Introduction
What DeepLM Ops is, who it is for, and how to get it running.
DeepLM Ops is a local, read-only assistant for GPU and Slurm questions. It runs Qwen3.5-4B on llama.cpp and serves a chat page on your own machine, so your questions and your cluster data stay with you.
It is built for GPU clusters. On a laptop it runs on the CPU, or on an NVIDIA GPU if you have one, and a sample demo cluster lets you try the cluster questions without a real cluster.
Start here
Install on WindowsDocker Desktop with WSL 2, then pull and run.Install on macOSDocker Desktop or Colima, then pull and run.Install on UbuntuDocker Engine, with optional NVIDIA GPU support.Your first questionsQuestions to try on the demo cluster.Security modelWhat stays on your machine and how the container is locked down.TroubleshootingFixes for the most common problems.
What you can ask
- Which GPUs are free right now?
- When will my job start?
- Who is using a given node?
- Did any of my jobs fail this week?
- Will a model fit on a given GPU?