Skip to content

Dynamedia/ai-horde-paperspace

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

33 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

AI Horde Worker for Paperspace Gradient

Purpose

This repo allows users to generate images and contribute to the AI Horde using Paperspace Gradient (Pro/Growth). If you do not have a Paperspace account you can sign up here for $10 free credit.

To learn more about AI Horde, see the project website.

Setup

Create a new notebook with the following advanced settings:

  • Workspace URL: https://github.com/Dynamedia/ai-horde-paperspace
  • Container Name: dynamedia/ai-horde-paperspace:latest

Configuration

Once your container is running you will need to do a small amount of configuration in the file browser. You only need to do this once as all settings will be stored in your workspace for future use.

Required

  • Edit worker_name in config/worker.yaml
  • Edit dreamer_name in config/dreamer.yaml

Optional

  • Set cache_home to '/notebooks/cache/ in config/dreamer.yaml for persistent model storage. Beware - Paperspace may charge you for using too much storage. Check their T&C.

  • Edit the files in config/gpu/ for better performance. Default settings are quite conservative and avoid picking jobs that will take a long time to process. Each file in this directory relates to a machine type on the Paperspace Gradient Pro plan.

Running

To run the worker you can either press run in dreamer.ipynb or open a terminal and type dreamer.sh

The first time you run the worker you will be asked for your AI Horde API key which will then be saved in /notebooks/.env - This is to avoid ever storing it where it could be accidentally viewed if the notebook is set to public.

Managing your Worker

The dreamer.sh script will automatically respawn your worker if it should die unexpectedly.

You can use any of the AI Horde UI's to set or remove maintenance mode. I recommend ArtBot.

About

Paperspace notebook environment with Dockerfile for the image

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors