MLCommons introduces a new speech-to-text benchmark based on established industry methods such as Whisper-Large-V3 and the LibriSpeech dataset.| MLCommons