microsoft/machine-learning-collection

Aggregates Microsoft's ML/AI libraries across specialized domains—boosting (LightGBM, EBM), AutoML (NNI, FLAML), vision transformers, NLP models (DeBERTa, UniLM), graph neural networks, time-series forecasting, and responsible AI—with implementation examples and benchmark tools. Spans multiple frameworks (PyTorch, TensorFlow, CNTK) and integrates with platforms like Hugging Face and Azure ML. Includes sample code, workshops, and datasets alongside research prototypes for tasks like anomaly detection, neural architecture search, and causal inference.

451 stars. No commits in the last 6 months.