Automated ML¶

automatically run all models and provid the scores for each model and also show the best.

models¶

AutoML's forecasting regression models assume that all features provided by the user are known into the future, at least up to the forecast horizon.

rerun failed experiment¶

from azureml.core import Workspace, Experiment

# Load the workspace and experiment objects
workspace = Workspace.from_config()
experiment = Experiment(workspace, "experiment-name')

# Get the failed runs
failed_runs = experiment.get_runs(type='Failed')

# Re-run the experiment
for failed_run in failed_runs:
    new_run = failed_run.re_run()

deploy using python¶

https://towardsdatascience.com/how-to-deploy-scikit-learn-models-to-azure-container-instances-a0a59d0d07a1

Setup Training Environment from terminal

conda create -n ml-10 python=3.10
conda activate ml-10
#install packahes
azureml-core==1.39
pandas==1.3.5
scikit-learn==0.23.2
mlflow==1.24.0
cloudpickle==2.0.0
psutil==5.9.0

get ml ws config

# config.json
  {
      "subscription_id": "subscription-id",
      "resource_group": "resource-group-name",
      "workspace_name": "workspace-name"
  }

experiment¶

import time
from azureml.core import Workspace, Experiment

automl_settings = {
    "name": f'auto_feature_engineering_{time.time()}',
    "task": "regression",
    "iterations": 10,
    "iteration_timeout_minutes": 10,    
    "max_cores_per_iteration": 1,
    "max_concurrent_iterations": 10,
    "primary_metric": 'r2_score',    
    "experiment_exit_score": 0.985,
    "debug_log": f'automl_errors{time.time()}.log',
    "verbosity": logging.ERROR,
}
# Local compute
automl_config = AutoMLConfig(
   preprocess=False,
   X=X_train,
   y=y_train,
   X_valid=X_valid,
   y_valid=y_valid,
   path=project_folder,
   **automl_settings,
)

# Training the model
experiment = Experiment(ws, experiment_name)
local_run = experiment.submit(automl_config, show_output=True)