Flavors of Moonshine: Tiny Specialized ASR Models for Edge Devices
This enables accurate on-device ASR for languages with limited support, though it is incremental as it builds on existing ASR methods.
The paper tackled the problem of automatic speech recognition for underrepresented languages on edge devices by developing tiny specialized monolingual models, achieving error rates 48% lower than comparably sized models and matching or outperforming much larger models.
We present the Flavors of Moonshine, a suite of tiny automatic speech recognition (ASR) models specialized for a range of underrepresented languages. Prevailing wisdom suggests that multilingual ASR models outperform monolingual counterparts by exploiting cross-lingual phonetic similarities. We challenge this assumption, showing that for sufficiently small models (27M parameters), training monolingual systems on a carefully balanced mix of high-quality human-labeled, pseudo-labeled, and synthetic data yields substantially superior performance. On average, our models achieve error rates 48% lower than the comparably sized Whisper Tiny model, outperform the 9x larger Whisper Small model, and in most cases match or outperform the 28x larger Whisper Medium model. These results advance the state of the art for models of this size, enabling accurate on-device ASR for languages that previously had limited support. We release Arabic, Chinese, Japanese, Korean, Ukrainian, and Vietnamese Moonshine models under a permissive open-source license.