Databricks-ML-Assoc Model Development — All Questions With Answers

{
  "run_id": "d7a8b9c0e1f2",
  "artifact_uri": "dbfs:/databricks/mlflow/789/d7a8b9c0e1f2/artifacts",
  "params": {"learning_rate": "0.01", "n_estimators": "100"},
  "tags": {"mlflow.source.name": "train_script.py", "mlflow.user": "admin"}
}

MLflow error: 'Invalid parameter: model_name must be a valid string, received: None'

Error: mlflow.exceptions.RestException: INVALID_PARAMETER_VALUE: Model signature is missing. Please provide a signature when logging the model.

{
  "model_name": "fraud_detection_model",
  "artifact_path": "model",
  "pip_requirements": ["scikit-learn==1.2.2", "pandas==2.0.0"],
  "metadata": {
    "framework": "sklearn",
    "python_version": "3.10.12"
  }
}

log_data = {'learning_rate': 0.01, 'max_depth': 5}
mlflow.log_params(log_data)
# Error: TypeError: 'dict' object is not callable

log_model_config = {
  "artifact_path": "model",
  "signature": infer_signature(X_train, y_pred),
  "input_example": X_train.iloc[:5]
}
mlflow.sklearn.log_model(model, **log_model_config)

log_model(
  model=model,
  artifact_path="model",
  signature=model_signature,
  input_example=input_data
)

Error: 'MLflowException: The model signature is missing. Please provide a signature to ensure schema validation.'

log_model(model=model, artifact_path='model', signature=sig, input_example=input_ex)

MLflow error: 'Cannot find module: sklearn.utils.extmath'

import mlflow
mlflow.sklearn.log_model(model, 'model', registered_model_name='my_model')