AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

DrVoice: Parallel Speech-Text Voice Conversation Model via Dual-Resolution Speech Representations

Chao-Hong Tan, Qian Chen, Wen Wang, Chong Deng, Qinglin Zhang, Luyao Cheng, Hai Yu, Xin Zhang, Xiang Lyu, Tianyu Zhao, Chong Zhang, Yukun Ma and 5 more

ICLR 2026 · Large language models › Language, safety and interpretability · approach: Speech and audio language models · 7 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add continuous speech encoder (Whisper features) for speech input (ablation)DrVoice-Small · Llama Questions S2Tscore: 62.33 -> 72.33Improved
add continuous speech encoder (ablation)DrVoice-Small · Llama Questions S2M (text)score: 61.67 -> 68.67Improved
add continuous speech encoder (ablation)DrVoice-Small · Llama Questions T2Tscore: 74.00 -> 75.33No clear effect
add Speech Refined Head (SRH) (ablation)DrVoice-Small · Llama Questions S2M (text)score: 21.67 -> 38.33Improved
add Speech Refined Head (SRH) (ablation)DrVoice-Small · Llama Questions T2M (text)score: 45.22 -> 59.33Improved
add Speech Refined Head (SRH) (ablation)DrVoice-Small · Llama Questions T2Tscore: 73.00 -> 73.33No clear effect
add SRH pretraining (ablation)DrVoice-Small · Llama Questions S2M (text)score: 38.33 -> 61.67Improved
add CoM-Mixing training (ablation)DrVoice-Small · Llama Questions S2M (text)score: 58.00 -> 68.67Improved
raise grouping factor from 1 to 5 (ablation)DrVoice-Small · Grouping sweep S2Tscore: 55.67 -> 63.33Improved
raise grouping factor from 5 to 7 (ablation)DrVoice-Small · Grouping sweep S2M (text/speech)score: 37.67 / 28.00 -> 36.00 / 16.67Hurt
use grouping factor 5 instead of 1 (ablation)DrVoice-Small · Training GPU hoursGPU hours: nearly 50% reductionImproved
add Stage 2 low-LR fine-tuning after merging (ablation)DrVoice 7B · VoiceBench (Inhouse v2, MagpiePro, InfGen)average: 70.19 -> 74.73Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy