How to solve the torch.masked_select problem

In this issue, the editor will bring you about how to solve the torch.masked_select problem. The article is rich in content and analyzes and narrates it from a professional point of view. I hope you can get something after reading this article.

Brief introduction:

When studying the official documents of pytorch, I found that the program of the mask was posted incorrectly. I wrote one myself, which we can refer to.

Torch.masked_select (input, mask, out=None) → Tensor

According to the binary value in the mask tensor mask, take the specified term in the input tensor (mask is a ByteTensor), and return the value to a new 1D tensor.

The tensor mask must have the same number of elements as the input tensor, but the shape or dimension does not need to be the same.

Note: the returned tensor does not share memory space with the original tensor.

Parameters:

Input (Tensor)-input tensor

Mask (ByteTensor)-Mask tensor, which contains binary index values

Out (Tensor, optional)-objective tensor

Experimental phenomenon

X = torch.randn (3pr 4)

Mask = torch.ByteTensor (x > 0)

Torch.masked_select (XMagneMask)

The above is the editor for you to share how to solve the torch.masked_select problem, if you happen to have similar doubts, you might as well refer to the above analysis to understand. If you want to know more about it, you are welcome to follow the industry information channel.