search

Tencent / NeuralNLP-NeuralClassifier

An Open-source Neural Hierarchical Multi-label Text Classification Toolkit
Other
1.85k stars 406 forks source link

About Focal Loss #35

Closed tc-yue closed 4 years ago

tc-yue commented 4 years ago

您好，代码中有SIgmoidFocal Loss和 Softmax FocalLoss，请问有这两种loss的实验结果对比吗

liqunhit commented 4 years ago

有过一些离线实验，对于单标签分类任务，这两个loss效果基本持平；对于多标签分类任务，必须使用SigmoidFocal Loss。