Vision-Language Modeling on InfoVQA
84.6AccuracyZPPO
Evaluation Results
| Method | Links | |
|---|---|---|
| ZPPOModel Scale=2B, Prompt Replay Buffer Augmentation=true2026.06 | 84.6 | |
| GRPOModel Scale=2B, Prompt Replay Buffer Augmentation=true2026.06 | 83.1 | |
| GRPOModel Scale=2B, Prompt Replay Buffer Augmentation=false2026.06 | 82.4 | |
| Qwen3.5Model Scale=2B, Prompt Replay Buffer Augmentation=false2026.06 | 81.2 | |
| On-DistillModel Scale=2B, Prompt Replay Buffer Augmentation=false2026.06 | 81.1 | |
| On-DistillModel Scale=2B, Prompt Replay Buffer Augmentation=true2026.06 | 80.9 | |
| Off-DistillModel Scale=2B, Prompt Replay Buffer Augmentation=true2026.06 | 80.4 | |
| Off-DistillModel Scale=2B, Prompt Replay Buffer Augmentation=false2026.06 | 80.2 | |
| ZPPOModel Scale=0.8B, Prompt Replay Buffer Augmentation=true2026.06 | 75.3 | |
| GRPOModel Scale=0.8B, Prompt Replay Buffer Augmentation=true2026.06 | 72.4 | |
| GRPOModel Scale=0.8B, Prompt Replay Buffer Augmentation=false2026.06 | 71 | |
| Qwen3.5Model Scale=0.8B, Prompt Replay Buffer Augmentation=false2026.06 | 68.6 | |
| On-DistillModel Scale=0.8B, Prompt Replay Buffer Augmentation=true2026.06 | 67.6 | |
| On-DistillModel Scale=0.8B, Prompt Replay Buffer Augmentation=false2026.06 | 67.1 | |
| Off-DistillModel Scale=0.8B, Prompt Replay Buffer Augmentation=true2026.06 | 66.8 | |
| Off-DistillModel Scale=0.8B, Prompt Replay Buffer Augmentation=false2026.06 | 66 |