---
pipeline_tag: text-to-video
---

# TurboT2VA

TurboT2VA is a fast text-to-video-audio generation model distilled from LTX-2 19B, producing synchronized visual and acoustic content in just 4 steps. This repository contains the TurboT2AV-14BVideo-5BAudio checkpoint.

- **Paper**: [TurboT2VA: Fast Large-Scale Text-to-Video-Audio Generation via Score-Regularized Consistency Distillation](https://huggingface.co/papers/2608.24674)
- **Code**: [TurboDiffusion/TurboT2AV](https://github.com/thu-ml/TurboDiffusion/tree/main/turbot2va)

For inference instructions and demos, please refer to the GitHub repository.