llama-cpp-cuda on AUR (Arch User Repository)

Last updated: October 05,2024

1. Install "llama-cpp-cuda" effortlessly via AUR with the help of a helper(YAY)

a. Install YAY (https://github.com/Jguer/yay)


    
        
            $
        

        
            
                sudo pacman -S --needed git base-devel && git clone https://aur.archlinux.org/yay.git && cd yay && makepkg -si
            
                    
    
    Copied

b. Install llama-cpp-cuda on Arch using YAY


    
        
            $
        

        
            
                yay -S
            
                        
                llama-cpp-cuda
            
                    
    
    Copied

* (Optional) Uninstall llama-cpp-cuda on Arch using YAY


    
        
            $
        

        
            
                yay -Rns
            
                        
                llama-cpp-cuda
            
                    
    
    Copied

2. Manually Install "llama-cpp-cuda" via AUR

a. Ensure you have the base development tools and git installed with:


    
        
            $
        

        
            
                sudo pacman -S --needed
            
                        
                base-devel git
            
                    
    
    Copied

b. Clone llama-cpp-cuda's git locally


    
        
            $
        

        
            
                git clone https://aur.archlinux.org/llama-cpp-cuda.git 
            
                        
                ~/llama-cpp-cuda
            
                    
    
    Copied

c. Go to ~/llama-cpp-cuda folder and install it


   
    
        
            $
        

        
            
                cd
            
                        
                ~/llama-cpp-cuda
            
                    
    
    Copied

    

   
    
        
            $
        

        
            
                makepkg -si
            
                    
    
    Copied

3. Information about the llama-cpp-cuda package on Arch User Repository (AUR)

ID: 1308458
Name: llama-cpp-cuda
PackageBaseID: 195852
PackageBase: llama-cpp
Version: c3e53b4-1
Description: Port of Facebook's LLaMA model in C/C++ (with CUDA)
URL: https://github.com/ggerganov/llama.cpp
NumVotes: 1
Popularity: 0.490555
OutOfDate: 1692967398
Maintainer: Freed
Submitter: Freed
FirstSubmitted: 1689667143
LastModified: 1692877217
URLPath: /cgit/aur.git/snapshot/llama-cpp-cuda.tar.gz