Closed 975150313 closed 4 years ago
你好,这是两类分割,如果想要多分类的话,可以参考这个项目https://github.com/junqiangchen/BraTS18-Challege
你好,这是两类分割,如果想要多分类的话,可以参考这个项目https://github.com/junqiangchen/BraTS18-Challege
感谢您的回复,我查阅资料了解到,通过修改损失函数,我已经实现了,多标签分类和二分类损失函数是相同的,我把损失函数修改为binary_crossentropy ,类别数量改成我自己的就可以
请问一下如何进行多标签分类,n_class设置不起作用 ,logits,labels怎么处理以适应多标签分类?