Scikit-learn

​​
​​
​​
​​
​​
​​
​See code examples on GitHub​
​See results in Neptune​
​Run example in Google Colab​
​Scikit-learn is an open-source machine learning framework commonly used for building predictive models. Neptune helps with keeping track of model training metadata.
What will you get with this integration?
With Neptune + Sklearn integration you can track your classifiers, regressors, and k-means clustering results, specifically:
log classifier and regressor parameters,
log pickled model,
log test predictions,
log test predictions probabilities,
log test scores,
log classifier and regressor visualizations, like confusion matrix, precision-recall chart, and feature importance chart,
log KMeans cluster labels and clustering visualizations,
log metadata including git summary info.
You can log many other run metadata like interactive charts, video, audio, and more.
See the full list.​
Where to start?
To get started with this integration follow the quickstart below (recommended as a first step).
You can also go to the demonstration of the functions that log regressor, classifier, or K-Means summary information to Neptune. Such summary includes parameters, pickled model, visualizations, and much more:
​classifier summary,
​regressor summary,
​k-means clustering summary.
Finally, if you want to log only specific information to Neptune you can make use of the convenience functions, such as:
get_estimator_parameters,
get_pickled_model,
create_prediction_error_chart.
Before you start
Make sure that:
You have Python 3.7+ in your system,
You are already registered user, so that you can log metadata to your private projects.
You have your NEPTUNE_API_TOKEN set to the environment variable: check this docs page.
Install neptune-client, scikit-learn, and neptune-sklearn
Depending on your operating system open a terminal or CMD and run this command. All required libraries are available via pip and conda:
pip
conda
1
pip install scikit-learn neptune-client neptune-sklearn
Copied!
1
conda install -c conda-forge neptune-client scikit-learn neptune-sklearn
Copied!
For more help see installing neptune-client.
This integration has been tested with neptune-client==0.9.18, scikit-learn==0.24.1, and neptune-sklearn==0.9.5.
You also need minimal familiarity with scikit-learn. Have a look at this scikit-learn guide to get started.
Quickstart
This quickstart will show you how to use Neptune with scikit-learn:
Create the first run in a project,
Log estimator parameters and scores, and
Explore results in the Neptune UI.
Step 0: Create and fit example estimator
Prepare a fitted estimator that will be further used to log its summary. Below snippet shows the idea:
1
parameters = {'n_estimators': 70,
2
              'max_depth': 7,
3
              'min_samples_split': 3}
4
​
5
estimator = RandomForestRegressor(**parameters)
6
X, y = load_boston(return_X_y=True)
7
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.20)
8
estimator.fit(X_train, y_train)
Copied!
Step 1: Initialize Neptune and create a new run
Add to your script (at the top):
1
import neptune.new as neptune
2
​
3
run = neptune.init(project='common/sklearn-integration',
4
                   api_token='ANONYMOUS')
Copied!
This opens a new “run” in Neptune to which you can log various objects.
You need to tell Neptune who you are and where you want to log things. To do that you specify:
project=my_workspace/my_project: your workspace name and project name,
api_token=YOUR_API_TOKEN : your Neptune API token.
If you configured your Neptune API token correctly, as described in this docs page, you can skip api_token argument.
If you are using.pyscripts for training Neptune will also log your training script automatically.
Step 2: Log parameters
To log parameters of your model training run you just need to pass them to the base_namespace of your choice.
1
run['parameters'] = parameters
Copied!
Step 3: Log estimator scores
Log scores on the test data under the base_namespace of your choice.
1
y_pred = estimator.predict(X_test)
2
​
3
run['scores/max_error'] = max_error(y_test, y_pred)
4
run['scores/mean_absolute_error'] = mean_absolute_error(y_test, y_pred)
5
run['scores/r2_score'] = r2_score(y_test, y_pred)
Copied!
Step 4: Stop logging
Once you are done logging, you should stop tracking the run using the stop() method. This is needed only while logging from a notebook environment. While logging through a script, Neptune automatically stops tracking once the script has completed execution.
1
run.stop()
Copied!
​See example in Neptune​
More Options
The Neptune - scikit-learn integration also lets you log regressor, classifier or K-Means summary information to Neptune. Such summary includes parameters, pickled model, visualizations and much more:
​classifier summary,
​regressor summary,
​k-means clustering summary.
Finally, if you want to log only specific information to Neptune you can make use of the convenience functions, such as:
get_estimator_parameters,
get_pickled_model,
create_prediction_error_chart.
