Image2Code Generation on Synthetic Eval Dataset Graph
0.636BLEULlama-VL-TUG
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Llama-VL-TUG2026.01 | 0.636 | 63.8309 | 0.7667 | 86.2771 | -0.0717 | 0.7047 | |
| GPT-4o-mini2026.01 | 0.1672 | 18.5799 | 0.488 | 57.8477 | -0.088 | 0.5753 | |
| Gemma3-12B-Instruction-Tuned2026.01 | 0.137 | 17.157 | 0.4911 | 62.573 | 0.075 | 0.5705 | |
| Llama3.2-11B-Vision-Instruct2026.01 | 0.108 | 14.7753 | 0.4323 | 62.3889 | 0.2716 | 0.622 | |
| Qwen2.5-VL-7B-Instruct2026.01 | 0.0861 | 12.3727 | 0.4571 | 56.7718 | -0.0477 | 0.5189 | |
| MiniCPM-V-2-62026.01 | 0 | 2.4965 | 0.1444 | 7.8484 | -1.2013 | 0.0544 |