torch.Tensor的4种乘法
torch.Tensor有4种常见的乘法:*, torch.mul, torch.mm, torch.matmul. 本文抛砖引玉,简单叙述一下这4种乘法的区别,具体使用还是要参照官方文档。
点乘
a与b做*乘法,原则是如果a与b的size不同,则以某种方式将a或b进行复制,使得复制后的a和b的size相同,然后再将a和b做element-wise的乘法。
下面以*标量和*一维向量为例展示上述过程。
* 标量
Tensor与标量k做*乘法的结果是Tensor的每个元素乘以k(相当于把k复制成与lhs大小相同,元素全为k的Tensor).
>>> a = torch.ones(3,4)
>>> a
tensor([[1., 1., 1., 1.],
[1., 1., 1., 1.],
[1., 1., 1., 1.]])
>>> a * 2
tensor([[2., 2., 2., 2.],
[2., 2., 2., 2.],
[2., 2., 2., 2.]])
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
* 一维向量
Tensor与行向量做*乘法的结果是每列乘以行向量对应列的值(相当于把行向量的行复制,成为与lhs维度相同的Tensor). 注意此时要求Tensor的列数与行向量的列数相等。
>>> a = torch.ones(3,4)
>>> a
tensor([[1., 1., 1., 1.],
[1., 1., 1., 1.],
[1., 1., 1., 1.]])
>>> b = torch.Tensor([1,2,3,4])
>>> b
tensor([1., 2., 3., 4.])
>>> a * b
tensor([[1., 2., 3., 4.],
[1., 2., 3., 4.],
[1., 2., 3., 4.]])
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
Tensor与列向量做*乘法的结果是每行乘以列向量对应行的值(相当于把列向量的列复制,成为与lhs维度相同的Tensor). 注意此时要求Tensor的行数与列向量的行数相等。
>>> a = torch.ones(3,4)
>>> a
tensor([[1., 1., 1., 1.],
[1., 1., 1., 1.],
[1., 1., 1., 1.]])
>>> b = torch.Tensor([1,2,3]).reshape((3,1))
>>> b
tensor([[1.],
[2.],
[3.]])
>>> a * b
tensor([[1., 1., 1., 1.],
[2., 2., 2., 2.],
[3., 3., 3., 3.]])
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
- 15
* 矩阵
经Arsmart在评论区提醒,增补一个矩阵 * 矩阵的例子,感谢Arsmart的热心评论!
如果两个二维矩阵A与B做点积A * B,则要求A与B的维度完全相同,即A的行数=B的行数,A的列数=B的列数
>>> a = torch.tensor([[1, 2], [2, 3]])
>>> a * a
tensor([[1, 4],
[4, 9]])
- 1
- 2
- 3
- 4
broadcast
点积是broadcast的。broadcast是torch的一个概念,简单理解就是在一定的规则下允许高维Tensor和低维Tensor之间的运算。broadcast的概念稍显复杂,在此不做展开,可以参考官方文档关于broadcast的介绍. 在torch.matmul里会有关于broadcast的应用的一个简单的例子。
这里举一个点积broadcast的例子。在例子中,a是二维Tensor,b是三维Tensor,但是a的维度与b的后两位相同,那么a和b仍然可以做点积,点积结果是一个和b维度一样的三维Tensor,运算规则是:若c = a * b
, 则c[i,*,*] = a * b[i, *, *]
,即沿着b的第0维做二维Tensor点积,或者可以理解为运算前将a沿着b的第0维也进行了expand操作,即a = a.expand(b.size()); a * b
。
>>> a = torch.tensor([[1, 2], [2, 3]]) >>> b = torch.tensor([[[1,2],[2,3]],[[-1,-2],[-2,-3]]]) >>> a * b tensor([[[ 1, 4], [ 4, 9]],
<span class="token punctuation">[</span><span class="token punctuation">[</span><span class="token operator">-</span><span class="token number">1</span><span class="token punctuation">,</span> <span class="token operator">-</span><span class="token number">4</span><span class="token punctuation">]</span><span class="token punctuation">,</span> <span class="token punctuation">[</span><span class="token operator">-</span><span class="token number">4</span><span class="token punctuation">,</span> <span class="token operator">-</span><span class="token number">9</span><span class="token punctuation">]</span><span class="token punctuation">]</span><span class="token punctuation">]</span><span class="token punctuation">)</span>
>>> b * a
tensor([[[ 1, 4],
[ 4, 9]],
<span class="token punctuation">[</span><span class="token punctuation">[</span><span class="token operator">-</span><span class="token number">1</span><span class="token punctuation">,</span> <span class="token operator">-</span><span class="token number">4</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token punctuation">[</span><span class="token operator">-</span><span class="token number">4</span><span class="token punctuation">,</span> <span class="token operator">-</span><span class="token number">9</span><span class="token punctuation">]</span><span class="token punctuation">]</span><span class="token punctuation">]</span><span class="token punctuation">)</span>
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
其实,上面提到的二维Tensor点积标量、二维Tensor点积行向量,都是发生在高维向量和低维向量之间的,也可以看作是broadcast.
torch.mul
官方文档关于torch.mul的介绍. 用法与*乘法相同,也是element-wise的乘法,也是支持broadcast的。
下面是几个torch.mul的例子.