Log classification summary
You can log a classification summary that includes:
​all classifier parameters,
​logged pickled model,
​logged test predictions,
​logged test predictions probabilities,
​logged test scores,
​logged classifier visualizations,
​logged metadata,
​logged code snapshot and git metadata.
​See example in Neptune​
Step 0: Create and fit example classifier
Prepare a fitted classifier that will be further used to log its summary. Below snippet shows the idea:
1
parameters = {'n_estimators': 120,
2
              'learning_rate': 0.12,
3
              'min_samples_split': 3,
4
              'min_samples_leaf': 2}
5
​
6
gbc = GradientBoostingClassifier(**parameters)
7
​
8
X, y = load_digits(return_X_y=True)
9
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)
10
​
11
gbc.fit(X_train, y_train)
Copied!
gbc object will be later used to log various metadata to the run.
Step 1: Initialize Neptune and create a run
Add the following snippet at the top of your script.
1
import neptune.new as neptune
2
​
3
run = neptune.init(project='common/sklearn-integration',
4
                   api_token='ANONYMOUS',
5
                   name='classification-example',
6
                   tags=['GradientBoostingClassifier', 'classification'])
Copied!
This creates a link to the run. Open the link in a new tab.
The run will currently be empty, but keep the window open. You will be able to see the estimator summary there.
When you create a run, Neptune will look for the .git directory in your project and get the last commit information saved.
If you are using.pyscripts for training Neptune will also log your training script automatically.
Step 2: Log classifier summary
Log classifier summary under the base_namespace of your choice.
1
import neptune.new.integrations.sklearn as npt_utils
2
​
3
run['cls_summary'] = npt_utils.create_classifier_summary(gbc, X_train, X_test, y_train, y_test)
Copied!
Step 3: Stop logging
Once you are done logging, you should stop tracking the run using the stop() method. This is needed only while logging from a notebook environment. While logging through a script, Neptune automatically stops tracking once the script has completed execution.
1
run.stop()
Copied!
Step 4: See results in Neptune
Once data is logged you can switch to the Neptune tab which you had opened previously to explore results. You can check:
​all classifier parameters,
​logged pickled model,
​logged test predictions,
​logged test predictions probabilities,
​logged test scores,
​logged classifier visualizations,
​logged metadata,
​logged code snapshot and git metadata.
Remember that you can try it out with zero setup:
​​
​​
​​
​​
​​
​​
​See code examples on GitHub​
​See results in Neptune​
​Run example in Google Colab​
Log regression summary
You can log a regression summary that includes:
​all regressor parameters,
​logged pickled model,
​logged test predictions,
​logged test scores,
​logged regressor visualizations,
​logged metadata,
​logged code snapshot and git metadata.
​See example in Neptune​
Step 0: Create and fit example regressor
Prepare a fitted regressor that will be further used to log its summary. The snippet below shows the idea:
1
parameters = {'n_estimators': 70,
2
              'max_depth': 7,
3
              'min_samples_split': 3}
4
​
5
rfr = RandomForestRegressor(**parameters)
6
​
7
X, y = load_boston(return_X_y=True)
8
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)
9
​
10
rfr.fit(X_train, y_train)
Copied!
rfr object will be later used to log various metadata to the run.
Step 1: Initialize Neptune and create a run
Add the following snippet at the top of your script.
1
import neptune.new as neptune
2
​
3
run = neptune.init(project='common/sklearn-integration',
4
                   api_token='ANONYMOUS',
5
                   name='regression-example',
6
                   tags=['RandomForestRegressor', 'regression'])
Copied!
This creates a link to the run. Open the link in a new tab.
The run will currently be empty, but keep the window open. You will be able to see the estimator summary there.
When you create a run, Neptune will look for the .git directory in your project and get the last commit information saved.
If you are using.pyscripts for training Neptune will also log your training script automatically.
Step 2: Log regressor summary
Log regressor summary under the base_namespace of your choice.
1
import neptune.new.integrations.sklearn as npt_utils
2
​
3
run['rfr_summary'] = npt_utils.create_regressor_summary(rfr, X_train, X_test, y_train, y_test)
Copied!
Step 3: Stop logging
Once you are done logging, you should stop tracking the run using the stop() method. This is needed only while logging from a notebook environment. While logging through a script, Neptune automatically stops tracking once the script has completed execution.
1
run.stop()
Copied!
