guoyww/animatediff — reverse-engineered prompt

Build me a local text to animation tool based on this repo. I want to run it on my machine, install the Python requirements, and then either use a simple web app or a basic command to generate short animated clips from text prompts. Please keep it aligned with the main branch setup for Stable Diffusion 1.5 style models, not the XL version.

I want the normal flow to be very simple, type a prompt, pick a model or example config, generate, and save the result into a samples folder. If this repo already auto downloads the needed checkpoints on first run, keep that working. Please make the Gradio app usable and easy to launch locally, and include a couple of ready to try examples for realistic and cartoon outputs.

If the repo already supports it, also expose the extra motion controls and the sparse image or sketch guidance in a basic optional way, but do not make that the default path. Make sure everything is actually runnable end to end, and look up current docs online if you need to.