HF model

videosdk-live / Namo-Turn-Detector-v1-Japanese

A DistilBERT‑based, quantized ONNX model that detects Japanese speech turns with 93.5 % accuracy and <14 ms latency, ready for Python, ONNX Runtime, or VideoSDK integration.

Description in other languages

Statistics

Statistics as of Nov 16, 2026

Hugging Face downloads (last 30 days)
2,000
Hugging Face likes
2
Listed
Oct 13, 2026

Resources in similar research fields

Learn from slides & articles