Efficiency Metrics for Time Series Forecasting on Weather
0.021Inference TimeTSMixer
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| TSMixerSamples per inference=2562024.12 | 0.021 | — | — | — | — | — | |
| FSMLPSamples per inference=2562024.12 | 0.021 | — | — | — | — | — | |
| FreTSSamples per inference=2562024.12 | 0.023 | — | — | — | — | — | |
| iTransformerSamples per inference=2562024.12 | 0.061 | — | — | — | — | — | |
| PatchTSTSamples per inference=2562024.12 | 0.134 | — | — | — | — | — | |
| DLinearInput length=96, Prediction length=96, Batch size=16, Number of layers=1, FLOPs Unit=G, T.T. Unit=ms/iter, I.M. Unit=MB2026.01 | 0.2 | 0.02 | 0.0004 | 1.38 | 0.02 | 9.06 | |
| TimesNetSamples per inference=2562024.12 | 0.244 | — | — | — | — | — | |
| FreTSInput length=96, Prediction length=96, Batch size=16, Number of layers=2, FLOPs Unit=G, T.T. Unit=ms/iter, I.M. Unit=MB2026.01 | 0.48 | 0.42 | 0.01 | 3.53 | 0.04 | 24.37 | |
| AutoformerSamples per inference=2562024.12 | 0.659 | — | — | — | — | — | |
| OLinearInput length=96, Prediction length=96, Batch size=16, Number of layers=2, FLOPs Unit=G, T.T. Unit=ms/iter, I.M. Unit=MB2026.01 | 1.18 | 4.52 | 0.1 | 7.47 | 0.21 | 39.26 | |
| vLinearInput length=96, Prediction length=96, Batch size=16, Number of layers=2, FLOPs Unit=G, T.T. Unit=ms/iter, I.M. Unit=MB2026.01 | 1.43 | 2.77 | 0.06 | 9.82 | 0.19 | 27.6 | |
| iTrans.Input length=96, Prediction length=96, Batch size=16, Number of layers=2, FLOPs Unit=G, T.T. Unit=ms/iter, I.M. Unit=MB2026.01 | 1.66 | 4.83 | 0.1 | 9.45 | 0.22 | 34.88 | |
| PatchTSTInput length=96, Prediction length=96, Batch size=16, Number of layers=2, FLOPs Unit=G, T.T. Unit=ms/iter, I.M. Unit=MB2026.01 | 2.96 | 3.76 | 0.82 | 9.89 | 0.27 | 90.78 | |
| LeddamInput length=96, Prediction length=96, Batch size=16, Number of layers=2, FLOPs Unit=G, T.T. Unit=ms/iter, I.M. Unit=MB2026.01 | 3.15 | 8.56 | 0.34 | 27.55 | 0.21 | 70.06 | |
| CARDInput length=96, Prediction length=96, Batch size=16, Number of layers=2, FLOPs Unit=G, T.T. Unit=ms/iter, I.M. Unit=MB2026.01 | 4.11 | 0.03 | 0.004 | 28.69 | 0.05 | 14.45 | |
| TimeMix.Input length=96, Prediction length=96, Batch size=16, Number of layers=2, FLOPs Unit=G, T.T. Unit=ms/iter, I.M. Unit=MB2026.01 | 5.1 | 0.1 | 0.05 | 33.3 | 0.16 | 36.03 | |
| Fredf.m.Input length=96, Prediction length=96, Batch size=16, Number of layers=2, FLOPs Unit=G, T.T. Unit=ms/iter, I.M. Unit=MB2026.01 | 7.19 | 0.5 | 0.01 | 37.85 | 0.05 | 18.28 | |
| TimeMix.++Input length=96, Prediction length=96, Batch size=16, Number of layers=2, FLOPs Unit=G, T.T. Unit=ms/iter, I.M. Unit=MB2026.01 | 30.18 | 2.37 | 8.36 | 102.04 | 0.86 | 316.49 | |
| TimesNetInput length=96, Prediction length=96, Batch size=16, Number of layers=2, FLOPs Unit=G, T.T. Unit=ms/iter, I.M. Unit=MB2026.01 | 156.27 | 299.97 | 291.21 | 505.88 | 5.8 | 1,397.18 |