How do you keep your local model collection organized across machines?
Between GGUFs, safetensors, LoRAs and a few quants of the same model, my homelab model folder is a junk drawer. Each machine has its own Hugging Face cache, and I can't easily tell which box has which revision. What's working for you?