作者：?頭孢就酒的快樂神仙

轉載地址：https://bbs.huaweicloud.com/forum/thread-146576-1-1.html

在復現High-resolution NetWork（HRNet）用于語義分割時，Cityscapes數據集不同類別的物體在計算損失時賦有不同的權重。

weights_list = [0.8373, 0.918, 0.866, 1.0345, 1.0166,0.9969, 0.9754, 1.0489,0.8786, 1.0023, 0.9539, 0.9843,1.1116, 0.9037, 1.0865, 1.0955,1.0865, 1.1529, 1.0507]

在PyTorch中提供torch.nn.CrossEntropyLoss(weight=weight, ignore_index=ignore_label)可用于實現不同類別的加權計算。

MindSpore的r1.1和r1.2版本并未提供類似功能的損失函數。可以用以下代碼實現：

class CrossEntropyLossWithWeights(_Loss):def __init__(self, weights, num_classes=19, ignore_label=255):super(CrossEntropyLossWithWeights, self).__init__()self.weights = weightsself.resize = F.ResizeBilinear(cfg.train.image_size)self.one_hot = P.OneHot(axis=-1)self.on_value = Tensor(1.0, mstype.float32)self.off_value = Tensor(0.0, mstype.float32)self.cast = P.Cast()self.ce = nn.SoftmaxCrossEntropyWithLogits()self.zeros = F.Zeros()self.fill = F.Fill()self.equal = F.Equal()self.select = F.Select()self.num_classes = num_classesself.ignore_label = ignore_labelself.mul = P.Mul()self.argmax = P.Argmax(output_type=mstype.int32)self.sum = P.ReduceSum(False)self.div = P.RealDiv()self.transpose = P.Transpose()self.reshape = P.Reshape()def construct(self, logits, labels):logits = self.resize(logits)labels_int = self.cast(labels, mstype.int32)labels_int = self.reshape(labels_int, (-1,))logits_ = self.transpose(logits, (0, 2, 3, 1)) # (12, 1024, 2048, 19)logits_ = self.reshape(logits_, (-1, self.num_classes))labels_float = self.cast(labels_int, mstype.float32)weights = self.zeros(labels_float.shape, mstype.float32)for i in range(self.num_classes):fill_weight = self.fill(mstype.float32, labels_float.shape, self.weights[i])equal_ = self.equal(labels_float, i)weights = self.select(equal_, fill_weight, weights)one_hot_labels = self.one_hot(labels_int, self.num_classes, self.on_value, self.off_value)loss = self.ce(logits_, one_hot_labels)loss = self.mul(weights, loss)loss = self.div(self.sum(loss), self.sum(weights))return loss

總結

以上是生活随笔為你收集整理的CV语义分割，类别加权损失函数的全部內容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網站內容還不錯，歡迎將生活随笔推薦給好友。