fix: add torch.no_grad() reducing memory size

feat: remove gradio and spaces from requirements.txt

feat: add finally block, it created a block for inference

feat: update default images from 1 to 4

feat: update default inference steps to 35

feat: set inference steps to 35

feat: change max batch from 4 to 1 to avoid OOM

feat: change name

fix: remove gr blocks

fix: detach values calling numpy

fix: detach values calling numpy

deps: update requirements file

load variables from environ

feat: gradio app with ZeroGPU support

initial commit