OpenClaw Step by Step Local Private AI Homelab Tutorial using a Dell PowerEdge R730 and Discord

Follow along and stand up your own local, private AI Homelab. This video walks through, step by step, how to install and configure OpenClaw.

If you watch the video, you will see the local AI model stuggling when receiving OpenClaw prompts speed wise, but not when querying the AI model directly. This led me down the path of studying OpenClaw prompts.. And I figured out the WHY on the slowness. It sends a ~70Kb prompt on EVERY request BEFORE it even handles the actual prompt. The next thing to do is to figure out HOW to stop that from happening.. Stay tuned..

FIXED! Add these parameters to your llama-server startup command: --cache-prompt -ctk q4_0 -ctv q4_0

First prompt will be slow while it caches things, subsequent will be FAST!

Hardware: Dell PowerEdge R730
GPUS: Tesla Nvidia P100 12Gb
OS: Ubuntu VM on Proxmox
Model: Qwen
Chat integration: Discord

Видео OpenClaw Step by Step Local Private AI Homelab Tutorial using a Dell PowerEdge R730 and Discord канала ptnerd