Pith. sign in

REVIEW 35 cited by

Freeze-Omni: A Smart and Low Latency Speech-to-speech Dialogue Model with Frozen LLM

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2411.00774 v5 pith:7XOWG4QH submitted 2024-11-01 cs.SD cs.AIcs.CLeess.AS

Freeze-Omni: A Smart and Low Latency Speech-to-speech Dialogue Model with Frozen LLM

classification cs.SD cs.AIcs.CLeess.AS
keywords freeze-omniabilitydatadialoguespeechspeech-to-speechtrainingfrozen
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Rapidly developing large language models (LLMs) have brought tremendous intelligent applications. Especially, the GPT-4o's excellent duplex speech interaction ability has brought impressive experience to users. Researchers have recently proposed several multi-modal LLMs in this direction that can achieve user-agent speech-to-speech conversations. This paper proposes a novel speech-text multimodal LLM architecture called Freeze-Omni. Our main contribution is that the speech input and output modalities can be easily connected to a textual LLM while keeping the LLM's parameters frozen throughout the training process. We design a three-stage training strategy for modeling both the speech input and output, enabling Freeze-Omni to obtain speech-to-speech conversation ability using text-speech paired data (such as ASR and TTS data) and only 60,000 multi-round text Q&A data on 8 GPUs. Moreover, we can effectively ensure that the intelligence of the Freeze-Omni in the speech modality is at the same level compared with that in the text modality of its backbone LLM, while achieving low latency end-to-end spoken response. In addition, we also designed a method to achieve duplex dialogue ability through multi-task training, giving Freeze-Omni a more natural style of dialogue ability between users and agents. In summary, Freeze-Omni holds great potential to conduct speech-to-speech dialogue based on a multimodal LLM under the condition of a frozen LLM, avoiding the catastrophic forgetting problem caused by limited data and training resources.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 35 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Unlocking Speech-Text Compositional Powers: Instruction-Following Speech Language Models without Instruction Tuning

    cs.CL 2026-07 unverdicted novelty 7.0

    SpeechCombine produces instruction-following SLMs via speech pre-training followed by direct weight combination with the text LLM instruction delta, without any speech instruction tuning.

  2. FlexiSLM: A Dynamic and Controllable Frame Rate Spoken Language Model

    cs.SD 2026-06 unverdicted novelty 7.0

    FlexiSLM is the first spoken language model supporting dynamic and controllable frame rates on speech input and output, outperforming fixed-rate 7B models at high quality and enabling faster inference at lower rates l...

  3. A Survey of Full-Duplex Spoken Dialogue Systems: Architectural Hierarchy, Interaction Ontology, and Decision State Machine

    eess.AS 2026-06 accept novelty 7.0

    A survey proposing an L0-L3 architectural hierarchy, T×I×R interaction ontology, and IDLE/LISTEN/SPEAK/WAIT/DUAL decision state machine for full-duplex spoken dialogue systems, documenting a realization gap between ar...

  4. DuplexSLA: A Full-Duplex Spoken Language Model with Synchronized Speech, Language, and Action

    eess.AS 2026-05 unverdicted novelty 7.0

    DuplexSLA introduces a three-channel full-duplex architecture that synchronizes continuous user audio, discrete assistant audio, and rate-limited textual actions inside a single backbone for native turn-taking and in-...

  5. Instruct-FD: Can Your Full-Duplex Speech System Follow Turn-Taking Instructions?

    cs.CL 2026-05 conditional novelty 7.0

    Current full-duplex speech systems largely fail to follow explicit turn-taking instructions: the best of six models scores 64.4% adherence on the new Instruct-FD benchmark, and proactive behaviors like interruption an...

  6. VITA-QinYu: Expressive Spoken Language Model for Role-Playing and Singing

    cs.CL 2026-05 unverdicted novelty 7.0

    VITA-QinYu is the first expressive end-to-end spoken language model supporting role-playing and singing alongside conversation, trained on 15.8K hours of data and outperforming prior models on expressiveness and conve...

  7. The Silent Thought: Modeling Internal Cognition in Full-Duplex Spoken Dialogue Models via Latent Reasoning

    eess.AS 2026-03 conditional novelty 7.0

    Causal full-duplex speech LLMs improve response quality by recursively feeding soft vocabulary embeddings as latent thoughts during listening, trained by matching a non-causal expert posterior via ELBO.

  8. The Silent Thought: Modeling Internal Cognition in Full-Duplex Spoken Dialogue Models via Latent Reasoning

    eess.AS 2026-03 unverdicted novelty 7.0

    FLAIR enables spoken dialogue AI to conduct continuous latent reasoning while perceiving speech through recursive latent embeddings and an ELBO-based finetuning objective.

