All Tools
P
DataFreeOpen Source
PADDLETS
Deep time series forecasting, anomaly detection, and AutoTS on PaddlePaddle
Apache-2.0
ABOUT
Deep time series stacks usually glue datasets, transforms, and SOTA models by hand across forecasting, anomaly detection, and representation tasks. PaddleTS gives a single TSDataset API, AutoTS tuning, and PaddlePaddle models so teams train, backtest, and ensemble temporal pipelines without rewriting data loaders for each architecture.
INSTALL
pip install paddletsINTEGRATION GUIDE
1. Forecast multivariate demand or energy load with N-BEATS, N-HiTS, or Informer
2. Detect anomalies in industrial sensors with Autoencoder, VAE, or AnomalyTransformer
3. Run AutoTS hyperparameter search across models and datasets
4. Learn time series representations with TS2Vec or CoST for downstream classifiers
5. Build a PaddlePaddle pipeline from missing-value fills through ensemble forecasts
TAGS
pythontime-seriesforecastinganomaly-detectionpaddlepaddleautotsdeep-learning