📄️ Model Training Guide
This guide explains how to retrieve datasets for training inside Michelangelo AI workflows and how to optionally scale training using RayTask and the Lightning Trainer SDK.
📄️ Distributed Training
This guide is for ML engineers and data scientists who have already trained a model with the SDK and now need to scale it across multiple workers or GPUs — by the end, you'll know exactly which LightningTrainerParam knob to reach for, whether that's picking a distributed strategy, warm-starting from existing weights, or recovering an interrupted run. If you haven't run a training task yet, start with the Model Training Guide — this page picks up where its Lightning section leaves off.
📄️ Model Registry Guide
Save, version, and manage trained models using Michelangelo AI's model packaging system.
📄️ Deploy a Model
Serve a trained model from Michelangelo AI so applications can call it for predictions.
📄️ Native Feature Transforms
This guide covers michelangelo.lib.native_transform — a library of PyTorch