Summary of basic knowledge recently used in python (8)

1. The scatter() function of pytorch

 The role of scatter()  and  scatter_() is the same, except that scatter() will not directly modify the original Tensor, while scatter_() will.

There are 3 parameters for scatter(dim, index, src)

• dim: along which dimension to index
• index: the element index used for scatter
• src: The source element used for scatter, which can be a scalar or a tensor

>>> torch.zeros(3, 5).scatter_(0, torch.tensor([[0, 1, 2, 0, 0], [2, 0, 0, 1, 2]]), 2)
tensor([[2., 2., 2., 2., 2.],
        [0., 2., 0., 2., 0.],
        [2., 0., 2., 0., 2.]])

 In the above example, the first parameter '0' represents the 0th dimension, which is the row in the example. The second parameter'torch.tensor([[0, 1, 2, 0, 0], [2, 0, 0, 1, 2]])' represents the index of the first element. In short, those places need to be changed. Value operations, for example, the [0,1,2,0,0] in the first row in the first row means to change the values ​​of the elements at 0,1,2,0,0 in the same column. The third parameter '2' means that the final change value is set to 2.

Reference: https://www.cnblogs.com/dogecheng/p/11938009.html

Extend to 3d matrix

>>> torch.zeros(2, 3, 5).scatter_(1, torch.tensor([[[0, 1, 2, 0, 0], [2, 0, 0, 1, 2]],[[0, 1, 2, 0, 0], [2, 0, 0, 1, 2]]]), 2)
tensor([[[2., 2., 2., 2., 2.],
         [0., 2., 0., 2., 0.],
         [2., 0., 2., 0., 2.]],

        [[2., 2., 2., 2., 2.],
         [0., 2., 0., 2., 0.],
         [2., 0., 2., 0., 2.]]])

Compare it with the first dimension parameter to determine which dimension to transform. (Generally used when the semantic segmentation mask is converted to one-hot label)