IExperimentTracker<T>

Interfaces · AiDotNet.Interfaces

Defines the contract for experiment tracking systems that log machine learning experiments.

Known implementations: ExperimentTrackerBase<T>

How It Works

An experiment tracker records information about machine learning experiments including parameters, metrics, and artifacts to enable reproducibility and comparison of different training runs.

For Beginners: Think of an experiment tracker as a lab notebook for machine learning. Just like a scientist records their experimental conditions and results, an experiment tracker logs all the details of your machine learning model training - what settings you used, how well it performed, and what models you created.

Key capabilities include:

  • Creating and managing experiments (groups of related training runs)
  • Logging hyperparameters (settings used for training)
  • Recording metrics (performance measurements over time)
  • Storing artifacts (models, plots, data files)
  • Comparing different training runs
  • Reproducing previous experiments

Why experiment tracking matters:

  • Helps you keep track of what you’ve tried
  • Makes it easy to reproduce good results
  • Enables comparison between different approaches
  • Provides audit trail for model development

Methods

Method Summary
IEnumerable<IExperiment> ListExperiments(string filter) Lists all experiments, optionally filtered by criteria.
IEnumerable<IExperimentRun<T>> ListRuns(string experimentId, string filter) Lists all runs in an experiment, optionally filtered by criteria.
IEnumerable<IExperimentRun<T>> SearchRuns(string filter, int maxResults) Searches for runs across all experiments based on criteria.
IExperiment GetExperiment(string experimentId) Gets an existing experiment by its ID.
IExperimentRun<T> GetRun(string runId) Gets an existing run by its ID.
IExperimentRun<T> StartRun(string experimentId, string runName, Dictionary<string, string> tags) Starts a new training run within an experiment.
string CreateExperiment(string name, string description, Dictionary<string, string> tags) Creates a new experiment to organize related training runs.
void DeleteExperiment(string experimentId) Deletes an experiment and all its associated runs.
void DeleteRun(string runId) Deletes a specific run.