DuplexJev 不經解碼即可批次處理語音決策,單 Token 監督讓凍結 LLM 聽見逐字稿以外的資訊
Batched Speech Decisions Without Decoding: Single-Token Supervision Lets a Frozen LLM Hear Beyond the Transcript
DuplexJev 提出將 ASR 編碼器的隱藏狀態經小型連接器送入凍結的 LLM,直接以單個 Token 對選項作分佈預測,無需自回歸解碼。在 8-GPU 節點上,系統約 0.1 s 內可針對八段語音作出 80 項決策;使用最後一層連接器時,語音問答準確率為 90%,接近讀取逐字稿的 91%。
Published on Oct 2
Authors:
,
,
,
,
,
Abstract
Full-duplex voice agents make many small, closed decisions, which current systems answer by slow autoregressive decoding. We propose DuplexJev, which feeds ASR-encoder hidden states through a small connector into a frozen LLM and reads each question as a single-token distribution over its options. Nothing is decoded, and an 8-GPU node answers 80 decisions about eight utterances in about 0.1 s. With a last-layer connector, spoken QA stays close to reading the transcript (90% vs. 91%). DuplexJev also hears the speaker: gender and emotion accuracy both reach 90% (from 55% and 28%) with a cross-attention connector, whose spoken QA drops by only 1 point (83% to 82%). We train decisions with cross-entropy on the read-out answer token, instead of the usual transcript distillation, whose teacher never hears the voice, and keep distillation for content. Encoders and LLMs are interchangeable; we release weights, training recipe, a batched-inference pipeline for full-duplex serving and a bilingual spoken-QA set.
View arXiv page View PDF Add to collection
Get this paper in your agent:
hf papers read 2610.02638
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash
Models citing this paper 11
Audio-Text-to-Text • 4B • Updated about 23 hours ago • 80 • 1adventists-ai/DuplexJev-4B
Audio-Text-to-Text • 33B • Updated about 23 hours ago • 38adventists-ai/DuplexJev-32B
Audio-Text-to-Text • 31B • Updated about 23 hours ago • 27adventists-ai/DuplexJev-Gemma-31B
Audio-Text-to-Text • 4B • Updated about 23 hours ago • 40adventists-ai/DuplexJev-4B-Turn
Datasets citing this paper 0
No dataset linking this paper
Cite arxiv.org/abs/2610.02638 in a dataset README.md to link it from this page.
Spaces citing this paper 3
Collections including this paper 1
來源:HuggingFace Daily Papers(社區熱門論文) · huggingface.co