Skip to content

PaddleTS 1.0.0 Release Note EN

kehuo edited this page Nov 3, 2022 · 13 revisions

Release Highlights for PaddleTS 1.0.0

New Features

  • Support time series model for anomaly detection. PaddleTS provides a collection of abilities such as anomaly detection metrics, visualization analysis, third-party library integration, etc. Meanwhile, PaddleTS involves a set of state-of-the-art models for anomaly detection, see below New Models list.
  • Support joint training by multiple time series datasets. Developers are now able to build time series models by using a list of time series datasets associated with related identifiers such as device ids.
  • Support time series ensemble model.
  • Representation models is now able to solve time series forecasting tasks.
  • RNN now supports categorical type features and static covariates.
  • TSDataset.load_from_dataframe() now requires significant less time than before.
  • More time series analysis.

New Models

Time series anomaly detection is now availalbe in PaddleTS, below are newly released models:

CPU & GPU Capable Docker Images for PaddleTS 1.0.0

  • registry.baidubce.com/paddlets/paddlets:1.0.0
  • registry.baidubce.com/paddlets/paddlets:1.0.0-gpu-paddle2.3.0-cuda11.2-cudnn8
  • registry.baidubce.com/paddlets/paddlets:1.0.0-gpu-paddle2.3.1-cuda11.2-cudnn8
  • registry.baidubce.com/paddlets/paddlets:1.0.0-gpu-paddle2.3.2-cuda11.2-cudnn8

Changelog for PaddleTS 1.0.0

Legend for changelogs:

Feature: new features that PaddleTS does not support previously.

Optimization: an existing feature now either requires less computation or memory.

API Change: some changes that require you to change your code.

Fix: some features that previously did not work as stated in documents, or for reasonable expectations that should now work.

Changelog

  • API Change autots.AutoTS::fit() now allows the train_tsdataset and valid_tsdataset parameters to be either one TSDataset or a list of TSDataset instance(s) to support multiple time series joint training.
  • Feature PaddleTS introduces AutoEncoder, VAE, and Anomaly Transformer deep time series models for anomaly detection.
  • Feature Third-party model integration module make_ml_model now supports building anomaly detection models based on third-party library pyod.
  • Feature RNN now supports categorical type features. It also supports using static covariates as features.
  • API Change make_ml_model module changed from models.forecasting.ml.ml_model_wrapper::make_ml_model to models.ml_model_wrapper::make_ml_model.
  • API Change paddle_base_impl.PaddleBaseModelImpl::fit() now allows the train_tsdataset and valid_tsdataset parameters to be either one TSDataset or a list of TSDataset instance(s) to support multiple time series joint training.
  • API Change pipeline.Pipeline::fit() now allows the train_tsdataset and.valid_tsdataset parameters to be either one TSDataset or a list of TSDataset instance(s) to support multiple time series joint training.
  • API Change The difference, fill, ksigma, lag, statistical, time_feature transformations now delete fit and transform functions, and add fit_one and transform_one functions to perform time series data transformation based on one TSDataset instance.