search

getumbrel / llama-gpt

A self-hosted, offline, ChatGPT-like chatbot. Powered by Llama 2. 100% private, with no data leaving your device. New: Code Llama support!
https://apps.umbrel.com/app/llama-gpt
MIT License
10.53k stars 666 forks source link

Installation on windows - please elaborate more #141

Open s-github-2 opened 6 months ago

s-github-2 commented 6 months ago

windows installation are minimal for newcomers - please elaborate more - here are some suggestions that got me going

  1. install WSL https://learn.microsoft.com/en-us/windows/wsl/install 2.Be warmed that I needed bitlocker recovery key on my win11 pro system to boot back after installing WSL because it detected changes in the boot system and wouldnt boot without recovery key. Had to log on to microsoft account to find my recovery key for bitlocker
  2. install docker dekstop
  3. enable docker integration into the wsl distro (Ubuntu is default installation of wsl) through the docker desktop app
  4. Verify Docker Installation by running docker --version in your terminal or command prompt
  5. start ubuntu (default linux distro) from windows - it should have an app installed after the installation
  6. Navigate to the Project Directory . Use the cd command to navigate to this directory, e.g., cd /mnt/c/Projects/llama-gpt remember /mnt/c is the path to c drive from ubuntu or linux
  7. you might get a "permission denied while trying to connect to the Docker daemon socket at unix:///var/run/docker.sock," if you try to run the command
    ./run.sh --model 7b in your linux terminal. To resolve this issue, you need to add your user to the docker group, which will grant the necessary permissions. Here's how you can do it: sudo usermod -aG docker $USER
  8. Apply Group Changes: by shutting down and restarting wsl. type wsl --shutdown in your Windows command prompt or PowerShell, this will shut down wsl
  9. restart wsl by typing ubuntu (default linux distro or whatever you have chosen) and check group membership by running groups command in your terminal
  10. Now you can run " ./run.sh --model 7b " command (assuming you are in correct directory where run.sh is located.
  11. if the run.sh command doesnt run , make sure its executable by running ' chmod +x run.sh ' in terminal