Visual Question Answering on ChartQA EvoCharts (test)
91.07AccuracySFT
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| SFTModel=IVL-8B2026.06 | 91.07 | 60.57 | 41.5 | 51 | 2.78 | |
| Att-CoTModel=IVL-8B2026.06 | 90.85 | 61.72 | 19.2 | 69 | 2.66 | |
| Att-CoTModel=Gemma-12B2026.06 | 90.72 | 57.22 | 28.2 | 74 | 0.32 | |
| SFTModel=Gemma-12B2026.06 | 89.45 | 55.21 | 48.2 | 75 | 1.63 | |
| LLaVA-RModel=Gemma-12B2026.06 | 88.57 | 53.69 | 38.7 | 65 | — | |
| LLaVA-RModel=IVL-8B2026.06 | 88.24 | 57.48 | 32.4 | 50 | — | |
| LLaVA-CoTModel=Gemma-12B2026.06 | 87.24 | 52.46 | 44.4 | 63 | — | |
| Att-CoTModel=IVL-4B2026.06 | 86.84 | 59.12 | 50.2 | 72 | 3.05 | |
| Att-CoTModel=QVL-7B2026.06 | 86.82 | 58.55 | 31.5 | 83 | 2.15 | |
| SFTModel=QVL-7B2026.06 | 86.61 | 56.25 | 57.2 | 60 | 2.05 | |
| SFTModel=IVL-4B2026.06 | 86.39 | 57.69 | 92.5 | 59 | 3.85 | |
| LLaVA-CoTModel=IVL-8B2026.06 | 85.66 | 55.21 | 40.7 | 43 | — | |
| LLaVA-RModel=IVL-4B2026.06 | 85.64 | 57.22 | 76.2 | 58 | — | |
| LLaVA-CoTModel=IVL-4B2026.06 | 84.81 | 55.21 | 84.3 | 44 | — | |
| LLaVA-RModel=QVL-7B2026.06 | 84.81 | 57.22 | 47.6 | 57 | — | |
| Att-CoTModel=QVL-3B2026.06 | 83.45 | 50.24 | 64.2 | 71 | 2.14 | |
| Att-CoTModel=Gemma-4B2026.06 | 82.54 | 48.22 | 62.4 | 78 | 1.04 | |
| SFTModel=QVL-3B2026.06 | 81.25 | 45.53 | 92.5 | 62 | 4.32 | |
| SFTModel=Gemma-4B2026.06 | 80.44 | 44.92 | 89.4 | 67 | 3.64 | |
| LLaVA-CoTModel=QVL-7B2026.06 | 80.35 | 55.21 | 56.1 | 48 | — | |
| LLaVA-RModel=QVL-3B2026.06 | 80.15 | 46.11 | 78.4 | 63 | — | |
| LLaVA-RModel=Gemma-4B2026.06 | 79.35 | 45.68 | 85.5 | 64 | — | |
| LLaVA-CoTModel=Gemma-4B2026.06 | 78.62 | 40.84 | 84.6 | 59 | — | |
| LLaVA-CoTModel=QVL-3B2026.06 | 78.33 | 42.84 | 89.2 | 55 | — |