Allie is a framework for building machine learning models from audio, text, image, video, or .CSV files.

Intended for both beginners and experts, Allie is designed to be easy-to-use for rapid prototyping and easy-to-extend to your desired modeling strategy.

Here are some things that Allie can do:

• find and download datasets (for quick experiments)
• annotate, clean, and/or augment audio, text, image, or video datasets (to prepare data for modeling)
• featurize files using a standard format (via audio, text, image, video, or csv featurizers)
• transform features (via scikit-learn preprocessing techniques)
• visualize featurized datasets (via yellowbrick, scikit-learn, and matplotlib libraries)
• train classification or regression machine learning models (via tpot, autokeras, autopytorch, ludwig, and 15+ other training scripts)
• make predictions from machine learning models (with all models trained in ./models directory)
• export data in .CSV file formats (for repeatable machine learning experiments across frameworks)
• compress machine learning models for deployment (including repositories with readmes)