https://github.com/ezroot/aiimagegen
AI Image generator server with a Discord Bot
https://github.com/ezroot/aiimagegen
ai bot discord generation image stablediffusion
Last synced: about 1 month ago
JSON representation
AI Image generator server with a Discord Bot
- Host: GitHub
- URL: https://github.com/ezroot/aiimagegen
- Owner: EZroot
- Created: 2024-12-26T20:48:36.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2025-03-05T22:06:21.000Z (over 1 year ago)
- Last Synced: 2025-03-05T23:19:50.145Z (over 1 year ago)
- Topics: ai, bot, discord, generation, image, stablediffusion
- Language: C#
- Homepage:
- Size: 1.3 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# AIImageGen
Create a /model/ folder with the corresponding models downloaded through civitai. Then in civittest.py you need to select the diffusion model the checkpoint is based on (SD, or SDXL).
Some civit models are pretty optimized, so you can generate large images. Some are not. Settings need to be independantly checked per model you load as well.
RECOMMENDED:
Create a python virtual env, to seperate out the pip modules (This is using python3 btw)
to create:
python3 -m venv sd-env (sd-env can be whatever)
to activate your env, do:
source sd-env/bin/activate
to deactivate:
deactivate
pip modules:
torch
diffusers
transformers
huggingface_hub
Run the AIDiscordBot using dotnet and fill out the config.json file it generates on first run to have a interface from discord to your image generation python server
------------------------
Flux is large, but one of the more accurate image gen models
animdiff is fast, but shitty results (I also need to try SDXL version)
img2vid is SDXL + LTVideo which is super slow but pretty good results (hour or so for a small vid)