From the source
From the source
The Open ASR Leaderboard has added multilingual and long-form transcription tracks.
A preprint analyzing 60+ models from 18 organizations across 11 datasets reveals that Conformer encoder + LLM decoder models achieve the best accuracy, CTC/TDT decoders are the fastest, multilingual models sacrifice single-language performance, and closed-source systems still lead in long-form transcription.
From the source
Over the past two years, the Open ASR Leaderboard has become a standard for comparing open and closed-source models on both accuracy and efficiency. Recently, multilingual and long-form transcription tracks have been added to the leaderboard 🎉
huggingface.co
