Add groq_whisper() function for fast speech-to-text transcription

[edubruell]

Description: Integrate Groq’s Whisper models for fast speech-to-text transcription in tidyllm. Add support for handling various audio file formats and optionally detecting languages.

Possible Actions:

• See whether we will need audio-preprocessing with a package like tuneR
• Implement a groq_whisper() function to send audio files to Groq’s fast Whisper models and retrieve text transcriptions.
• Add an argument to specify the language or enable automatic language detection.
• Optionally, support batch processing of multiple audio files in parallel, returning a tibble of file paths and their transcriptions.

[edubruell]

It is called groq_transcribe() and it works...