# TurboT2VA: Fast Large-Scale Text-to-Video-Audio Generation via Score-Regularized Consistency Distillation

arXiv: https://arxiv.org/abs/2608.24674 · HTML: https://mlsanity.com/paper/2608.24674/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 26 of its 32 references that are in the map.
- First posted on arXiv 25 Aug 2026.

## Builds on (most cited first, 26 of 26)

- [CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer](https://mlsanity.com/paper/2408.06072/) (2024)
- [Consistency Models](https://mlsanity.com/paper/2303.01469/) (2023)
- [VBench: Comprehensive Benchmark Suite for Video Generative Models](https://mlsanity.com/paper/2311.17982/) (2023)
- [Make-A-Video: Text-to-Video Generation without Text-Video Data](https://mlsanity.com/paper/2209.14792/) (2022)
- [ImageBind: One Embedding Space To Bind Them All](https://mlsanity.com/paper/2305.05665/) (2023)
- [One-step Diffusion with Distribution Matching Distillation](https://mlsanity.com/paper/2311.18828/) (2023)
- [Improved Distribution Matching Distillation for Fast Image Synthesis](https://mlsanity.com/paper/2405.14867/) (2024)
- [Open-Sora: Democratizing Efficient Video Production for All](https://mlsanity.com/paper/2412.20404/) (2024)
- [LTX-Video: Realtime Video Latent Diffusion](https://mlsanity.com/paper/2501.00103/) (2024)
- [Simplifying, Stabilizing and Scaling Continuous-Time Consistency Models](https://mlsanity.com/paper/2410.11081/) (2024)
- [LTX-2: Efficient Joint Audio-Visual Foundation Model](https://mlsanity.com/paper/2601.03233/) (2026)
- [SageAttention: Accurate 8-Bit Attention for Plug-and-play Inference Acceleration](https://mlsanity.com/paper/2410.02367/) (2024)
- [MMAudio: Taming Multimodal Joint Training for High-Quality Video-to-Audio Synthesis](https://mlsanity.com/paper/2412.15322/) (2024)
- [FoleyCrafter: Bring Silent Videos to Life with Lifelike and Synchronized Sounds](https://mlsanity.com/paper/2407.01494/) (2024)
- [JavisDiT: Joint Audio-Video Diffusion Transformer with Hierarchical Spatio-Temporal Prior Synchronization](https://mlsanity.com/paper/2503.23377/) (2025)
- [Large Scale Diffusion Distillation via Score-Regularized Continuous-Time Consistency](https://mlsanity.com/paper/2510.08431/) (2025)
- [Ovi: Twin Backbone Cross-Modal Fusion for Audio-Video Generation](https://mlsanity.com/paper/2510.01284/) (2025)
- [Diverse and Aligned Audio-to-Video Generation via Text-to-Video Model Adaptation](https://mlsanity.com/paper/2309.16429/) (2023)
- [SLA: Beyond Sparsity in Diffusion Transformers via Fine-Tunable Sparse-Linear Attention](https://mlsanity.com/paper/2509.24006/) (2025)
- [Read, Watch and Scream! Sound Generation from Text and Video](https://mlsanity.com/paper/2407.05551/) (2024)
- [The Power of Sound (TPoS): Audio Reactive Video Generation with Stable Diffusion](https://mlsanity.com/paper/2309.04509/) (2023)
- [TurboDiffusion: Accelerating Video Diffusion Models by 100-200 Times](https://mlsanity.com/paper/2512.16093/) (2025)
- [Speed by Simplicity: A Single-Stream Architecture for Fast Audio-Video Generative Foundation Model](https://mlsanity.com/paper/2603.21986/) (2026)
- [TTA-Bench: A Comprehensive Benchmark for Evaluating Text-to-Audio Models](https://mlsanity.com/paper/2509.02398/) (2025)
- [A Progressive Training Strategy for Embodied Vision-Language Models to Mitigate Spatio-Temporal Hallucinations](https://mlsanity.com/paper/2604.10506/) (2026)
- [TMD-Bench: A Multi-Level Evaluation Paradigm for Music-Dance Co-Generation](https://mlsanity.com/paper/2605.01809/) (2026)

## Abstract (arXiv)

Joint text-to-video-audio generation produces synchronized visual and acoustic content, but the long sampling trajectories and heterogeneous multimodal computation of large models make inference prohibitively expensive. We present TurboT2VA, a distillation and inference framework for accelerating a 19B-parameter joint video-audio model. Large-scale T2VA distillation is challenged by modality-imbalanced optimization, the difficulty of continuous-time consistency training at scale, and the quality--diversity trade-off. TurboT2VA addresses these issues with per-modality normalization and a progressive curriculum comprising discrete consistency warm-up, continuous consistency refinement, and joint consistency--distribution matching. The curriculum first establishes a stable, diverse generation trajectory and only then introduces distribution-level refinement. On LTX-2, four-step distillation reduces generator latency from 50.52s to 2.51s at the standard evaluation resolution of 512×768, achieving a 20.1× speedup while maintaining strong visual quality, audio fidelity, diversity, and video-audio synchronization. We further develop an architecture-aware inference stack that combines guarded W8A8 and fused operators, padded-text compaction, and modality-aware sparse attention while preserving dense cross-modal and text-conditioning paths. Under the high-resolution deployment setting at 1024×1792, the complete stack reduces generator latency from 318.74s to 5.83s on one NVIDIA H20, achieving a 54.67× generator-only speedup. Inference code and generation demos are available at https://github.com/thu-ml/TurboDiffusion/tree/main/turbot2va.