cli-automl-regression-task-hardware-perf.yml

$schema: https://azuremlsdk2.blob.core.windows.net/preview/0.0.1/autoMLJob.schema.json
type: automl

experiment_name: dpv2-cli-automl-regression-experiment
# name: dpv2-cli-regression-train-job-01
description: A regression job using hardware performance dataset

compute: azureml:cpu-cluster

task: regression
log_verbosity: debug
primary_metric: r2_score

target_column_name: "ERP"
n_cross_validations: 5

training_data:
  path: "./training-mltable-folder"
  type: mltable

limits:
  timeout_minutes: 180
  max_trials: 10
  max_concurrent_trials: 2
  trial_timeout_minutes: 20
  enable_early_termination: true

# This is example of custom featurization
# If you don't have this section, AutoML will default to automatic featurization
featurization:
  mode: custom
  blocked_transformers: [LabelEncoder]
  column_name_and_types:
    CHMIN: Categorical
  transformer_params:
    Imputer:
      - fields: ["CACH"]
        parameters:
          strategy: most_frequent
      - fields: ["PRP"]
        parameters:
          strategy: most_frequent

training:
  enable_model_explainability: true
  allowed_training_algorithms:
    - light_gbm
    - xg_boost_regressor