Effortless AI Tool Deployment and Guides for Flux1.0, Ollama, and Beyond
With the release of Flux1.0 by Black Forest Labs I regained my interest with image generation and started exploring how to train Lora's for Flux.
From my research, there is a lot of information in the web regarding dataset generation and training for Flux and other image generation models, but it is scatttered throughout numerous Reddit posts, GitHub issues, websites, etc.
Setting up the development environment is also a hassle, even more so for people without much programming knowledge, which may lead to many giving up before even starting.
To tackle both these issues, I created this repository to serve as a spot for enabling people to easely deploy containarized tools and guides for various Machine Learning related tasks.
🎉 Let's make model training easily accessible! 🎉
Make sure you have docker and docker compose correctly setup on you pc.
Follow the instructions at Docker's official documentation, mainly:
If you have Ollama locally installed, change the ollama host port to a different one in
docker-compose.ollama.yaml
docker-compose -f ./dockercompose/docker-compose.ollama.yaml up -d
# check health status of the services
docker ps -a
# download llama3.1 through the dockerized ollama
docker exec -it ollama ollama pull llama3.1
docker exec -it ollama ollama listThe services should become available at:
- WebUI: http://127.0.0.1:8081
- Ollama: http://127.0.0.1:11434
- Searxng: http://127.0.0.1:11435
Go to http://127.0.0.1:8081 and check if:
- you are able to select the downloaded Llama3.1 model
- you can toggle the web search by presssing the plus button on the left side of the prompt textbox
To allow WebUI to query searxng, you need to add json to the list of formats in searxng settings.yml file:
Enable editing the file:
sudo chmod 666 dockercompose/searxng/settings.ymlPerform the change:
# dockercompose/searxng/settings.yml
...
# remove format to deny access, use lower case.
# formats: [html, csv, json, rss]
formats:
- html
- json
...Restart docker compose:
docker-compose -f ./dockercompose/docker-compose.ollama.yaml down
docker-compose -f ./dockercompose/docker-compose.ollama.yaml up -d- Containarized Ollama + Webui + Search
- Containarized ComfyUI
- Containarized Flux1.0 Dev Lora training with sd-scripts
- ComfyUI Workflows for basic Flux usage
- ComfyUI Workflows for dataset generation
- ComfyUI Workflows for dataset captioning
- Lora dataset generation guides
- Flux Lora training guides
Feel free to contribute! I'm far from an expert but I'll do my best to keep up 🍻!