乘标量
>>> a = torch.ones(3,4)
>>> a
tensor([[1., 1., 1., 1.],
[1., 1., 1., 1.],
[1., 1., 1., 1.]])
>>> a * 2
tensor([[2., 2., 2., 2.],
[2., 2., 2., 2.],
[2., 2., 2., 2.]])
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
乘行向量
>>> a = torch.ones(3,4)
>>> a
tensor([[1., 1., 1., 1.],
[1., 1., 1., 1.],
[1., 1., 1., 1.]])
>>> b = torch.Tensor([1,2,3,4])
>>> b
tensor([1., 2., 3., 4.])
>>> torch.mul(a, b)
tensor([[1., 2., 3., 4.],
[1., 2., 3., 4.],
[1., 2., 3., 4.]])
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
乘列向量
>>> a = torch.ones(3,4)
>>> a
tensor([[1., 1., 1., 1.],
[1., 1., 1., 1.],
[1., 1., 1., 1.]])
>>> b = torch.Tensor([1,2,3]).reshape((3,1))
>>> b
tensor([[1.],
[2.],
[3.]])
>>> torch.mul(a, b)
tensor([[1., 1., 1., 1.],
[2., 2., 2., 2.],
[3., 3., 3., 3.]])
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
乘矩阵
例1:二维矩阵 mul 二维矩阵
>>> a = torch.tensor([[1, 2], [2, 3]])
>>> torch.mul(a,a)
tensor([[1, 4],
[4, 9]])
- 1
- 2
- 3
- 4
例2:二维矩阵 mul 三维矩阵(broadcast)
>>> a = torch.tensor([[1, 2], [2, 3]]) >>> b = torch.tensor([[[1,2],[2,3]],[[-1,-2],[-2,-3]]]) >>> torch.mul(a,b) tensor([[[ 1, 4], [ 4, 9]],
<span class="token punctuation">[</span><span class="token punctuation">[</span><span class="token operator">-</span><span class="token number">1</span><span class="token punctuation">,</span> <span class="token operator">-</span><span class="token number">4</span><span class="token punctuation">]</span><span class="token punctuation">,</span> <span class="token punctuation">[</span><span class="token operator">-</span><span class="token number">4</span><span class="token punctuation">,</span> <span class="token operator">-</span><span class="token number">9</span><span class="token punctuation">]</span><span class="token punctuation">]</span><span class="token punctuation">]</span><span class="token punctuation">)</span>
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
torch.mm
官方文档关于torch.mm的介绍. 数学里的矩阵乘法,要求两个Tensor的维度满足矩阵乘法的要求.
例子:
>>> a = torch.ones(3,4)
>>> b = torch.ones(4,2)
>>> torch.mm(a, b)
tensor([[4., 4.],
[4., 4.],
[4., 4.]])
- 1
- 2
- 3
- 4
- 5
- 6
torch.matmul
官方文档关于torch.matmul的介绍. torch.mm的broadcast版本.
例子:
>>> a = torch.ones(3,4) >>> b = torch.ones(5,4,2) >>> torch.matmul(a, b) tensor([[[4., 4.], [4., 4.], [4., 4.]],
<span class="token punctuation">[</span><span class="token punctuation">[</span><span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">,</span> <span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">]</span><span class="token punctuation">,</span> <span class="token punctuation">[</span><span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">,</span> <span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">]</span><span class="token punctuation">,</span> <span class="token punctuation">[</span><span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">,</span> <span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">]</span><span class="token punctuation">]</span><span class="token punctuation">,</span> <span class="token punctuation">[</span><span class="token punctuation">[</span><span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">,</span> <span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">]</span><span class="token punctuation">,</span> <span class="token punctuation">[</span><span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">,</span> <span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">]</span><span class="token punctuation">,</span> <span class="token punctuation">[</span><span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">,</span> <span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">]</span><span class="token punctuation">]</span><span class="token punctuation">,</span> <span class="token punctuation">[</span><span class="token punctuation">[</span><span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">,</span> <span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">]</span><span class="token punctuation">,</span> <span class="token punctuation">[</span><span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">,</span> <span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">]</span><span class="token punctuation">,</span> <span class="token punctuation">[</span><span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">,</span> <span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">]</span><span class="token punctuation">]</span><span class="token punctuation">,</span> <span class="token punctuation">[</span><span class="token punctuation">[</span><span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">,</span> <span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">]</span><span class="token punctuation">,</span> <span class="token punctuation">[</span><span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">,</span> <span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">]</span><span class="token punctuation">,</span> <span class="token punctuation">[</span><span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">,</span> <span class="token number">4</span><span class="token punctuation">.</span><span class="token punctuation">]</span><span class="token punctuation">]</span><span class="token punctuation">]</span><span class="token punctuation">)</span>
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
- 15
- 16
- 17
- 18
- 19
- 20
- 21
- 22
同样的a和b,使用torch.mm相乘会报错
>>> torch.mm(a, b)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
RuntimeError: matrices expected, got 2D, 3D tensors at /pytorch/aten/src/TH/generic/THTensorMath.cpp:2065