Description2Code Generation on Synthetic Eval Dataset Graph 1.0
0.9092BLEUGPT-4o-mini
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| GPT-4o-mini2026.01 | 0.9092 | 90.9247 | 0.9076 | 97.3794 | 0.9716 | 0.999 | |
| Llama-VL-TUG2026.01 | 0.8464 | 84.6445 | 0.8541 | 93.3428 | 0.7006 | 0.9403 | |
| Gemma3-12B-Instruction-TunedModel Size=12B, Variant=Instruction-Tuned2026.01 | 0.159 | 20.8206 | 0.5247 | 73.1211 | 0.2322 | 0.7391 | |
| Llama3.2-11B-Vision-InstructModel Size=11B, Variant=Vision-Instruct2026.01 | 0.1428 | 17.8969 | 0.5116 | 66.2555 | 0.1393 | 0.6608 | |
| Qwen2.5-VL-7B-InstructModel Size=7B, Variant=Instruct2026.01 | 0.0563 | 9.7717 | 0.4591 | 57.6757 | -0.0484 | 0.5707 | |
| MiniCPM-V-2-62026.01 | 0.0115 | 5.3132 | 0.3373 | 50.4762 | -0.3247 | 0.4465 |