replicate/cog

Containers for machine learning

76
/ 100
Verified

This tool helps machine learning engineers and researchers easily package their trained ML models into standardized, production-ready containers. You define the model's environment and how it processes inputs, and the tool generates a Docker image that can take data (like an image file) and return the model's output (like a transformed image). It's designed for anyone deploying machine learning models into live applications.

9,268 stars. Actively maintained with 83 commits in the last 30 days. Available on PyPI.

Use this if you need to deploy your machine learning model to production reliably and efficiently, without getting bogged down in complex Docker configurations or CUDA compatibility issues.

Not ideal if you are looking for a tool to train models or if you do not plan to deploy models using Docker.

machine-learning-deployment MLOps model-serving containerization ML-engineering
Maintenance 22 / 25
Adoption 10 / 25
Maturity 25 / 25
Community 19 / 25

How are scores calculated?

Stars

9,268

Forks

657

Language

Go

License

Apache-2.0

Last pushed

Mar 12, 2026

Commits (30d)

83

Dependencies

8

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/replicate/cog"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.

Compare