Skip to content

Files

Latest commit

 

History

History
15 lines (10 loc) · 625 Bytes

MLLearner_en.md

File metadata and controls

15 lines (10 loc) · 625 Bytes

MLLearner

MLLearner is a core class for model training. In principle, any core model-training logic in Angel should be implemented and called by this class. MLLearner is train's core class.

Function

  • Reading DataBlock constantly, training to obtain the MLModel

Core Interface

  1. train
    • Definition: def train(train: DataBlock[LabeledData], vali: DataBlock[LabeledData]):MLModel
    • Function: train the model on the training data using a specific algorithm
    • Parameters: train: DataBlock[LabeledData] training dataset; vali: DataBlock[LabeledData] validation dataset
    • Return value: MLModel