通过中间层对齐速度特征，让生成模型训练更快更准。
Deeply Supervised Flow-Based Generative Models
- 在Transformer分支间插入轻量级速度校正模块，实现深层监督
- ImageNet上收敛速度提升8倍，训练时间减半且FID降低2.6
- 适合追求高效生成与高质量图像的开发者和研究者
基于流的生成模型通过学习线性插值的速度表示，在多个视觉生成任务中表现优异。然而我们发现，仅从最终层输出训练速度表示，未能充分利用中间层丰富的表征，可能阻碍模型收敛。为此，我们提出DeepFlow框架，通过分枝结构与深度监督，引入轻量级速度校正模块VeRA，对齐各层内部速度特征。得益于内部速度对齐带来的增强深度监督，DeepFlow在ImageNet上收敛速度比之前流模型快8倍，训练时间减半，且保持相当性能的同时将FID降低2.6。在文本到图像生成任务中，DeepFlow在MSCOCO和零样本GenEval评估中也优于基线模型。
原文摘要 · Abstract (English)
Flow based generative models have charted an impressive path across multiple visual generation tasks by adhering to a simple principle: learning velocity representations of a linear interpolant. However, we observe that training velocity solely from the final layer output underutilizes the rich inter layer representations, potentially impeding model convergence. To address this limitation, we introduce DeepFlow, a novel framework that enhances velocity representation through inter layer communication. DeepFlow partitions transformer layers into balanced branches with deep supervision and inserts a lightweight Velocity Refiner with Acceleration (VeRA) block between adjacent branches, which aligns the intermediate velocity features within transformer blocks. Powered by the improved deep supervision via the internal velocity alignment, DeepFlow converges 8 times faster on ImageNet with equivalent performance and further reduces FID by 2.6 while halving training time compared to previous flow based models without a classifier free guidance. DeepFlow also outperforms baselines in text to image generation tasks, as evidenced by evaluations on MSCOCO and zero shot GenEval.
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。