Does a LLama (llama.cpp, huggingface) local deployable server exist yet, that lets me load different kinds of QLora finetunings on demand and does tasks in a priority queue?
So that I could have for instance and bulk archive processor that has a specific fintuning work, while the server remains useful in handling user requests with a higher priority?
Would be good if it would be locally deploy-able, but a cloud works too.
