Hanami

Loading...

680,000 hours of multilingual audio is what OpenAI trained Whisper on: 438,000 hours of English speech, 126,000 of non-E | Hanami