Find centralized, trusted content and collaborate around the technologies you use most.
Teams
Q&A for work
Connect and share knowledge within a single location that is structured and easy to search.
全卷积网络使用 softmaxloss 作为其损失层,但语义识别是一个多标签任务,不能将其视为概率问题。我理解错了吗?
softmax 损失是每像素:输出(预测)概率是长度为 #classes per-pixel的向量,损失是"SoftmaxWithLoss" per-pixel。
"SoftmaxWithLoss"