約164件1ページ目

日本語のみで絞り込む

A Benchmark for Evaluating Turn-Taking and Overlap Handling in Full-Duplex Spoken Dialogue Models - DanielLin94144/Full-Duplex-Bench.

In this paper, we survey the field of Full-Duplex Spoken Language Models (FD-SLMs), which enable synchronous human–AI dialogue via simultaneous speaking and ...

A curated list of full-duplex spoken dialogue models & benchmarks - Ruiqi-Yan/Awesome-Full-Duplex-SDM.

人間のように自然な発声ができ、相槌や割り込みにも対応できる音声対話モデル「Haru 4B」が登場した。higurashiが11月13日に発表したもので、事前学習済みベースモデルのオープンウェイト(モデルの重み)をApache Li...

PC Watch

Lychee-FD is a native end-to-end full-duplex speech language model designed for realtime spoken interaction. Instead of relying on cascaded ASR, LLM, TTS, and ...

2025/9/18 -True Full-Duplex (TFD) voice communication—enabling simultaneous listening and speaking with natural turn-taking, overlapping speech, ...

2026/2/3 -I wanted to see if I could build a full-duplex speech model that avoids the coherence degradation that plagues models of this type while ...

LSLM is an innovative approach to full duplex modeling in interactive speech language models, enabling real-time interaction and turn-taking in spoken dialogues ...

2026/3/10 -This repository provides the model for DuplexCascade, a full-duplex speech-to-speech dialogue system built on a cascaded ASR-LLM-TTS ...

MichiAI is a lightweight, multimodal speech large language model designed for full-duplex interaction. Unlike traditional serial pipelines (ASR → LLM ...

Hertz-dev: an open-source, first-of-its-kind base model for full-duplex conversational audio. ... Github: GitHub - Standard-Intelligence ...

RedditGothsim10