Hey there,
I am new to this and no dev either (yet extremely curious), so please forgive me for some pretty basic questions. I am trying to understand how a base model interacts with LoRA without them merging. For instance, I want to run stabilityai/stable-diffusion-2-1-base model on GCP, and run this LoRA model (lora-library/https-huggingface-co-lora-library-test · Hugging Face) in a different cloud. How do I shard this? How do I amend the code such that in inference, the base model runs its layers and calls the other server for the matrix evaluation (LoRA) before sending back that intermediate output to the base model running on GCP. Can anyone guide me? Many thanks!