Nixtla home pagelight logodark logo
    Local
    Callbacks
  • TimeGPT
  • StatsForecast
  • MLForecast
  • NeuralForecast
  • HierarchicalForecast
  • UtilsForecast
  • DatasetsForecast
  • CoreForecast
    • mlforecast
    Getting Started
    • Install
    • Quick start (local)
    • Quick start (distributed)
    • End to end walkthrough
    How-to guides
    • Exogenous features
    • Lag transformations
    • Hyperparameter optimization
    • Using scikit-learn pipelines
    • Sample weights
    • Cross validation
    • Probabilistic forecasting
    • Target transformations
    • Analyzing the trained models
    • MLflow
    • Transforming exogenous features
    • Custom training
    • Training with numpy arrays
    • One model per step
    • Custom date features
    • Predict callbacks
    • Predicting a subset of ids
    • Transfer Learning
    Tutorials
    • Electricity Load Forecast
    • Detect Demand Peaks
    • Prediction intervals
    API Reference
    • Local
      • MLForecast
      • Auto
      • LightGBMCV
      • Optimization
      • Utils
      • Core
      • Target transforms
      • Lag transforms
      • Feature engineering
      • Callbacks
    • Distributed
    Local

    Callbacks


    source

    ​
    SaveFeatures

     SaveFeatures ()
    

    Saves the features in every timestamp.


    source

    ​
    SaveFeatures.get_features

     SaveFeatures.get_features (with_step:bool=False)
    

    Retrieves the input features for every timestep

    TypeDefaultDetails
    with_stepboolFalseAdd a column indicating the step
    ReturnsUnionDataFrame with input features

    Was this page helpful?

    Feature engineeringDistributed Forecast
    githubslacktwitter
    Powered by Mintlify
    On this page
    • SaveFeatures
    • SaveFeatures.get_features