Loading original post…Open on X ↗
Voice turn-end detection demo using Jev: 0.5s after speech ends, the model scores whether the turn finished and conditionally adds hold time (e.g., timeout=3.0s vs None), with ~0.22s inference latency—still rough but promising.
Original post
Loading original post…Open on X ↗
jev
Start typing to find a use case.