Introduction
Experiment tracking in machine learning is the process of saving all experiment metadata in one central place (database or a repository). This includes model hyperparameters, model performance metrics, run logs, model artifacts, data artifacts, etc.
Experiment logging can be implemented in many different ways. It can be something as simple as spreadsheets (nobody does that these days!), or leverage GitHub to track experiments.
The easiest way to achieve experiment logging is by either using an open-source library / framework such as MLFlow or buy some enterprise tool platform offering these capabilities such as Weights & Biases, Comet, etc. This article lists down some really useful experiment logging tools for data scientists.
This is not a sponsored post.
MLFlow
MLflow is an open source platform to manage the machine learning lifecycle, including experimentation, reproducibility, deployment, and a central model registry. MLflow currently offers four components:
- Tracking experiments to record and compare parameters…
