Image-to-LaTeX Table Conversion on Table2LaTeX simple (test)
83.5TEDSQwen2.5-VL-7B-CSPO
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Qwen2.5-VL-7B-CSPOModel Backbone=Qwen2.5-VL, Parameter Count=7B, Reinforcement Learning Strategy=CSPO2026.04 | 83.5 | 37.1 | 69 | 78.4 | 46 | |
| Qwen2.5-VL-7B-GRPOModel Backbone=Qwen2.5-VL, Parameter Count=7B, Reinforcement Learning Strategy=GRPO2026.04 | 83.1 | 34.7 | 64.7 | 79.8 | 44.8 | |
| Qwen2.5-VL-7B-VSGRPO*Model Backbone=Qwen2.5-VL, Parameter Count=7B, Reinforcement Learning Strategy=VSGRPO2026.04 | 82.9 | 33.9 | 66.9 | 80.2 | 42.3 | |
| Qwen2.5-VL-3B-CSPOModel Backbone=Qwen2.5-VL, Parameter Count=3B, Reinforcement Learning Strategy=CSPO2026.04 | 82.3 | 26 | 64.3 | 74.4 | 33.9 | |
| Qwen2.5-VL-3B-VSGRPO*Model Backbone=Qwen2.5-VL, Parameter Count=3B, Reinforcement Learning Strategy=VSGRPO2026.04 | 81.5 | 22.6 | 63.3 | 74.2 | 30.7 | |
| Qwen2.5-VL-3B-GRPOModel Backbone=Qwen2.5-VL, Parameter Count=3B, Reinforcement Learning Strategy=GRPO2026.04 | 81.5 | 24.4 | 58.3 | 75.6 | 32.1 | |
| Qwen2.5-VL-72BModel Backbone=Qwen2.5-VL, Parameter Count=72B2026.04 | 74.7 | 7.5 | 64.9 | 66.5 | 11.7 | |
| Gemini-2.5 FlashModel Backbone=Gemini-2.5 Flash2026.04 | 71.5 | 10.5 | 72 | 78.4 | 12.9 | |
| GPT-4oModel Backbone=GPT-4o2026.04 | 70.6 | 3.4 | 60.9 | 66.1 | 4.6 | |
| Qwen2.5-VL-7BModel Backbone=Qwen2.5-VL, Parameter Count=7B2026.04 | 66.4 | 4 | 54.6 | 58.7 | 5.4 | |
| Qwen2.5-VL-3BModel Backbone=Qwen2.5-VL, Parameter Count=3B2026.04 | 63.5 | 2 | 39.1 | 44.2 | 3 | |
| NougatModel Backbone=Nougat2026.04 | 27 | 0.4 | 20.6 | 2.8 | 5.7 |