Text Image Reconstruction on StyledTextSynth
84.92AccuracyUniTok
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| UniTokResolution=High Resolution (1024), Data=1.28B, AR-Free=false2026.06 | 84.92 | 87.69 | 99.73 | 0.01 | |
| ChameleonVQ + RDAResolution=High Resolution (1024), Data=5M, AR-Free=true2026.06 | 73.74 | 9.99 | 77.03 | 9.7 | |
| LlamagenVQ + RDAResolution=High Resolution (1024), Data=5M, AR-Free=true2026.06 | 69.49 | 74.22 | 77.03 | 9.7 | |
| UniTokResolution=Low Resolution (512), Data=1.28B, AR-Free=false2026.06 | 62.11 | 68.54 | 92.43 | 11.32 | |
| LlamagenVQResolution=High Resolution (1024), Data=50M, AR-Free=false2026.06 | 55.58 | 60.51 | 75.58 | 8.95 | |
| ChameleonVQResolution=High Resolution (1024), Data=-, AR-Free=false2026.06 | 39.01 | 43.82 | 71.63 | 10.86 | |
| LlamagenVQ + RDAResolution=Low Resolution (512), Data=5M, AR-Free=true2026.06 | 13.57 | 18.18 | 67.67 | 22.23 | |
| ChameleonVQ + RDAResolution=Low Resolution (512), Data=5M, AR-Free=true2026.06 | 7.8 | 10.53 | 65.13 | 24.9 | |
| TA-TokResolution=High Resolution (1024), Data=200M, AR-Free=false2026.06 | 6.69 | 9.24 | 42.7 | 46.45 | |
| ChameleonVQResolution=Low Resolution (512), Data=-, AR-Free=false2026.06 | 5.52 | 7.44 | 63.81 | 26.89 | |
| LlamagenVQResolution=Low Resolution (512), Data=50M, AR-Free=false2026.06 | 4.12 | 6.55 | 66.33 | 22.39 | |
| TA-TokResolution=Low Resolution (512), Data=200M, AR-Free=false2026.06 | 2.25 | 3.93 | 44.35 | 47.73 |