Skip to content

Install

github-actions[bot] edited this page Oct 24, 2026 · 1 revision

Install

This takes about two minutes. You need the board switched on and connected to the internet, and a terminal on it: either a keyboard and screen plugged in, or SSH from your computer.

1. Open a terminal on the board

Over SSH from your computer (Windows, macOS and Linux all have ssh built in):

ssh your-user@board-ip

Don't know the board's address? Your router's device list shows it, or run hostname -I on the board itself.

2. Run the installer

Copy this line, paste it into the terminal and press Enter:

curl -fsSL https://raw.githubusercontent.com/anand34577/ggml-rknpu2/main/install.sh | bash

What it does, in order:

  1. Downloads the ready-made program package and checks it isn't damaged.
  2. Downloads Rockchip's NPU runtime from Rockchip's own GitHub page.
  3. Adds the llama-npu command.
  4. Checks your board and prints what it found.

Everything goes into ~/.local/share/ggml-rknpu2 in your home folder. It doesn't need your password and doesn't touch the rest of the system.

Older system? On Ubuntu older than 22.04 or Debian older than 12, the installer builds the program on the board instead. That takes 10 to 40 minutes and asks for your password once, to install the build tools. You can keep using the board meanwhile.

3. Read the check at the end

The installer ends with a check like this:

Chip                   OK    rk3588  (Radxa ROCK 5B)
NPU driver             OK    /dev/dri/renderD129
NPU access             OK    allowed for radxa
NPU runtime            OK    /home/radxa/.local/share/ggml-rknpu2/bin/librknnrt.so  (version 2.3.2)
llama.cpp              OK    can use RKNPU2: Rockchip NPU (rk3588)

All OK? You're done. A line saying FIX tells you what to do, and Troubleshooting explains each one. Even with a FIX line everything works, just on the CPU.

4. Open a new terminal

Close the terminal and open a new one (or run source ~/.bashrc) so it finds the new llama-npu command. Then go to Your first chat.

Installer options

Add them after bash -s --, for example curl -fsSL .../install.sh | bash -s -- --from-source.

Option What it does
--from-source Build on the board instead of downloading the ready build
--version v0.1.0 Install a specific release
--prefix DIR Install somewhere else
--system-runtime Use the Rockchip runtime already installed in /usr/lib
--runtime-version 2.5.0 Download a different Rockchip runtime version
--no-path Don't add ~/.local/bin to your PATH

Clone this wiki locally