用表示空间的弗雷谢距离优化生成模型，提升图像质量。
Representation Fréchet Loss for Visual Generation
- 将弗雷谢距离分解为采样数和梯度批量，实现有效训练
- 单步生成器在ImageNet上达到0.72 FID，优于传统方法
- 无需对抗训练或教师模型，适合追求高效生成的研究者
我们发现长期被视为不可行的弗雷谢距离（FD），实际上可在表示空间中有效优化。核心思想是将用于估计的总体规模（如50k）与用于梯度计算的批量大小（如1024）解耦。该方法称为FD-loss。优化FD-loss揭示多个意外发现：第一，在Inception特征空间中，仅需一次后训练的基线生成器即可在ImageNet 256x256上达到0.72 FID；第二，相同FD-loss可将多步生成器转化为强单步生成器，无需教师蒸馏、对抗训练或样本级目标；第三，FID可能误判视觉质量：现代表示虽使Inception FID更差，但生成样本更优。这促使我们提出多表示评估指标FDr$^k$。我们希望本工作能推动分布距离在多样化表示空间中作为训练目标和评估指标的进一步探索。
原文摘要 · Abstract (English)
We show that Fréchet Distance (FD), long considered impractical as a training objective, can in fact be effectively optimized in the representation space. Our idea is simple: decouple the population size for FD estimation (e.g., 50k) from the batch size for gradient computation (e.g., 1024). We term this approach FD-loss. Optimizing FD-loss reveals several surprising findings. First, post-training a base generator with FD-loss in different representation spaces consistently improves visual quality. Under the Inception feature space, a one-step generator achieves0.72 FID on ImageNet 256x256. Second, the same FD-loss repurposes multi-step generators into strong one-step generators without teacher distillation, adversarial training or per-sample targets. Third, FID can misrank visual quality: modern representations can yield better samples despite worse Inception FID. This motivates FDr$^k$, a multi-representation metric. We hope this work will encourage further exploration of distributional distances in diverse representation spaces as both training objectives and evaluation metrics for generative models.
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。