Enhance Whisper transcription with multiple model support and performance improvements

Refactor project structure and update dependencies

Remove vocals from ui code

Change UI

Change transcribe compute to float32

Add progress bar to transcribe script

Change transcription to fp16, and read if vocals separation is made into another script, if yes it doesn't separe vocals, if no it separe vocals

transcribe with whisperx

Manage files into transcribe.py

Set output transcription file as srt into transcrie.py

Create script to transcript audios