Quant VideoGen：通过2位KV缓存量化实现自回归长视频生成
Quant VideoGen: Auto-Regressive Long Video Generation via 2-Bit KV-Cache Quantization
AI总结：
本文提出Quant VideoGen，通过语义感知平滑和渐进残差量化技术，有效降低KV缓存内存消耗，提升长视频生成质量与效率。
摘要：
尽管自回归视频扩散模型在生成视频方面取得了快速进展，但KV缓存内存问题成为部署和生成能力的瓶颈。在自回归视频生成模型中，KV缓存随着生成历史增长，迅速占据GPU内存，常超过30GB，限制了在广泛可用硬件上的部署。更关键的是，受限的KV缓存预算直接降低了身份、布局和运动的长时一致性。为解决这一挑战，我们提出了Quant VideoGen（QVG），一种无需训练的KV缓存量化框架。QVG通过语义感知平滑利用视频时空冗余，生成低幅值、量化友好的残差。它进一步引入渐进残差量化，一种从粗到细的多阶段方案，减少量化误差的同时实现平滑的质量内存权衡。在LongCat Video、HY WorldPlay和Self Forcing基准测试中，QVG在质量和内存效率之间建立了新的帕累托前沿，将KV缓存内存减少高达7.0倍，同时端到端延迟开销低于4%，并且在生成质量上持续优于现有基线。代码可在：https://github.com/svg-project/Quant-VideoGen获取。
英文摘要：
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
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
