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 like 6.25 Hz.
hub Mixed citations
Funaudiollm: Voice understanding and generation foundation models for natural interaction between humans and llms.arXiv preprint arXiv:2407.04051
Mixed citation behavior. Most common role is background (57%).
26
Pith papers citing it
7
external citations · Pith
Background
57% of classified citations
abstract
This report introduces FunAudioLLM, a model family designed to enhance natural voice interactions between humans and large language models (LLMs). At its core are two innovative models: SenseVoice, which handles multilingual speech recognition, emotion recognition, and audio event detection; and CosyVoice, which facilitates natural speech generation with control over multiple languages, timbre, speaking style, and speaker identity. SenseVoice-Small delivers exceptionally low-latency ASR for 5 languages, and SenseVoice-Large supports high-precision ASR for over 50 languages, while CosyVoice excels in multi-lingual voice generation, zero-shot in-context learning, cross-lingual voice cloning, and instruction-following capabilities. The models related to SenseVoice and CosyVoice have been open-sourced on Modelscope and Huggingface, along with the corresponding training, inference, and fine-tuning codes released on GitHub. By integrating these models with LLMs, FunAudioLLM enables applications such as speech-to-speech translation, emotional voice chat, interactive podcasts, and expressive audiobook narration, thereby pushing the boundaries of voice interaction technology. Demos are available at https://fun-audio-llm.github.io, and the code can be accessed at https://github.com/FunAudioLLM.
hub tools
citation-role summary
background 4
method 3
citation-polarity summary
representative citing papers
PolySpeech-100 is a new benchmark for native-level speech comprehension across 110 linguistic variants that evaluates 22 models and reports E2E advantages on dialects, robustness gaps on low-resource languages, and degradation from Chain-of-Thought prompting.
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 conversational benchmarks.
AudioHijack generates imperceptible adversarial audio via gradient estimation, attention supervision, and reverberation blending to hijack 13 LALMs with 79-96% success on unseen contexts and real commercial agents.
RoleJudge is a multidimensional evaluation framework for speech-character alignment in audio LLMs, backed by the RoleChat dataset and multi-stage RL training with standard alignment to reduce reward issues.
Lychee-FD resolves modality interference in full-duplex spoken language models by separating acoustic and semantic parameters in deep layers and adding a dense semantic alignment channel, achieving state-of-the-art performance on spoken QA and interaction benchmarks.
A large-scale, hierarchically-annotated human-centric video dataset and a three-level benchmark (OHBench) are presented, showing that fine-tuning on a subset of the data improves audio-video generation model performance.
SwanVoice is a zero-shot TTS system for 1-4 speakers that reports higher richness and hierarchy scores than open-source baselines on monologue and dialogue tasks via mixed training and DiffusionNFT post-training.
VocalParse applies interleaved and Chain-of-Thought prompting to a Large Audio Language Model to jointly transcribe lyrics, melody and word-note alignments, achieving state-of-the-art results on multiple singing datasets.
MiniMind-O delivers a working 0.1B-scale open omni model with speech-native output, Thinker-Talker split, frozen encoders, and full release of code, checkpoints, and training data.
Mutual Forcing trains a single native autoregressive audio-video model with mutually reinforcing few-step and multi-step modes via self-distillation to match 50-step baselines at 4-8 steps.
TASU2 adds controllability over uncertainty and error rate to text-derived CTC simulation, enabling better cross-modal alignment and low-resource adaptation for speech LLMs than prior text-only or TTS methods.
OmniVoice introduces a diffusion language model-style non-autoregressive TTS system that directly maps text to multi-codebook acoustic tokens, scaling zero-shot synthesis to over 600 languages with SOTA results on multilingual benchmarks using 581k hours of open data.
HumanOmni-Speaker introduces a Visual Delta Encoder and VR-SDR benchmark that enable end-to-end speaker diarization and recognition by sampling video at 25 fps and compressing inter-frame motion residuals into 6 tokens per frame.
Omni Parsing framework converts complex multimodal signals into locatable, enumerable, and traceable structured knowledge via hierarchical detection, recognition, and interpreting with strict evidence alignment.
TW-Sound580K dataset plus Tai-LALM model with dynamic Dual-ASR arbitration lifts localized Taiwanese audio-language accuracy to 49.1% on the TAU benchmark.
Q2D2 uses 2D geometric grid projections to quantize feature pairs in neural audio codecs, yielding implicit codebooks that improve efficiency and utilization over RVQ, VQ, and FSQ while maintaining reconstruction quality.
SenSE adds language-model semantic guidance to flow-matching generative speech enhancement via a dual-path masked conditioning strategy and reports SOTA results on distorted speech.
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 before fine-tuning on conversational speech.
Using model confusion patterns from counterfactual audio conditions to curate training data improves audio question answering accuracy from 65.90% to 67.27% on the development set.
A survey of Large Audio Language Models that establishes a taxonomy of trustworthiness vulnerabilities and proposes a Defense-in-Depth roadmap for audio intelligence.
SQuTR is a large bilingual benchmark of 37,317 synthesized spoken queries under clean/low/medium/high noise, showing that retrieval quality steadily degrades as noise increases.
LLM-based multi-talker ASR with dual-encoder, feature interleaving, length-aware speaker loss, and adaptive ASR threshold achieves 18% and 24% relative gains over baselines on AliMeeting and Aishell4.
Small Qwen3-ASR models fine-tuned on LLM-rewritten spoken-to-formal datasets achieve up to 37.4% relative CER reduction versus verbatim baselines while improving formal text metrics.
citing papers explorer
-
OmniHuman: A Large-scale Dataset and Benchmark for Human-Centric Video Generation
A large-scale, hierarchically-annotated human-centric video dataset and a three-level benchmark (OHBench) are presented, showing that fine-tuning on a subset of the data improves audio-video generation model performance.
-
Mutual Forcing: Dual-Mode Self-Evolution for Fast Autoregressive Audio-Video Character Generation
Mutual Forcing trains a single native autoregressive audio-video model with mutually reinforcing few-step and multi-step modes via self-distillation to match 50-step baselines at 4-8 steps.
-
OmniVoice: Towards Omnilingual Zero-Shot Text-to-Speech with Diffusion Language Models
OmniVoice introduces a diffusion language model-style non-autoregressive TTS system that directly maps text to multi-codebook acoustic tokens, scaling zero-shot synthesis to over 600 languages with SOTA results on multilingual benchmarks using 581k hours of open data.