Step 4: See results in Neptune
Once data is logged you can switch to the Neptune tab which you had opened previously to explore results. You can check:
​all regressor parameters,
​logged pickled model,
​logged test predictions,
​logged test scores,
​logged regressor visualizations,
​logged metadata,
​logged code snapshot and git metadata.
Remember that you can try it out with zero setup:
​​
​​
​​
​​
​​
​​
​See code examples on GitHub​
​See results in Neptune​
​Run example in Google Colab​
Log K-Means clustering summary
You can log a K-Means clustering summary that includes:
​all KMeans parameters,
​logged cluster labels,
​logged KMeans clustering visualizations,
​logged metadata,
​logged code snapshot and git metadata.
​See example in Neptune​
Step 0: Create K-Means clustering object and example data
Prepare K-Means object and example data that will be further used to log its summary. Below snippet shows the idea:
1
parameters = {'n_init': 11,
2
              'max_iter': 270}
3
​
4
km = KMeans(**parameters)
5
X, y = make_blobs(n_samples=579, n_features=17, centers=7, random_state=28743)
Copied!
Step 1: Initialize Neptune and create a run
Add the following snippet at the top of your script.
1
import neptune.new as neptune
2
​
3
run = neptune.init(project='common/sklearn-integration',
4
                   api_token='ANONYMOUS',
5
                   name='clustering-example',
6
                   tags=['KMeans', 'clustering'])
Copied!
This creates a link to the run. Open the link in a new tab.
The run will currently be empty, but keep the window open. You will be able to see the estimator summary there.
When you create a run, Neptune will look for the .git directory in your project and get the last commit information saved.
If you are using.py scripts for training Neptune will also log your training script automatically.
Step 2: Log KMeans clustering summary
Log K-Means clustering summary under the base_namespace of your choice.
1
import neptune.new.integrations.sklearn as npt_utils
2
​
3
run['kmeans_summary'] = npt_utils.create_kmeans_summary(km, X, n_clusters=17)
Copied!
Step 3: Stop logging
Once you are done logging, you should stop tracking the run using the stop() method. This is needed only while logging from a notebook environment. While logging through a script, Neptune automatically stops tracking once the script has completed execution.
1
run.stop()
Copied!
Step 4: See results in Neptune
Once data is logged you can switch to the Neptune tab which you had opened previously to explore results. You can check:
​all KMeans parameters,
​logged cluster labels,
​logged KMeans clustering visualizations,
​logged metadata,
​logged code snapshot and git metadata.
Remember that you can try it out with zero setup:
​​
​​
​​
​​
​​
​​
​See code examples on GitHub​
​See results in Neptune​
​Run example in Google Colab​
Other logging options
Log estimator parameters
You can choose to only log estimator parameters.
1
import neptune.new.integrations.sklearn as npt_utils
2
​
3
rfc = RandomForestClassifier()
4
​
5
run = neptune.init(project='common/sklearn-integration',
6
                   api_token='ANONYMOUS',
7
                   name='other-options')
8
​
9
run['estimator/parameters'] = npt_utils.get_estimator_params(rfc)
10
​
11
run.stop()
Copied!
​See example in Neptune​
Log model
You can choose to log a fitted model as a pickled file.
1
import neptune.new.integrations.sklearn as npt_utils
2
​
3
rfc = RandomForestClassifier()
4
rfc.fit(X, y)
5
​
6
run = neptune.init(project='common/sklearn-integration',
7
                   api_token='ANONYMOUS',
8
                   name='other-options')
9
​
10
run['estimator/pickled-model'] = npt_utils.get_pickled_model(rfc)
11
​
12
run.stop()
Copied!
​See example in Neptune​
Log confusion matrix
You can choose to log a confusion matrix chart.
1
import neptune.new.integrations.sklearn as npt_utils
2
​
3
rfc = RandomForestClassifier()
4
X, y = load_digits(return_X_y=True)
5
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.20, random_state=28743)
6
rfc.fit(X_train, y_train)
7
​
8
run = neptune.init(project='common/sklearn-integration',
9
                   api_token='ANONYMOUS',
10
                   name='other-options')
11
​
12
run['confusion-matrix'] = npt_utils.create_confusion_matrix_chart(rfc, X_train, X_test, y_train, y_test)
13
​
14
run.stop()
Copied!
Remember that you can try it out with zero setup:
​​
​​
​​
​​
​​
​​
​See code examples on GitHub​
​See results in Neptune​
​Run example in Google Colab​
How to ask for help?
Please visit the Getting help page. Everything regarding support is there.
Other integrations you may like
You may also like these two integrations:
Optuna
Plotly