BentoML
BentoML copied to clipboard
feature: use LoRA as entity in bentoml model store
Feature request
We are using diffusers SDXL + LoRA Potentially we gonna have lots of LoRA files that we want to manage.
we would like to be able to save LoRA files to bentoml store something like: bentoml.lora.save_model(lora_file, "my_use_case_name")
it might be a pickle file, and might be safetensors
and load it dynamically on inference from: https://docs.bentoml.org/en/latest/frameworks/diffusers.html#using-lora sd21_runner = bento_model.with_options( pipeline_class=diffusers.StableDiffusionPipeline, lora_weights="my_use_case_name", ).to_runner()
Thank you!
P.S i would love to hear want you guys think and of course contribute!
Motivation
No response
Other
No response