Cog[-rust]: Containers for machine learning

Cog is an open-source tool that lets you package machine learning models in a standard, production-ready container.

Cog bundles Python models into a Docker image with a simple interface for loading and running models. This project aims to match that interface for Rust models, so they can be used interchangeably (on Replicate or Dyson).

WIP

To see the proposed DX, check the hello-world example.

• [x] Basic web server
• [x] Rust Cog interface
• [x] Make everything work
• [x] Request validation
• [x] Dockerfile
• [x] cargo cog login
• [x] cargo cog debug
• [x] cargo cog build
• [x] cargo cog push
• [x] cargo cog predict (for everything except arrays)
• [x] deploys to Replicate work
• [x] Run locally
• [ ] Run on Replicate