Skip to content

README.md

Latest commit

 

History

History
53 lines (32 loc) · 1.55 KB

README.md

File metadata and controls

53 lines (32 loc) · 1.55 KB
Raw

HyperCycle BlockHack 2023 Repo

This contains resources related to the HyperCycle challenges at BlockHack 2023.

Writing AI Machines

An AI Machine (AIM) is a packaged AI model that lets other machines run your model in a compatable way that includes cost logic, so other machines can know how to charge for using your model. The hackathon_aim_examples includes examples of how to create such a service, and DockerDiles for creating the corresponding docker images.

Examples given are:

Greeting AIM

This is a simple "Hello World!" AIM, using the pyhypercycle-aim library

Queue AIM

This is an example AIM that uses a builtin queue system to better manager GPU memory usage. See app/main.py for more information

ZTranslate AIM

This is a self-contained example AIM that doesn't use the pyhypercycle-aim library, and includes all the needed header and body information needed to get a custom aim to run.

The pyhypercycle-aim Package

The pyhypercycle-aim package is available at https://github.com/hypercycle-development/pyhypercycle-aim and includes the helper classes being used to create AIMs. To install it, include the following in your requirements.txt file:

git+https://github.com/hypercycle-development/pyhypercycle-aim.git#egg=pyhypercycle_aim

Frontend

For UI interaction we have prepared demos powered by Vite, React and Tailwind so you can quickly build your project. You need to have Node.js and yarn installed and then:

  1. Install NPM packages 
    yarn install
  2. Run application in develop mode 
    yarn dev