广播机制在什么条件下可以应用?

如果两个数组的后缘维度,即从末尾开始算起的维度的轴长度相符，或其中的一方的长度为1，则认为它们是广播兼容的。

1.后缘维度的轴长是相同的。

eg1:

x=torch.arange(6).reshape(2,3)
y=torch.tensor([1,2,3])
y.shape,x.shape,x,y,x+y
-------------------------------
(torch.Size([3]),
 torch.Size([2, 3]),
 tensor([[0, 1, 2],
         [3, 4, 5]]),
 tensor([1, 2, 3]),
 tensor([[1, 3, 5],
         [4, 6, 8]]))
#缺省也是没问题的

eg2:

x=torch.arange(6).reshape(2,3)
y=torch.tensor([[1,2,3]])
y.shape,x.shape,x,y,x+y
-------------------------------
(torch.Size([1, 3]),
 torch.Size([2, 3]),
 tensor([[0, 1, 2],
         [3, 4, 5]]),
 tensor([[1, 2, 3]]),
 tensor([[1, 3, 5],
         [4, 6, 8]]))
#其实可以看成第二种情况

eg3:

x=torch.arange(12).reshape(2,2,3)
y=torch.tensor([1,2,3])
y.shape,x.shape,x,y,x+y
-------------------------------
(torch.Size([3]),
 torch.Size([2, 2, 3]),
 tensor([[[ 0,  1,  2],
          [ 3,  4,  5]],
 
         [[ 6,  7,  8],
          [ 9, 10, 11]]]),
 tensor([1, 2, 3]),
 tensor([[[ 1,  3,  5],
          [ 4,  6,  8]],
 
         [[ 7,  9, 11],
          [10, 12, 14]]]))

eg4:

x=torch.arange(12).reshape(2,2,3)
y=torch.tensor([[1,2,3],[1,2,3]])
y.shape,x.shape,x,y,x+y
-------------------------------
(torch.Size([2, 3]),
 torch.Size([2, 2, 3]),
 tensor([[[ 0,  1,  2],
          [ 3,  4,  5]],
 
         [[ 6,  7,  8],
          [ 9, 10, 11]]]),
 tensor([[1, 2, 3],
         [1, 2, 3]]),
 tensor([[[ 1,  3,  5],
          [ 4,  6,  8]],
 
         [[ 7,  9, 11],
          [10, 12, 14]]]))

2.两个数组维数相同,有一方的长度为1。

eg1:

x=torch.arange(12).reshape(4,3)
y=torch.arange(4).reshape(4,1)
x.shape,y.shape,x,y,x+y
-------------------------------
(torch.Size([4, 3]),
 torch.Size([4, 1]),
 tensor([[ 0,  1,  2],
         [ 3,  4,  5],
         [ 6,  7,  8],
         [ 9, 10, 11]]),
 tensor([[0],
         [1],
         [2],
         [3]]),
 tensor([[ 0,  1,  2],
         [ 4,  5,  6],
         [ 8,  9, 10],
         [12, 13, 14]]))

eg2:

x=torch.arange(18).reshape(3,1,6)
y=torch.arange(15).reshape(3,5,1)
x.shape,y.shape,x,y,x+y
-------------------------------
(torch.Size([3, 1, 6]),
 torch.Size([3, 5, 1]),
 tensor([[[ 0,  1,  2,  3,  4,  5]],
 
         [[ 6,  7,  8,  9, 10, 11]],
 
         [[12, 13, 14, 15, 16, 17]]]),
 tensor([[[ 0],
          [ 1],
          [ 2],
          [ 3],
          [ 4]],
 
         [[ 5],
          [ 6],
          [ 7],
          [ 8],
          [ 9]],
 
         [[10],
          [11],
          [12],
          [13],
          [14]]]),
 tensor([[[ 0,  1,  2,  3,  4,  5],
          [ 1,  2,  3,  4,  5,  6],
          [ 2,  3,  4,  5,  6,  7],
          [ 3,  4,  5,  6,  7,  8],
          [ 4,  5,  6,  7,  8,  9]],
 
         [[11, 12, 13, 14, 15, 16],
          [12, 13, 14, 15, 16, 17],
          [13, 14, 15, 16, 17, 18],
          [14, 15, 16, 17, 18, 19],
          [15, 16, 17, 18, 19, 20]],
 
         [[22, 23, 24, 25, 26, 27],
          [23, 24, 25, 26, 27, 28],
          [24, 25, 26, 27, 28, 29],
          [25, 26, 27, 28, 29, 30],
          [26, 27, 28, 29, 30, 31]]]))

注意.从以上两种情况可以得出torch.Size([3, 5, 6]),
torch.Size([1, 6]),也是没问题的。

x=torch.arange(90).reshape(3,5,6)
y=torch.arange(6).reshape(1,6)
x.shape,y.shape,x,y,x+y
-------------------------------
(torch.Size([3, 5, 6]),
 torch.Size([1, 6]),
 tensor([[[ 0,  1,  2,  3,  4,  5],
          [ 6,  7,  8,  9, 10, 11],
          [12, 13, 14, 15, 16, 17],
          [18, 19, 20, 21, 22, 23],
          [24, 25, 26, 27, 28, 29]],
 
         [[30, 31, 32, 33, 34, 35],
          [36, 37, 38, 39, 40, 41],
          [42, 43, 44, 45, 46, 47],
          [48, 49, 50, 51, 52, 53],
          [54, 55, 56, 57, 58, 59]],
 
         [[60, 61, 62, 63, 64, 65],
          [66, 67, 68, 69, 70, 71],
          [72, 73, 74, 75, 76, 77],
          [78, 79, 80, 81, 82, 83],
          [84, 85, 86, 87, 88, 89]]]),
 tensor([[0, 1, 2, 3, 4, 5]]),
 tensor([[[ 0,  2,  4,  6,  8, 10],
          [ 6,  8, 10, 12, 14, 16],
          [12, 14, 16, 18, 20, 22],
          [18, 20, 22, 24, 26, 28],
          [24, 26, 28, 30, 32, 34]],
 
         [[30, 32, 34, 36, 38, 40],
          [36, 38, 40, 42, 44, 46],
          [42, 44, 46, 48, 50, 52],
          [48, 50, 52, 54, 56, 58],
          [54, 56, 58, 60, 62, 64]],
 
         [[60, 62, 64, 66, 68, 70],
          [66, 68, 70, 72, 74, 76],
          [72, 74, 76, 78, 80, 82],
          [78, 80, 82, 84, 86, 88],
          [84, 86, 88, 90, 92, 94]]]))