Skip to content

Latest commit

 

History

History
77 lines (56 loc) · 2.56 KB

README.md

File metadata and controls

77 lines (56 loc) · 2.56 KB
Raw

Neptune + scikit-learn integration

Experiment tracking for scikit-learn–trained models.

What will you get with this integration?

  • Log, organize, visualize, and compare ML experiments in a single place
  • Monitor model training live
  • Version and query production-ready models and associated metadata (e.g., datasets)
  • Collaborate with the team and across the organization

What will be logged to Neptune?

  • classifier and regressor parameters,
  • pickled model,
  • test predictions,
  • test predictions probabilities,
  • test scores,
  • classifier and regressor visualizations, like confusion matrix, precision-recall chart, and feature importance chart,
  • KMeans cluster labels and clustering visualizations,
  • metadata including git summary info,
  • other metadata

image

Resources

Example

# On the command line:
pip install neptune-sklearn

# In Python, prepare a fitted estimator
parameters = {
    "n_estimators": 70, "max_depth": 7, "min_samples_split": 3
}

estimator = ...
estimator.fit(X_train, y_train)

# Import Neptune and start a run
import neptune

run = neptune.init_run(
    project="common/sklearn-integration",
    api_token=neptune.ANONYMOUS_API_TOKEN,
)

# Log parameters and scores
run["parameters"] = parameters

y_pred = estimator.predict(X_test)

run["scores/max_error"] = max_error(y_test, y_pred)
run["scores/mean_absolute_error"] = mean_absolute_error(y_test, y_pred)
run["scores/r2_score"] = r2_score(y_test, y_pred)

# Stop the run
run.stop()

Support

If you got stuck or simply want to talk to us, here are your options:

  • Check our FAQ page
  • You can submit bug reports, feature requests, or contributions directly to the repository.
  • Chat! When in the Neptune application click on the blue message icon in the bottom-right corner and send a message. A real person will talk to you ASAP (typically very ASAP),
  • You can just shoot us an email at [email protected]