torch.matmul

import torch
x = torch.tensor([[[1,  2,  3, 4],
         [5, 6, 7, 8],
         [ 9, 10, 11, 12]],

        [[-1, -2, -3, -4],
         [-5, -6, -7, -8],
         [-9, -10, -11, -12]]])

y = torch.tensor([1,1,1])
print(x.size(), y.size())
print(torch.matmul(y, x),'\n',torch.matmul(y, x).size()) #1D*3D


输出:
torch.Size([2, 3, 4]) torch.Size([3])
tensor([[ 15,  18,  21,  24],
        [-15, -18, -21, -24]]) 
 torch.Size([2, 4])


x = torch.tensor([[[1,  2,  3, 4],
         [5, 6, 7, 8],
         [ 9, 10, 11, 12]],

        [[-1, -2, -3, -4],
         [-5, -6, -7, -8],
         [-9, -10, -11, -12]]])

y = torch.tensor([1,1,1,1])
print(x.size(), y.size())
print(torch.matmul(x, y),'\n',torch.matmul(x, y).size()) # 3D*1D

输出:
torch.Size([2, 3, 4]) torch.Size([4])
tensor([[ 10,  26,  42],
        [-10, -26, -42]]) 
 torch.Size([2, 3])

x维度为[2,3,4] 的时候，y维度为[3]的时候，y可以左乘x，y扩充为[2, 1,3] 维度，输出结果为 [2, 4]矩阵

x维度为[2,3,4] 的时候，y维度为[4]的时候，y可以左乘x，y扩充为[2, 4,1] 维度，输出结果为 [2, 3]矩阵，类似于下面的方法，最后输出减去一维

x为3维 y为2维不能相乘

import torch
x = torch.tensor([[[1,  2,  3, 4],
         [5, 6, 7, 8],
         [ 9, 10, 11, 12]],

        [[-1, -2, -3, -4],
         [-5, -6, -7, -8],
         [-9, -10, -11, -12]]])

y = torch.tensor([[[1],[1],[1],[1]],
                [[1],[1],[1],[1]]])
print(x.size(), y.size())
print(torch.matmul(x, y),'\n',torch.matmul(x, y).size()) # 3D*1D


输出:

torch.Size([2, 3, 4]) torch.Size([2, 4, 1])
tensor([[[ 10],
         [ 26],
         [ 42]],

        [[-10],
         [-26],
         [-42]]]) 
 torch.Size([2, 3, 1])