Local Ai Pc 7663 Dual Epyc / Dual 9709

CPU Information
Name
AMD EPYC 7663
Topology
2 Processors, 112 Cores, 224 Threads

Memory Information
RDiMM 2933Mhz
Size 1007.61 GB

System Information
Operating System
Ubuntu 24.04.5 LTS

Motherboard
Giga Computing MZ72-HB2-00

GPUs
2x ASUS Turbo R9700 AI Pro 32 GB newest BIOS low Fan Profile (throttled to 210w currently)
ROCm version: 7.2.3

Beside that baby I have a Strix Halo M5 128Gb

Now wanted to setup that big boy for local llm

For myself want to use it for coding . But
I am also looking for something where I can also easily switch Model within the UI . Can Openwebui reload the model and what I read is that vllm is best for tensor split formte two cards . Also want to use it for family for image creation within the ui and also Image checking kind of Allrounder as chatgpt

Is that possible with vLLM?

I am also looking for docker setups so i can keep my host clean

Thank you for you suggestions