Skip to content

Latest commit

 

History

History
38 lines (36 loc) · 3.44 KB

supported_xgboost_parameters_python.md

File metadata and controls

38 lines (36 loc) · 3.44 KB
Raw

Supported Parameters

This is a description of all the parameters available when you are running examples in this repo:

  1. All xgboost parameters are supported.
    • Please use the camelCase, e.g., --treeMethod=gpu_hist.
    • lambda is replaced with lambda_, because lambda is a keyword in Python.
  2. --mainClass=[app class]: The entry class of the application to be started. Available value is one of the below classes.
    • com.nvidia.spark.examples.agaricus.cpu_main
    • com.nvidia.spark.examples.agaricus.gpu_main
    • com.nvidia.spark.examples.mortgage.cpu_main
    • com.nvidia.spark.examples.mortgage.gpu_main
    • com.nvidia.spark.examples.taxi.cpu_main
    • com.nvidia.spark.examples.taxi.gpu_main
    • com.nvidia.spark.examples.mortgage.etl_main
    • com.nvidia.spark.examples.taxi.etl_main
  3. --format=[csv|parquet|orc]: The format of the data for training/transforming, now only supports 'csv', 'parquet' and 'orc'. Required.
  4. --mode=[all|train|transform]. The behavior of the XGBoost application (meaning CPUMain and GPUMain), default is 'all' if not specified.
    • all: Do both training and transforming, will save model to 'modelPath' if specified
    • train: Do training only, will save model to 'modelPath' if specified.
    • transform: Do transforming only, 'modelPath' is required to locate the model data to be loaded.
  5. --dataPath=[prefix]::[path]: Path to input data file(s), or path to output data files. Use it repeatly to specify multiple data paths.
    • --dataPath=train::[path]: Path to the training data file(s), required when mode is NOT 'transform'.
    • --dataPath=trans::[path]: Path to the transforming data file(s), required when mode is NOT 'train'.
    • --dataPath=eval::[path]: Path to the evaluation data file(s) for training. Optional.
    • --dataPath=rawTrain::[path]: Path to the raw data files for training, only used by taxi/CPUMain, taxi/GPUMain now to support E2E train.
    • --dataPath=rawTrans::[path]: Path to the raw data files for transforming, only used by taxi/CPUMain, taxi/GPUMain now to support E2E tranformation.
    • --dataPath=rawEval::[path]: Path to the raw data files being used as evaluation data for training. Optional.
    • --dataPath=raw::[path]: Path to the raw data files to be transformed by taxi/ETLMain.
    • --dataPath=perf::[path],-dataPath=acq::[path]: Paths to the raw data files to be transformed by mortgage/ETLMain.
    • --dataPath=out::: Path where to place the output data files for both mortgage/ETLMain and taxi/ETLMain.
  6. --modelPath=[path]: Path to save model after training, or where to load model for transforming only. Required only when mode is 'transform'.
  7. --overwrite=[true|false]: Whether to overwrite the current model data under 'modelPath'. Default is false. You may need to set to true to avoid IOException when saving the model to a path already exists.
  8. --hasHeader=[true|false]: Indicate whether the csv file has header.
  9. --numRows=[int value]: The number of the rows to be shown after transforming done. Default is 5.
  10. --showFeatures=[true|false]: Whether to show the features columns after transforming done. Default is true.
  11. --dataRatios=[trainRatio:transformRatio]: The ratios of data for train and transform, then the ratio for evaluation is (100-train-test). Default is 80:20, no evaluation. This is only used by taxi/ETLMain now to generate the output data.