Juneberry improves the experience of machine learning experimentation by providing a framework for automating the training, evaluation, and comparison of multiple models against multiple datasets, thereby reducing errors and improving reproducibility.
This README describes how to use the Juneberry framework to execute machine learning tasks. Juneberry follows a (mostly) declarative programming model composed of sets of config files (dataset, model, and experiment configurations) and Python plugins for features such as model construction and transformation.
If you're looking for a slightly more in depth description of Juneberry see What Is Juneberry.
Other resources can be found at the Juneberry Home Page
The Getting Started documentation explains how to install Juneberry. It also includes a simple test command you can use to verify the installation.
The Workspace and Experiment Overview documentation contains information about the structure of the Juneberry workspace and how to organize experiments.
The Juneberry Basic Tutorial describes how to create a model, train the model, and run an experiment.
The Juneberry Configuration Guide describes various ways to configure Juneberry.
During normal use of Juneberry, you may encounter warning messages. The Known Warnings in Juneberry documentation contains information about known warning messages and what (if anything) should be done about them.
The vignettes directory contains detailed walkthroughs of various Juneberry tasks. The vignettes provide helpful examples of how to construct various Juneberry configuration files, including datasets, models, and experiments. A good start is Replicating a Classic Machine Learning Result with Juneberry.
Copyright 2022 Carnegie Mellon University. See LICENSE.txt file for license terms.