Urban Sensing on Beijing AQI (test)
78.84Top-1 AccuracyChronosBase
Evaluation results
| Method | Links | |
|---|---|---|
| ChronosBase# Parameters=200M, # Time Points=12.8B, Evaluation Protocol=linear-probing2026.03 | 78.84 | |
| SLIPBase# Parameters=120M, # Time Points=1.7B, Evaluation Protocol=linear-probing2026.03 | 77.47 | |
| PatchTST (Supervised Learning)# Parameters=0.9M, Evaluation Protocol=Supervised Learning2026.03 | 76.11 | |
| Normwear# Parameters=136M, # Time Points=0.31B, Evaluation Protocol=linear-probing2026.03 | 74.74 | |
| Statistical MLEvaluation Protocol=linear-probing2026.03 | 68.6 | |
| SimMTM# Parameters=0.5M, # Time Points=1.72M, Evaluation Protocol=linear-probing2026.03 | 67.23 | |
| SundialBase# Parameters=128M, # Time Points=1032B, Evaluation Protocol=linear-probing2026.03 | 66.55 | |
| ChatTS# Parameters=8B, # Time Points=0.23B, Evaluation Protocol=linear-probing2026.03 | 59.73 | |
| Chronos2# Parameters=120M, # Time Points=242.8B, Evaluation Protocol=linear-probing2026.03 | 49.49 | |
| TF-C# Parameters=12M, # Time Points=1.72M, Evaluation Protocol=linear-probing2026.03 | 46.08 |