Tencent-Hunyuan/HunyuanVideo ? reverse-engineered prompt

Hey, can you set up the HunyuanVideo project for me? I want to be able to generate high-quality videos from text prompts, like "a cinematic shot of a spaceship flying through a nebula." I also saw that it can do image-to-video, so I'd love to be able to upload a starting image and have the AI animate it based on my description.

Please install all the Python dependencies and download the necessary pre-trained model weights. The repository seems to include a script for a Gradio web demo. I'd like you to get that running so I have a simple interface to type in my prompts and see the results. It looks like the model has a special feature for rewriting prompts to make them better, so make sure that is enabled.