Examples

Pick a section below to jump straight to its examples.

Agents

Build agents that use tools, browse the web, run code in sandboxes, and chain reasoning across calls.

• Agents — guides and end-to-end builds

Deploy and inference

Run inference engines, deploy containers, and serve models at scale with custom servers.

• Inference — deploy a model, serve from containers, run multi-node inference

Train models

Custom training, fine-tuning, and pretraining — from low-code recipes to full-code Trainer builds.

• Examples (full code) — wrap any PyTorch model in a Trainer, build your own training loop with Fabric
• Examples (low code) — train models with minimal scaffolding via Lightning AI's high-level APIs

Prep data

Optimise datasets for fast loading, transform large datasets at scale, and stream data into training jobs.

• Datasets — create datasets, import existing ones, stream from cloud storage

Host web apps

Serve Gradio apps, Streamlit dashboards, demos, and any web app from a Studio.

• Host web apps — Gradio, Streamlit, and custom web apps