  9. Preserving Speech-to-Text LLM Capabilities in Speech-to-Speech Generation

    eess.AS 2026-06 unverdicted novelty 6.0

    PRIME-Speech adds low-latency speech output to frozen S2T LLMs by synchronizing a causal post-decoder with intermediate hidden states and using mixed conditioning plus turn-level KV-cache packing, preserving original ...

  10. FacePlex: Full-Duplex Joint Speech-Facial Motion Generation for Conversational Avatars

    cs.AI 2026-06 unverdicted novelty 6.0

    FacePlex introduces a unified streaming model with Rolling Flow Matching and Rolling Cross-Attention to enable full-duplex joint real-time generation of speech and facial motion tokens.

  11. Audio Interaction Model

    cs.SD 2026-06 unverdicted novelty 6.0

    Audio-Interaction unifies offline and online audio tasks into one streaming model via the SoundFlow framework and a new 2.6M-item streaming corpus, enabling real-time instruction following and proactive responses.

  12. Learning When to Think While Listening in Large Audio-Language Models

    cs.CL 2026-05 unverdicted novelty 6.0

    A wait-think-answer controller for LALMs is trained via SFT followed by six-reward DAPO, raising row-weighted accuracy from 67.6% to 70.3% and cutting post-endpoint thinking length by 14% on synthetic spoken QA while ...

  13. FastTurn: Unifying Acoustic and Streaming Semantic Cues for Low-Latency and Robust Turn Detection

    cs.SD 2026-04 unverdicted novelty 6.0

    FastTurn unifies acoustic features and streaming CTC decoding for low-latency, robust turn detection in full-duplex dialogue systems and releases a realistic human-dialogue test set.

  14. StableToken: A Noise-Robust Semantic Speech Tokenizer for Resilient SpeechLLMs

    cs.CL 2025-09 unverdicted novelty 6.0

    StableToken introduces a multi-branch architecture with bit-wise voting to create noise-robust semantic speech tokens, achieving lower Unit Edit Distance and better SpeechLLM robustness than prior single-path tokenizers.

  15. Training-Free Multimodal Large Language Model Orchestration

    cs.CL 2025-08 unverdicted novelty 6.0

    LLM Orchestration integrates modality experts via an LLM controller, cross-modal memory, and interaction layer to enable multimodal input-output without gradient-based training.

  16. Step-Audio 2 Technical Report

    cs.CL 2025-07 unverdicted novelty 6.0

    Step-Audio 2 integrates a latent audio encoder, reasoning-centric reinforcement learning, and discrete audio token generation into language modeling to deliver state-of-the-art performance on audio understanding and c...

  17. Step-Audio: Unified Understanding and Generation in Intelligent Speech Interaction

    cs.CL 2025-02 unverdicted novelty 6.0

    Step-Audio introduces a 130B-parameter unified speech-text model with open-sourced components for understanding, generation, affordable voice cloning, and dynamic control, claiming SOTA human evaluation results on a n...

  18. GLM-4-Voice: Towards Intelligent and Human-Like End-to-End Spoken Chatbot

    cs.CL 2024-12 conditional novelty 6.0

    GLM-4-Voice builds an end-to-end spoken chatbot by deriving a 175bps single-codebook tokenizer from ASR, synthesizing interleaved speech-text data, and continuing pre-training of GLM-4-9B on up to 1 trillion tokens be...

  19. Raon-Speech Technical Report

    cs.CL 2026-04 conditional novelty 5.5

    A 9B SpeechLM trained on 1.38M hours of English/Korean data plus a full-duplex chat extension trained on 119K hours of time-aligned dialogue outperform same-size audio models on speech tasks and FDB turn-taking metrics.

  20. Adaptive Turn-Taking for Real-time Multi-Party Voice Agents

    eess.AS 2026-06 unverdicted novelty 5.0

    ModeratorLM conditions a streaming speech LLM on assigned roles for adaptive turn-taking in multi-party settings, reporting over 40% higher precision and 70% higher recall than non-role baselines on real meetings and ...

  21. Which Speech Representation Better Matches Text-Native Reasoning? A Study of Speech-Text Alignment on Frame Rate and Representation

    eess.AS 2026-06 unverdicted novelty 5.0

    Empirical sweep finds 4.17 Hz frame rate plus intermediate-layer alignment optimal for speech QA under frozen text LLM backbone.

