advimman/lama ? reverse-engineered prompt

Build me a simple image inpainting app that uses LaMa so I can remove unwanted objects from photos and fill in the missing area realistically.

I want to upload an image, paint a mask over the part I want removed, and then get back a cleaned up version that works well even for big masks and higher resolution photos. Please include a clear way to run it locally and a demo flow that makes it easy to test with my own images. If it helps, add support for generating random masks for quick testing and a refinement option for better results. Use the existing model checkpoints if they’re available, and make sure the setup instructions are straightforward enough for someone who is not a machine learning expert. If you need to check current install or model download details, look up the latest docs online and wire it up so it just works.