Question 263143 in Azure Question posted in
The official documentation to get you started can be found here.

Azure – Getting empty experiment run using mlflow databricks

I am trying to run the following code

iter_max = 100
CC = 1
save_model = False
with mlflow.start_run(run_name=run_name, experiment_id=experiment.experiment_id, nested=True) as run:
    text_clf = Pipeline([
        ('vect', CountVectorizer()),
        ('tfidf', TfidfTransformer()),
        ('clf',LogisticRegression(max_iter=iter_max, C=CC))])
    text_clf.fit(X_train, y_train)
    run_id = mlflow.active_run().info.run_id`

But I get no model and no dataset…

enter image description here

I would like to get a model so I can register it and use it on databricks

I would like to get a model, register it and use it

Tags:

2

Answers


  1. Chosen as BEST ANSWER
    0

    thanks for the reply! it worked, but I'm facing a new challenge....I used a balanced training set, nevertheless, the model created by mlflow is overfitting towards 1 outcome. To see if I was doing something wrong, I trained the model on databricks and run it on memory, turns out it isn't overfitting so...idk what is going on..


  2. 0

    You need to register the model like below. Use below code.

    save_model = True
with mlflow.start_run(run_name=run_name, experiment_id=experiment_id, nested=True) as run:
    text_clf = Pipeline([
        ('vect', CountVectorizer()),
        ('tfidf', TfidfTransformer()),
        ('clf', LogisticRegression(max_iter=iter_max, C=CC))
    ])
    text_clf.fit(X_train, y_train)


    if save_model:
        mlflow.sklearn.log_model(text_clf, "model")

    mlflow.log_param("iter_max", iter_max)
    mlflow.log_param("CC", CC)

    run_id = mlflow.active_run().info.run_id

if save_model:
    mlflow.register_model(f"runs:/{run_id}/model", "My Registered Model")

    enter image description here

    Output:

    enter image description here

    You can use below code for prediction.

    import mlflow
logged_model = 'runs:/{run_id}/model'

# Load model as a PyFuncModel.
loaded_model = mlflow.pyfunc.load_model(logged_model)

# Predict on a Pandas DataFrame.
import pandas as pd
loaded_model.predict(pd.DataFrame(data))

    You can also see the model in models section.

    enter image description here

    Login or Signup to reply.
Please signup or login to give your own answer.

Back To Top
Search