OpenAI参与，重卷ImageNet：终于把FID做成训练
中文摘要
OpenAI参与研究，小模型图像生成质量显著提升，FID分数突破0.8。
English Summary
OpenAI and others improved ImageNet results by incorporating FID into training, enabling small models to achieve FID scores below 0.8.
原文节选
新方法让小模型FID也能破0.8
中文摘要
OpenAI参与研究，小模型图像生成质量显著提升，FID分数突破0.8。
English Summary
OpenAI and others improved ImageNet results by incorporating FID into training, enabling small models to achieve FID scores below 0.8.
新方法让小模型FID也能破0.8