Quick install guide

The installation has two steps.

1. Install the yzma command line tool.
2. Use the yzma command to install the llama.cpp libraries for your platform.

Install the yzma command

You need Go 1.26 or later. You do not need a C compiler.

go install github.com/hybridgroup/yzma@latest

Check that the command works:

yzma version

Install the llama.cpp libraries

The yzma command downloads the prebuilt llama.cpp libraries for your machine. Follow the page for your platform.

• macOS
• Linux
• Windows
• Raspberry Pi
• NVIDIA Jetson Orin
• Arduino UNO Q
• Browser
• Manual installation

Set the YZMA_LIB variable

yzma must know where the libraries are. Set the YZMA_LIB environment variable to that directory.

• Linux and macOS
• Windows

export YZMA_LIB=/path/to/lib

set YZMA_LIB=C:\yzma\lib

Add this line to your shell profile to keep the setting.

Next steps

Download a model, then run your first program. See Download models.