针对视频生成模型，提出保护首尾层的量化方法，实现零精度损失。
Boundary-Protection W8A8 HiFloat8 Quantization for Large-Scale Text-to-Video Diffusion Transformers
- 按各层激活分布差异，保护首两末三共五层不量化
- 在5个评测维度上达到原精度基准水平，无明显性能下降
- 适合部署大规模文生视频模型到昇腾910B硬件的场景
我们提出一种针对140亿参数文生视频扩散模型Wan2.1-T2V-14B的后训练量化（PTQ）方法，目标为在昇腾910B NPU上采用W8A8 HiFloat8（HiF8）格式。视频DiT模型中，不同层的激活分布差异显著：首尾几层与中间层统计特性完全不同，导致统一量化效果差。通过对全部40个WanAttentionBlock进行逐层激活分析，提出边界保护策略——保留前两个和后三个块为BF16，其余35个块采用W8A8 HiF8量化。该方法在五个VBench评测维度上与原始BF16基线持平或略有超越，5个提示词测试集下无明显精度损失。消融实验表明，完整边界保护配置获得最高平均VBench分数，验证了数据驱动选择的有效性。此外，研究了量化感知训练（QAT）作为补充微调阶段，发现其在单卡硬件条件下难以超越纯PTQ。
原文摘要 · Abstract (English)
We present a post-training quantization (PTQ) approach for Wan2.1-T2V-14B, a 14-billion-parameter text-to-video diffusion transformer, targeting the W8A8 HiFloat8 (HiF8) format on Ascend 910B NPUs. A central challenge in quantizing video DiT models is the heterogeneous activation distribution across transformer blocks: boundary blocks (the first and last few blocks) exhibit fundamentally different statistical properties from middle blocks, making uniform quantization ineffective. We conduct a systematic per-block activation analysis across all 40 WanAttentionBlocks and use the findings to motivate a boundary-protection strategy that retains the first two and last three blocks in BF16 while quantizing the remaining 35 blocks with W8A8 HiF8. The proposed PTQ method matches or marginally exceeds the BF16 baseline on all five VBench dimensions evaluated, indicating no measurable accuracy loss within the 5-prompt evaluation set. An ablation study over four protection configurations confirms that full boundary protection yields the highest average VBench score, validating the data-driven block selection. We additionally investigate quantization-aware training (QAT) as a complementary fine-tuning stage and analyze the conditions under which it fails to outperform plain PTQ on single-card hardware.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。