https://github.com/mpolinowski/ray-deployments
Use Ray to deploy your remote services.
https://github.com/mpolinowski/ray-deployments
deployment python ray ray-serve
Last synced: over 1 year ago
JSON representation
Use Ray to deploy your remote services.
- Host: GitHub
- URL: https://github.com/mpolinowski/ray-deployments
- Owner: mpolinowski
- Created: 2023-01-29T10:15:00.000Z (over 3 years ago)
- Default Branch: master
- Last Pushed: 2023-01-29T10:15:03.000Z (over 3 years ago)
- Last Synced: 2025-01-28T19:17:33.644Z (over 1 year ago)
- Topics: deployment, python, ray, ray-serve
- Language: Python
- Homepage: https://mpolinowski.github.io/docs/IoT-and-Machine-Learning/AIOps/2023-01-29-python-ray-deployments/2023-01-29
- Size: 1.95 KB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
## Deployments with Ray
Use Ray to deploy your remote services.
> Source: [Scaling Python with Ray](https://github.com/scalingpythonml/scaling-python-with-ray)
Use __Ray Serve__ for implementing a general-purpose microservice framework and how to use this framework for model serving. __Ray Serve__ is implemented on top of Ray with Ray actors. Three kinds of actors are created to make up a Serve instance:
| Ray Actor | Description |
| -- | -- |
| __Controller__ | The controller is responsible for creating, updating, and destroying other actors. All of the __Serve API__ calls (e.g., creating or getting a deployment) use the controller for their execution. |
| __Router__ | There is one router per node. Each router is a HTTP server that accepts incoming requests, forwards them to replicas, and responds after they are completed. |
| __Worker Replica__ | Worker replicas execute the user-defined code in response to a request. |