Logging a run
In this exercise, you will train a model using scikit-learn's Linear Regression
to predict profit from the Unicorn
dataset. You have created an experiment called Unicorn Sklearn Experiment
and started a new run. You will log metrics for r2_score
and parameters for n_jobs
as well as log the training code as an artifact.
The Linear Regression
model will be trained with n_jobs
parameter set to 1
. The r2_score
metric has been produced using the r2_score()
from scikit-learn based on y_pred
variable which came from predictions of X_test
.
model = LinearRegression(n_jobs=1)
model.fit(X_train, y_train)
y_pred = model.predict(X_test)
r2_score = r2_score(y_test, y_pred)
The mlflow
module as well as the LinearRegression
, train_test_split
, and metrics
modules from scikit-learn will be imported.
This exercise is part of the course
Introduction to MLflow
Exercise instructions
- Log the
r2_score
variable as a metric called"r2_score"
. - Log a parameter called
"n_jobs"
to the Tracking Server. - Log the
"train.py"
file as an artifact in the run.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Log the metric r2_score as "r2_score"
____.____("____", ____)
# Log parameter n_jobs as "n_jobs"
____.____("____", ____)
# Log the training code
____.____("train.py")