Implement async batch predictions

[atiorh]

Adopting Core ML's async batch prediction APIs will be crucial for high-throughput batch transcription with WhisperKit.

Reference: https://developer.apple.com/videos/play/wwdc2023/10049/

[ZachNagengast]

@jkrukowski has expressed interest in taking this on, here is a suggested approach:

Starting with CLI:

• --audio-path accepts an array of audio files now
• --batch-size param which stands for maximum concurrent inference that can happen in the queue (default is unlimited). Open to suggestions on terminology, because batch isn't exactly the same concept as concurrent async
• Accuracy is within 1%, should not need to tradeoff here (whisperkittools can run this analysis)
• This code may help for reference https://github.com/ZachNagengast/similarity-search-kit/blob/ef737c4aad3af904a9925c476dc37c0dddc6ad8c/Sources/SimilaritySearchKit/Core/Index/SimilarityIndex.swift#L231-L244