Typhoon ASR Streaming: Steerable Low-Latency Thai Speech Recognition with Real-Time Shallow Fusion
A deployable system for streaming Thai ASR that lets a user steer its vocabulary at decode time, without retraining is presented, which lifts keyword recall from 16.6% to 20.7% at no accuracy cost and negligible overhead.
Warit Sirichotedumrong, Tanawin Samutsin, Shahira Wani et al.
· 0 citations