Overview

This page provides steps for installing Comet's Python SDK and describes the main objects and tools it provides.

Install the SDK¶

There are several ways to install the Comet Python SDK in your system. The easiest way is using pip.

pip install comet_ml

conda install -c anaconda -c conda-forge -c comet_ml comet_ml

%pip install comet_ml

See next steps to start logging your data to Comet.

Experiment¶

Experiment is the core class of Comet. An Experiment represents a unit of measurable research that defines a single execution of code with some associated data for example training a model on a single set of hyperparameters.

Learn more about Comet's Experiment Management capabilities in our Experiment Management guides.

Artifact¶

Artifact objects live in a Comet Workspace and are identified by their name. Each artifact can have multiple versions, identified by their version string numbers. Typical artifact types could be dataset, image, training-data, validation-data, testing-data, and so on.

Learn more in the Artifact Overview.

Optimizer¶

The Comet Optimizer is used to dynamically find the best set of hyperparameter values that will minimize a Hyper Parameter Optimization tool (HPO) that can be used to maximize a particular metric. Optimizer provides many benefits over traditional hyperparameter optimizer search services because of its integration with Comet's Experiments. In addition, Comet's hyperparameter search has a powerful architecture for customizing your search or sweep. You can easily switch search algorithms, or perform phased searches.

Learn more in the Optimizer Overview.

Command-line utilities¶

The Comet installation package contains a collection of CLI utilities:

Learn more in Command-line utilities.

Extras¶

• Troubleshooting addresses errors and warnings you may see while working with the SDK.
• Comet Python SDK Releases lists all public-facing additions, changes, and deprecations of released comet_ml versions.