🤖 Nexus is robot software for use on a simple Pi or other GPIO-based system. It is made possible with the assistance of an LLM (large language model) like LLaMA 3.2.
- It communicates to a remote and central AI server on your network (for low-latency communication),
https://robot.local
for example. - The central AI server identifies your robot and places it inside it's hub for management.
- The central AI server communicates with the robot through HTTP securely (authorization secure).
- Your robot receives requests and sends requests, helping things like speech data reach the central AI server.
👁️
It uses YOLOv8 software to detect objects out of a camera🦾
It uses OpenPaw software as a proxy🖥️
Requires a local machine (a good control station) that can run large language models efficiently, otherwise it will be super slow⏰
Completely real-time (*unless your network is slow, minimum Wi-Fi 4 or 5-gigabit Ethernet connection recommended)
- It is recommended that you get a GPU that can run both of these models efficiently and effectively. Aim for a GPU that has at least 12GB of VRAM and a lot of CUDA cores for your OpenPaw server software.
No contributors yet.