Video Prediction on Cityscapes (Semantic and Generative Metrics)
69.76mIoU (A)Re2Pix (Stage-1)
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Re2Pix (Stage-1)Stage=1, Model parameters=1.1B2026.04 | 69.76 | 69.66 | 88.03 | 0.1262 | — | — | |
| Re2PixModel parameters=1.1B2026.04 | 63.53 | 62.29 | 85.72 | 0.1413 | 9.9 | 52.66 | |
| Baseline-LargeModel parameters=1.5B2026.04 | 61.69 | 59.65 | 85.43 | 0.1453 | 11.99 | 56.81 | |
| Baseline w/ REPAModel parameters=792M2026.04 | 61.45 | 59.63 | 85.35 | 0.1465 | 12.34 | 55.15 | |
| Baseline w/ VideoREPAModel parameters=792M2026.04 | 60.98 | 58.61 | 85.22 | 0.1466 | 12.95 | 59.03 | |
| BaselineModel parameters=782M2026.04 | 60.55 | 57.64 | 85.15 | 0.146 | 12.86 | 60.7 |