  22. DuplexSLA: A Full-Duplex Spoken Language Model with Synchronized Speech, Language, and Action

    eess.AS 2026-05 unverdicted novelty 5.0

    DuplexSLA is a dual-stream three-channel full-duplex model that synchronizes continuous user audio, discrete assistant audio, and rate-limited action text for native turn-taking and in-conversation tool calling.

  23. A Survey of Large Audio Language Models: Generalization, Trustworthiness, and Outlook

    cs.SD 2026-05 unverdicted novelty 5.0

    A survey of Large Audio Language Models that establishes a taxonomy of trustworthiness vulnerabilities and proposes a Defense-in-Depth roadmap for audio intelligence.

  24. Full-Duplex Interaction in Spoken Dialogue Systems: A Comprehensive Study from the ICASSP 2026 HumDial Challenge

    eess.AS 2026-04 unverdicted novelty 5.0

    A new HumDial-FDBench benchmark and real human-recorded dual-channel dataset are released to assess full-duplex dialogue systems on interruptions and conversational flow.

  25. FastTurn: Unifying Acoustic and Streaming Semantic Cues for Low-Latency and Robust Turn Detection

    cs.SD 2026-04 unverdicted novelty 5.0

    FastTurn fuses streaming CTC semantics with acoustic cues for lower-latency, more accurate turn detection and ships a real-dialogue test set.

  26. FastTurn: Unifying Acoustic and Streaming Semantic Cues for Low-Latency and Robust Turn Detection

    cs.SD 2026-04 unverdicted novelty 5.0

    FastTurn fuses streaming CTC semantics with acoustic cues for lower-latency, more accurate turn detection in full-duplex dialogue and releases a real-dialogue test set.

  27. Sharp spectral estimates for free boundary problems arising in plasma physics

    math.AP 2026-04 unverdicted novelty 5.0

    For a constrained superlinear free-boundary plasma model, the non-local first eigenvalue σ₁ is always positive on balls in every dimension N≥2, despite lacking a general Faber–Krahn property.

  28. Towards Building Speech Large Language Models for Multitask Understanding in Low-Resource Languages

    cs.SD 2025-09 unverdicted novelty 5.0

    Introduces XLSR-Thai encoder, U-Align alignment, and Thai-SUP data pipeline to enable multitask speech understanding SLLMs for Thai.

  29. Training-Free Multimodal Large Language Model Orchestration

    cs.CL 2025-08 unverdicted novelty 5.0

    A training-free orchestration framework integrates off-the-shelf modality experts via an LLM controller, text-centric cross-modal memory, and unified interaction layer to enable multimodal input-output without joint training.

  30. Beyond Words: Multimodal LLM Knows When to Speak

    cs.CV 2025-05 unverdicted novelty 5.0

    MM-When2Speak reformulates conversational timing as dense response-type prediction and achieves up to 3x better performance by integrating video, audio, and text cues on top of an LLM backbone using a new dyadic conve...

  31. Kimi-Audio Technical Report

    eess.AS 2025-04 unverdicted novelty 5.0

    Kimi-Audio is an open-source audio foundation model that achieves state-of-the-art results on speech recognition, audio understanding, question answering, and conversation after pre-training on more than 13 million ho...

  32. Adaptive Turn-Taking for Real-time Multi-Party Voice Agents

    eess.AS 2026-06 unverdicted novelty 4.0

    ModeratorLM conditions a chunk-wise streaming speech LLM on assigned roles (with optional CoT) to raise turn-taking precision over 40% and recall over 70% versus non-role baselines on synthetic RolePlayConv data and r...

  33. IRAF: Interference-Resilient Adaptive Fusion for Noise-Robust End-to-End Full-Duplex Spoken Dialogue Systems

    cs.SD 2026-06 unverdicted novelty 4.0

    IRAF introduces an adaptive fusion module that uses a predicted scalar reliability gate to reduce the impact of interfering speakers on user audio representations in end-to-end full-duplex spoken dialogue systems, wit...

  34. The Silent Thought: Modeling Internal Cognition in Full-Duplex Spoken Dialogue Models via Latent Reasoning

    eess.AS 2026-03 unverdicted novelty 4.0

    FLAIR enables simultaneous latent reasoning during speech input in full-duplex dialogue models via recursive latent embeddings and an ELBO-based training objective without added latency.

  35. VITA-1.5: Towards GPT-4o Level Real-Time Vision and Speech Interaction

    cs.CV 2025-01 conditional novelty 4.0

    VITA-1.5 integrates vision and speech into a single LLM through multi-stage training, delivering competitive benchmark results on image, video, and speech tasks with near real-time response speed.