====== Local Ai Install ======
===== All in one =====
* [[https://lmstudio.ai/|LMstudio]] multi platform
===== local service install on MacOS =====
You need a **__mac Silicon__**
==== Ollama ====
First, install ollama\\
-> [[https://ollama.com/]]\\
Unzip et run the .app\\
Once installed it proposed to run ollama run llama3.2
Do it...in a terminal\\
Once the model is pulled and command prompt is **>>>** test it (say hi)\\
==== Homebrew ====
If not already done, install homebrew:
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
:!: Don't forget to add homebrew to your PATH as indicated at the end of the install :!:
=== git/conda ===
brew install git
grab latest conda for the platform from [[https://github.com/conda-forge/miniforge/releases/]]\\
cd ~/Downloads && wget https://github.com/conda-forge/miniforge/releases/download/24.11.2-1/Miniforge3-24.11.2-1-MacOSX-arm64.sh
sh Miniforge3-24.11.2-1-MacOSX-arm64.sh
Read / accept and install by default [enter] ...
eval "$(/Users/$USER/miniforge3/bin/conda shell.zsh hook)"
python --version #Python 3.12.8
pip --version #pip 24.3.1 from /Users/fortiera/miniforge3/lib/python3.12/site-packages/pip (python 3.12)
pip install open-webui
=== web server ===
open-webui serve
==== Stable diffusion ====
eval "$(/Users/$USER/miniforge3/bin/conda shell.zsh hook)"
conda create -n stable python=3.10.6
conda activate stable
python --version #Must be 3.10.6
cd ~/git # or wherever you put your git stuff
git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui.git
cd stable-diffusion-webui/models/Stable-diffusion
wget https://huggingface.co/stable-diffusion-v1-5/stable-diffusion-v1-5/blob/main/v1-5-pruned.safetensors
cd ../..
./webui.sh
In the web interface select v1-5-pruned.safetensors and check generation
**Ctrl+C** to quit after testing
==== Integration in open-WebUI ====
Let's generate image in our nice local Ai\\
relaunch stable-diffusion:
eval "$(/Users/$USER/miniforge3/bin/conda shell.zsh hook)"
conda activate stable
cd ~/git/stable-diffusion-webui
./webui.sh --listen --api
Click on your account icon then Settings > Admin Settings > Images then:
* Image Generation Engine set to AUTOMATIC1111
* URL to http://127.0.0.1:7860
* Activate "Image Generation (Experimental)"
===== Update =====
==== Open-WebUI ====
eval "$(/Users/$USER/miniforge3/bin/conda shell.zsh hook)"
conda
===== References =====
* https://www.youtube.com/watch?v=6E2-HQMSs5k
* https://github.com/mateo-cogeanu/local-ai-mac-setup
* https://www.youtube.com/watch?v=R0XuVqMSHYk
* https://www.fahdmirza.com/2024/12/magicquill-installation-on-windows.html