4 ms·
Judging from OP's comment, they used OpenAI's Whisper speech-to-text transcription model, which was recently released open-source: https://openai.com/blog/whisp
by mkaic 4y ago
Judging from OP's comment, they used OpenAI's Whisper speech-to-text transcription model, which was recently released open-source: https://openai.com/blog/whisper/ https://openai.com/blog/whisper/