CoolFace
Apppublic

jbilcke-hf/media-server

sourceHugging Faceupdated 3y agoView on Hugging Face
8likes
App README

Media server ๐Ÿ“ก

The main code of the webtv is located inside the media-server :

manual steps:

  • โ€”human input to write a short paragraph describing a multi-shot video sequence
  • โ€”manual submit it to GPT-4 to generate a list of video captions for each shot (the system instructions are extracts from a stable diffusion guide)
  • โ€”commit the captions to the playlist database

Inside the media-server space (generation process running in the background):

  • โ€”for each prompt in the database
  • โ€”generate a silent 3 seconds video clip with Zeroscope V2 576w (hosted on Hugging Face Spaces)
  • โ€”upscale the clip with Zeroscope V2 XL (also a HF Space)
  • โ€”perform frame interpolation with FILM (also a HF Space)
  • โ€”storage in the Persistent Storage of the media-server Space

Inside the media-server space (streaming process running in the foreground):

  • โ€”for each video file in the persistent storage folder
  • โ€”add it to a new FFmpeg playlist (it's just a .txt file)
  • โ€”broadcast it over the RTMP protocol using FFmpeg (in FLV format)
  • โ€”diffusion of the stream using node-media-server

Inside the AI-WebTV space:

  • โ€”display the stream using mpegts.js
  • โ€”this doesn't work on iPhone, but now there is also a Twitch mirror