torch.sigmoid()
torch.nn.Sigmoid()
CODE
torch.sigmoid()
import torch
a = torch.randn(4)
print(a)
print(torch.sigmoid(a))
tensor([ 1.2622, -0.3935, -0.0859, 1.0130])
tensor([0.7794, 0.4029, 0.4785, 0.7336])
torch.nn.Sigmoid()
import torch
a = torch.tensor([1.2622, -0.3935, -0.0859, 1.0130])
m = torch.nn.Sigmoid()
b = m(a)
print(a)
print(b)
tensor([ 1.2622, -0.3935, -0.0859, 1.0130])
tensor([0.7794, 0.4029, 0.4785, 0.7336])
torch.nn.Sigmoid,首字母S大写,是一个类。
torch.sigmoid,首字母s小写,是一个方法。
举例:
def structure_loss(pred, mask):
weit = 1+5*torch.abs(F.avg_pool2d(mask, kernel_size=31, stride=1, padding=15)-mask)
wbce = F.binary_cross_entropy_with_logits(pred, mask, reduction='none')
wbce = (weit*wbce).sum(dim=(2,3))/weit.sum(dim=(2,3))
pred = torch.sigmoid(pred)
inter = ((pred * mask) * weit).sum(dim=(2, 3))
union = ((pred + mask) * weit).sum(dim=(2, 3))
wiou = 1-(inter+1)/(union-inter+1)
return (wbce+wiou).mean()
Pytorch:激活函数及其梯度 (Tanh函数 ReLU函数)
Tanh函数
ReLU函数