1.用法:
G_skeleton, D_skeleton, start_epoch = self.build_models()
optimizerG_skeleton, optimizerD_skeleton = self.define_optimizers(G_skeleton, D_skeleton)
criterion = nn.BCELoss()
其中G_skeleton
和 D_skeleton
是我们用到的模型。使用以下代码打印参数总数:
# 打印G和D的总参数数量
print("Total number of param in Generator is ", sum(x.numel() for x in G_skeleton.parameters()))
print("Total number of param in Discriminator is ", sum(x.numel() for x in D_skeleton.parameters()))
2.解析:
my_model.parameters()
:用来返回模型中的参数
numel()
:获取tensor中一共包含多少个元素
例:
sum()
:python内置函数,对元组或列表求和