Pytorch相关笔记 torch.nn nn.linear：输入张量的形状: [batch_size, in_features] 输出张量的形状: [batch_size, out_features] 输入 x (1,4)： A的转置 (4,2)：