29 papers
Beyond Reconstruction: Full-Context Generative DiT for Music Generation
Yunjia Li, Menglin Wu, Junyu Dai +13
Hybrid music generators combine the long-range planning of an autoregressive language model with the fidelity of a diffusion- or flow-based acoustic renderer. Yet renderers are tra…
Pushing the Frontier of Full-Song Generation: Hierarchical Autoregressive Planning Meets Flow-Matching Rendering
Junyu Dai, Xinyue Fan, Weiqin Li +14
In this report, we present a unified song generation framework capable of producing high-quality full-length music from lyrics, text descriptions, and musical attributes. The propo…
Qwen-Audio-3.0-TTS: Freely Controllable and Highly Robust Speech Synthesis with Multi-Stage Training Paradigm
Bajian Xiang, Cheng Wen, Han Zhao +12
In this report, we present Qwen-Audio-3.0-TTS, a production-oriented speech synthesis system that jointly advances content consistency, speaker similarity, prosodic naturalness, au…
TF-MossFormer: Integrating Convolution Gated Local-Global Attentions for Enhanced Time-Frequency Domain Monaural Speech Separation
Shengkui Zhao, Zexu Pan, Haoxu Wang +3
Transformers with global attention capture long-range dependencies but can miss the fine-grained local continuity crucial for speech separation. We propose TF-MossFormer, a time-fr…
Vidu S1: A Real-Time Interactive Video Generation Model
Jintao Zhang, Kai Jiang, Jintao Chen +24
We introduce Vidu S1, a real-time interactive video generation model supporting voice control of digital characters. Users can control video generation content at any moment throug…
FlowTTS-GRPO: Online Reinforcement Learning with Multi-Objective Reward Optimization for Flow-Matching Based Text-to-Speech
Haoxu Wang, Biao Tian, Weiqin Li +3
Existing Reinforcement Learning (RL) research for Text-to-Speech (TTS) focuses on large language models (LLMs), leaving Flow-Matching (FM) under-explored. We present FlowTTS-GRPO,…
TurboServe: Serving Streaming Video Generation Efficiently and Economically
Youhe Jiang, Haoxu Wang, Haotong Bao +5
Streaming video generation is emerging as a new serving workload in which users interact with long-lived sessions that generate video progressively, chunk by chunk. Unlike offline…
KernelBenchX: A Comprehensive Benchmark for Evaluating LLM-Generated GPU Kernels
Han Wang, Jintao Zhang, Kai Jiang +3
LLM-based Triton kernel generation has attracted significant interest, yet a fundamental empirical question remains unanswered: where does this capability break down, and why? We p…
HieraSparse: Hierarchical Semi-Structured Sparse KV Attention
Haoxuan Wang, Chen Wang
The deployment of long-context Large Language Models (LLMs) poses significant challenges due to the intense computational cost of self-attention and the substantial memory overhead…
SageBwd: A Trainable Low-bit Attention
Jintao Zhang, Marco Chen, Haoxu Wang +5
Low-bit attention, such as SageAttention, has emerged as an effective approach for accelerating model inference, but its applicability to training remains poorly understood. In pri…
SLA2: Sparse-Linear Attention with Learnable Routing and Quantization-aware Training
Jintao Zhang, Haoxu Wang, Kai Jiang +6
Sparse-Linear Attention (SLA) combines sparse and linear attention to accelerate diffusion models and has shown strong performance in video generation. However, (i) SLA relies on a…
LuSeeL: Language-queried Binaural Universal Sound Event Extraction and Localization
Zexu Pan, Shengkui Zhao, Yukun Ma +4
Most universal sound extraction algorithms focus on isolating a target sound event from single-channel audio mixtures. However, the real world is three-dimensional, and binaural au…
E2E-AEC: Implementing an end-to-end neural network learning approach for acoustic echo cancellation
Yiheng Jiang, Biao Tian, Haoxu Wang +4
We propose a novel neural network-based end-to-end acoustic echo cancellation (E2E-AEC) method capable of streaming inference, which operates effectively without reliance on tradit…
FlowSE-GRPO: Training Flow Matching Speech Enhancement via Online Reinforcement Learning
Haoxu Wang, Biao Tian, Yiheng Jiang +5
Generative speech enhancement offers a promising alternative to traditional discriminative methods by modeling the distribution of clean speech conditioned on noisy inputs. Post-tr…
TurboDiffusion: Accelerating Video Diffusion Models by 100-200 Times
Jintao Zhang, Kaiwen Zheng, Kai Jiang +5
We introduce TurboDiffusion, a video generation acceleration framework that can speed up end-to-end diffusion generation by 100-200x while maintaining video quality. TurboDiffusion…
SLA: Beyond Sparsity in Diffusion Transformers via Fine-Tunable Sparse-Linear Attention
Jintao Zhang, Haoxu Wang, Kai Jiang +10
In Diffusion Transformer (DiT) models, particularly for video generation, attention latency is a major bottleneck due to the long sequence length and the quadratic complexity. We f…
MELA-TTS: Joint transformer-diffusion model with representation alignment for speech synthesis
Keyu An, Zhiyu Zhang, Changfeng Gao +7
This work introduces MELA-TTS, a novel joint transformer-diffusion framework for end-to-end text-to-speech synthesis. By autoregressively generating continuous mel-spectrogram fram…
FLASepformer: Efficient Speech Separation with Gated Focused Linear Attention Transformer
Haoxu Wang, Yiheng Jiang, Gang Qiao +2
Speech separation always faces the challenge of handling prolonged time sequences. Past methods try to reduce sequence lengths and use the Transformer to capture global information…
Exploring Efficient Directional and Distance Cues for Regional Speech Separation
Yiheng Jiang, Haoxu Wang, Yafeng Chen +2
In this paper, we introduce a neural network-based method for regional speech separation using a microphone array. This approach leverages novel spatial cues to extract the sound s…
SageAttention3: Microscaling FP4 Attention for Inference and An Exploration of 8-Bit Training
Jintao Zhang, Jia Wei, Pengle Zhang +6
The efficiency of attention is important due to its quadratic time complexity. We enhance the efficiency of attention through two key contributions: First, we leverage the new FP4…
Understanding Zero-shot Rare Word Recognition Improvements Through LLM Integration
Haoxuan Wang
In this study, we investigate the integration of a large language model (LLM) with an automatic speech recognition (ASR) system, specifically focusing on enhancing rare word recogn…
ZipEnhancer: Dual-Path Down-Up Sampling-based Zipformer for Monaural Speech Enhancement
Haoxu Wang, Biao Tian
In contrast to other sequence tasks modeling hidden layer features with three axes, Dual-Path time and time-frequency domain speech enhancement models are effective and have low pa…
Memory-Efficient and Secure DNN Inference on TrustZone-enabled Consumer IoT Devices
Xueshuo Xie, Haoxu Wang, Zhaolong Jian +4
Edge intelligence enables resource-demanding Deep Neural Network (DNN) inference without transferring original data, addressing concerns about data privacy in consumer Internet of…
Robust Wake Word Spotting With Frame-Level Cross-Modal Attention Based Audio-Visual Conformer
Haoxu Wang, Ming Cheng, Qiang Fu +1
In recent years, neural network-based Wake Word Spotting achieves good performance on clean audio samples but struggles in noisy environments. Audio-Visual Wake Word Spotting (AVWW…