Hi there,
I am looking for some ideas and inputs.
Who is using a notebook to develop models, and deploys it for big dataset training to a server or heavy workstation?
I work on an Apple device and run it on my Ubuntu GPU workstation but somehow the process is not that smooth.
what is your workflow, or do you prefer to use an Ubuntu notebook? What do you do about the Office suite.
I’m not an expert in this area, but I can share my current setup.
My main workstation is a Windows notebook, and I also run a small Proxmox server (2×RTX 5090). On that server, I use several LXC containers with GPU passthrough, each dedicated to specific tasks — for example, one for data-science experiments (with different Python environments), another for Ollama, and others as needed. The containers can also communicate with each other.
Each container has its own software stack Linux image, Python, PyTorch, etc. On the Proxmox host, I maintain shared folders for AI models, datasets, and related files, which I can easily mount inside containers when needed.
I usually connect from my notebook using SSH and VS Code Remote Server, which lets me work directly inside those containers as if they were local environments. So far, it’s been a flexible and efficient workflow for managing different projects and workloads.