Add support for Faster Whisper models

[a-huk]

They offer an unbeatable accuracy to transcription time.

[ML-CSI]

They offer an unbeatable accuracy to transcription time.

Please check https://huggingface.co/spaces/hf-audio/open_asr_leaderboard
especially look for column "WER" and "RTFx"

side notes:
on "https://github.com/SYSTRAN/faster-whisper" they say "This implementation is up to 4 times faster than openai/whisper for the same accuracy"

RTFx (higher=better)
OpenAI-whisper: 350; (times 4 as stated above: 1400 RTFx)
nvidia/parakeet-tdt-0.6b-v2: 3400 RTFx

WER("accuracy"; lower=better):
openai/whisper-large-v3: 7.44
nvidia/parakeet-tdt-0.6b-v2: 6.05