.6 | |
| Base modelTraining Approach=Training-free, VLM=LLaVA 1.62026.04 | 57.59 | |
| Entropy-Gradient GroundingTraining Approach=Training-free, VLM=LLaVA 1.52026.04 | 56.02 | |
| ViCropTraining Approach=Training-free, VLM=LLaVA 1.52026.04 | 47.64 | |
| Base modelTraining Approach=Training-free, VLM=InternVL 3.52026.04 | 47.64 | |
| Base modelTraining Approach=Training-free, VLM=LLaVA 1.52026.04 | 46.07 |