feat: add models locally from_pretrained

add models locally

removed wtf error

update: update return type of inference

fix: commute inference to gpu and touches

fix: fix inference endpoint

fix: fix running handler and add test file for catching potential bugs

fix: attempt to fix inference endpoint with working sdxl

refactor: remove notebook file

feat: add sdxl with controlnet

refactor: add venv locally and num_of_images parameter and return list of images instead of string

update readme and add util to handler for faster generations

touch: update request and diffusers version

remove unneccessary files

Duplicate from philschmid/ControlNet-endpoint