WhisperX containerization #19
Description
The rational is to have diarization implemented in the BrainBox, so the routine of the automatic extraction of voices of the character can be built.
WhisperX https://github.com/m-bain/whisperX needs to be containerized and implemented as a BrainBox decider. The most similar decider is VideoToImages.
The input should be video. The output should be
- the JSON with all the information about fragments WhisperX provides
- The fragments themselves as aac files.
The amount of fragments might be capped to avoid overusing of the disk space.