- Install the tool(s)
- Install Windows Subsystem for Linux (WSL)
- Install Docker
- Install Ollama and Ollama Web UI
- Test connectivity to the Ollama API and UI
- Visual Studio Code to execute all PowerShell, Python and Linux code (download: https://code.visualstudio.com/download)
- Install the following Visual Studio Code extensions from the Extension Manager:
- PowerShell
- Check if virtualisation is enabled for your pc/laptop. Check this article: https://www.technorms.com/8208/check-if-processor-supports-virtualization
- If it's disabled but your CPU is able to, enable it in your BIOS. Check this article: https://bce.berkeley.edu/enabling-virtualization-in-your-pc-bios.html
- Go to Control Panel > "Turn Windows features on or off"
- Enable 'Virtual Machine Platform'
- Enable 'Windows Subsystem for Linux'
- Run the install command from "install-wsl-ubuntu.ps1" in a Powershell terminal.
- After the installation run the commands in wsl-ubuntu-post-install.sh in a Linux (Ubuntu WSL) terminal.
- We will run Ollama and the UI in a docker instance in Ubuntu.
- Run the commands in "install-docker.sh" in a Linux (Ubuntu WSL) terminal.
- Run the commands in "install-ollama-webgui.sh" in a Linux (Ubuntu WSL) terminal.
- Go to http://localhost:11434/ to see that the Ollama API is running
- Go to http://localhost:3000/ to see if the Ollama WebUI is running
- Register with Ollama-WebUI
- Download a model, e.g.:
- llama2
- mistral-7b
- llava (image recognition)
- If you have a paid account and credits at OpenAI (ChatGPT) > Create a OpenAI API key and set it under Settings > External