igel

Why we included this project

igel is for people who want a working machine learning model without writing Python. You describe the task in a YAML or JSON config file, point it at your data, and it handles the preprocessing, model selection, and training. It covers regression, classification, and clustering. Since it wraps scikit-learn and related libraries, the models it produces are familiar and easy to inspect, and it can run cross-validation, hyperparameter search, and parallel jobs across multiple processes. The auto-ML mode is the most useful part for quick experiments: give it a folder of raw images or a text dataset and it will handle preprocessing and pick a model on its own. For teams that want a fast draft model or a way to let non-engineers run experiments, the command-line interface beats assembling a training script by hand.