Faster and cheaper Whisper-AI Large-v2 responses. JAX implementation of OpenAI's Whisper model for up to 15x speed-up (doesn't support TPU).
