Image-to-Text Style Generation on VLM Dataset
97.4AUCFor-Value
Evaluation results
| Method | Links | ||
|---|---|---|---|
| For-ValueVLM Model=Llama-3.2-11B-vision2025.08 | 97.4 | 99.7 | |
| For-ValueVLM Model=Qwen2.5-VL-3B-Instruct2025.08 | 89.5 | 91.6 | |
| EmbVLM Model=Qwen2.5-VL-3B-Instruct2025.08 | 56 | 19.8 | |
| EmbVLM Model=Llama-3.2-11B-vision2025.08 | 55.3 | 34 | |
| DataInfVLM Model=Qwen2.5-VL-3B-Instruct2025.08 | 52 | 76 | |
| HyperINFVLM Model=Qwen2.5-VL-3B-Instruct2025.08 | 51.6 | 86 | |
| Hessian-freeVLM Model=Qwen2.5-VL-3B-Instruct2025.08 | 51.5 | 79.9 | |
| Hessian-freeVLM Model=Llama-3.2-11B-vision2025.08 | 51.5 | 82.4 | |
| DataInfVLM Model=Llama-3.2-11B-vision2025.08 | 51.5 | 78.5 | |
| HyperINFVLM Model=Llama-3.2-11B-vision2025.08 | 49 | 82.1 |