S2Q-VDiT: Accurate Quantized Video Diffusion Transformer with Salient Data and Sparse Token Distillation

Weilun Feng; Haotong Qin; Chuanguang Yang; Xiangqi Li; Han Yang; Yuqi Li; Zhulin An; Libo Huang; Michele Magno; Yongjun Xu

doi:10.48550/arxiv.2508.04016

NeurIPS2025顶会

S2Q-VDiT: Accurate Quantized Video Diffusion Transformer with Salient Data and Sparse Token Distillation

Weilun Feng, Haotong Qin, Chuanguang Yang, Xiangqi Li, Han Yang, Yuqi Li, Zhulin An, Libo Huang, Michele Magno, Yongjun Xu

arXiv DOI 出版方

2025年份

1被引次数

3顶会引用

摘要

a stylish woman walks down a Tokyo street filled with warm glowing neon and animated city signage.

Figure 1: We present S 2 Q-VDiT, a post-training quantization method for video diffusion transformers. We quantize HunyuanVideo [24] to 4-bit weights and 6-bit activations without compromising visual quality. S 2 Q-VDiT can further achieve 3.9× model compression and 1.3× inference acceleration.