🎤 Enhance speech recognition with GLM-ASR-Nano-2512, a 1.5B parameter model that excels in transcribing low-volume and dialectal speech accurately.