LocalAI_Mac   

A step-by-step guide to set up local AI tools on Mac, including detailed instructions for installing and configuring Ollama, Homebrew, Conda, Open WebUI, and Stable Diffusion.

Chapter 1: Setting Up Ollama   

Install the Ollama tool to run AI models like Llama locally. This foundational step ensures the required software environment is set up.

1. Download and Install Ollama

   • Visit the official website to download: ollama.com

2. Run Ollama

   • Open Terminal and execute the following command to start a model:

     ollama run llama3.2:3b

3. Test Ollama

   • Enter a message in the terminal to test functionality:

     Hello!
     

---

Chapter 2: Installing Homebrew

Install Homebrew, a package manager for macOS that simplifies the installation of required software.

1. Install Homebrew

   Run the following command in Terminal:

   /bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"

2. Add Homebrew to Your System Path
   Append Homebrew's path to your shell profile (e.g., .zprofile for Zsh):

   echo 'eval "$(/opt/homebrew/bin/brew shellenv)"' >> ~/.zprofile

3. Restart Terminal and Verify Installation

   • Close and reopen Terminal.

   • Run this command to verify Homebrew installation:

     brew --help

---

Chapter 3: Installing Conda

Obtain the Conda installer suitable for your system to manage Python environments and dependencies.

1. Download Miniforge3

   • Get the installer for macOS (ARM64):
     Miniforge3 for MacOS ARM64

2. Run the Installer

   • Navigate to the download directory and execute the script:

     cd Downloads
     sh Miniforge3-MacOSX-arm64.sh

3. Activate Conda

   • Add Conda to your shell by running:

     eval "$($(conda info --base)/bin/conda shell.zsh hook)"

4. Verify Installation

   • Run this command to check Python version:

     python --version

---

Chapter 4: Installing and Running Open WebUI

Launch the WebUI to provide a local interaction interface and add required AI models to extend WebuUI functionality.

1. Install Open WebUI     Use pip to install Open WebUI:

   pip install open-webui

2. Start Open WebUI

   open-webui serve

3. Access the Interface

   • Open your browser and navigate to: localhost:8080.
   • Complete the registration process.

---

Chapter 5: Setting Up Stable Diffusion

Download model files to serve as the foundation for image generation.

1. Create a Conda Environment

   • Open a new terminal

   eval "$($(which conda) shell.zsh hook)"
   conda create -n stable python=3.10.6
   conda activate stable  

2. Clone the Stable Diffusion Repository

   • Create a working directory and clone the repository:

     mkdir stablediff
     cd stablediff
     git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui.git
     cd stable-diffusion-webui

3. Download the Stable Diffusion Model

   • Get the model file (e.g., v1.5 Model) and place it in:

     models/Stable-diffusion  
     

4. Run the WebUI Script

   ./webui.sh

5. Generate Images

   • Open the interface in a browser, select the model file, and enter prompts to generate images.

---

Chapter 6: Linking Stable Diffusion to Open WebUI

Enable API mode to control Stable Diffusion via WebUI.

Start Stable Diffusion with API Mode

1. Start Stable Diffusion with API Mode

   ./webui.sh --listen --api

2. Configure Open WebUI

   • Go to Settings > Admin Settings > Images.

   • Set the Image Generation Engine to AUTOMATIC1111.

   • Input the following URL:

     http://127.0.0.1:7860
     

3. Test the Connection

   • In Open WebUI, click Refresh under Image Settings and save the configuration.

---

Chapter 7: Automating Setup with AI_Start.sh

To simplify the startup and configuration process for Open WebUI and Stable Diffusion, use the provided AI_Start.sh script. This script automates:

• Starting Open WebUI and initializing the Conda environment.
• Guiding users through necessary configurations (e.g., setting up the image generation engine).
• Launching the Stable Diffusion WebUI and enabling API mode.

Instructions

1. Download the Script
   Place AI_Start.sh in your home directory or another convenient location.

2. Grant Execute Permission

   chmod +x AI_Start.sh

3. Run the Script

   ./AI_Start.sh

4. Follow the Prompts

   • The script will prompt you at key steps, such as configuring the image generation engine in Open WebUI.
   • Complete each step and press Enter to proceed.

Script Details

• Automated Conda Initialization: Ensures dependencies run in the correct environment.
• Multi-Window Management: Runs Open WebUI in a new terminal window for easier process handling.
• Stable Diffusion Automation: Eliminates manual steps for starting the WebUI and enabling API.

---

Final Testing

1. Generate Images from Open WebUI

   • Enter a text prompt and click the Image Icon under the response.

2. Verify Results

   • Wait for the image to be generated and check the output.

---

This guide is designed to help anyone set up a fully functional local AI environment on macOS. Contributions and improvements are free to open a pull request! 😻😻