Quant VideoGen: Auto-Regressive Long Video Generation via 2-Bit KV-Cache Quantization
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866915982816051200 |
|---|---|
| author | Xi, Haocheng Yang, Shuo Zhao, Yilong Li, Muyang Cai, Han Li, Xingyang Lin, Yujun Zhang, Zhuoyang Zhang, Jintao Li, Xiuyu Xu, Zhiying Wu, Jun Xu, Chenfeng Stoica, Ion Han, Song Keutzer, Kurt |
| author_facet | Xi, Haocheng Yang, Shuo Zhao, Yilong Li, Muyang Cai, Han Li, Xingyang Lin, Yujun Zhang, Zhuoyang Zhang, Jintao Li, Xiuyu Xu, Zhiying Wu, Jun Xu, Chenfeng Stoica, Ion Han, Song Keutzer, Kurt |
| contents | Despite rapid progress in autoregressive video diffusion, an emerging system algorithm bottleneck limits both deployability and generation capability: KV cache memory. In autoregressive video generation models, the KV cache grows with generation history and quickly dominates GPU memory, often exceeding 30 GB, preventing deployment on widely available hardware. More critically, constrained KV cache budgets restrict the effective working memory, directly degrading long horizon consistency in identity, layout, and motion. To address this challenge, we present Quant VideoGen (QVG), a training free KV cache quantization framework for autoregressive video diffusion models. QVG leverages video spatiotemporal redundancy through Semantic Aware Smoothing, producing low magnitude, quantization friendly residuals. It further introduces Progressive Residual Quantization, a coarse to fine multi stage scheme that reduces quantization error while enabling a smooth quality memory trade off. Across LongCat Video, HY WorldPlay, and Self Forcing benchmarks, QVG establishes a new Pareto frontier between quality and memory efficiency, reducing KV cache memory by up to 7.0 times with less than 4% end to end latency overhead while consistently outperforming existing baselines in generation quality. Code is available at: https://github.com/svg-project/Quant-VideoGen |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_02958 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Quant VideoGen: Auto-Regressive Long Video Generation via 2-Bit KV-Cache Quantization Xi, Haocheng Yang, Shuo Zhao, Yilong Li, Muyang Cai, Han Li, Xingyang Lin, Yujun Zhang, Zhuoyang Zhang, Jintao Li, Xiuyu Xu, Zhiying Wu, Jun Xu, Chenfeng Stoica, Ion Han, Song Keutzer, Kurt Machine Learning Despite rapid progress in autoregressive video diffusion, an emerging system algorithm bottleneck limits both deployability and generation capability: KV cache memory. In autoregressive video generation models, the KV cache grows with generation history and quickly dominates GPU memory, often exceeding 30 GB, preventing deployment on widely available hardware. More critically, constrained KV cache budgets restrict the effective working memory, directly degrading long horizon consistency in identity, layout, and motion. To address this challenge, we present Quant VideoGen (QVG), a training free KV cache quantization framework for autoregressive video diffusion models. QVG leverages video spatiotemporal redundancy through Semantic Aware Smoothing, producing low magnitude, quantization friendly residuals. It further introduces Progressive Residual Quantization, a coarse to fine multi stage scheme that reduces quantization error while enabling a smooth quality memory trade off. Across LongCat Video, HY WorldPlay, and Self Forcing benchmarks, QVG establishes a new Pareto frontier between quality and memory efficiency, reducing KV cache memory by up to 7.0 times with less than 4% end to end latency overhead while consistently outperforming existing baselines in generation quality. Code is available at: https://github.com/svg-project/Quant-VideoGen |
| title | Quant VideoGen: Auto-Regressive Long Video Generation via 2-Bit KV-Cache Quantization |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2602.02958 |