Speechlib

1. added regular whisper support
2. users should provide their own huggingface access tokens
3. updated the api

using faster-whisper==1.0.2

Adds MPS support for faster inferencing on Macs

changelog:

log file saved as filename_HHMMSS_language.txt

change log:

- added faster-whisper
- now able to directly input mp3, flac files to transcribe function without converting to wav
- updated readme