Web UI

[cocktailpeanut]

I wrote a simple gradio web ui to call the inference. I did my best to not modify the code as much as possible.

1. Just added a scripts/app.py file that does the gradio stuff
2. The inference.py file returns the output file path, and also prints each step in the loop (otherwise it's impossible to know how far along I am when running this over a long audio clip)

Here's what the UI looks like: https://x.com/cocktailpeanut/status/1802377545733529840

Additionally, fixed the audio issue where the generated video wouldn't have sound, and you can't upload the video to social media because the video is detected as corrupt. This was because the write_videofile doesn't specify audio_codec and ends up using None. So by specifying aac the generated video has sound everywhere and can be uploaded to any website. (code here: main...peanutcocktail:hallo:main#diff-53d5d3645d0d3f299313331b6b6dec81f7f7f10028f879edab54baaed2a2a3d8R318)

[k18coldwl]

I don't think anyone will accept your Pull Request, please do some research on what a good Pull Request is.

[AricGamma]

Thank you for your contributing, and we are very glad to merge this pr. 💕💕

Before this, we have a few simple changes requests:

1. Please squash commits and use a meaningful commit message.
2. Move app.py to a new demo folder and add gradio dependency in requirements.txt.

[cocktailpeanut]

@AricGamma thank you, I just forked from scratch and sent another PR here #51 adding the gradio dependency as well. Closing this one.