Image2Description on Synthetic Eval Dataset Packet
0.6302BLEULlama-VL-TUG
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Llama-VL-TUG2026.01 | 0.6302 | 63.024 | 0.7891 | 81.736 | 0.4825 | 0.7232 | |
| Gemma3-12B-Instruction-Tuned2026.01 | 0.4553 | 45.5292 | 0.6432 | 70.5237 | 0.2239 | 0.6365 | |
| Llama3.2-11B-Vision-Instruct2026.01 | 0.4044 | 40.441 | 0.6921 | 74.9078 | 0.2912 | 0.5578 | |
| MiniCPM-V-2-62026.01 | 0.4044 | 40.441 | 0.6921 | 74.9078 | 0.2912 | 0.5578 | |
| GPT-4o-mini2026.01 | 0.3828 | 38.2805 | 0.6982 | 75.008 | 0.1336 | 0.547 | |
| Qwen2.5-VL-7B-Instruct2026.01 | 0.3441 | 34.413 | 0.6457 | 71.2078 | 0.1079 | 0.5171 |