Skip to main content
The init command generates a project template with everything you need to build and deploy a model on Reactor.

​
Usage

reactor init --model <model-name>

​
What it generates

my-model/
├── Dockerfile           # Optimized for Reactor deployments
├── model.py             # Example ReactorPipeline implementation
├── requirements.txt     # Python dependencies
├── reactor.yaml         # Model configuration
└── README.md            # Getting started instructions

​
Dockerfile

A multi-stage Dockerfile configured for GPU-based inference. It includes CUDA base images, Python dependency installation, and the correct entrypoint for Reactor’s runtime.

​
model.py

A minimal ReactorPipeline implementation you can use as a starting point. Replace the inference logic with your own model.

​
reactor.yaml

Configuration file specifying your model name, version, and runtime settings. The CLI reads this file for commands like model publish.

​
Next steps

Runtime Overview

Learn how to build models with Reactor Runtime.

Docker Deployment

Build and push your Docker image.