Output variable creation in depthwise convolution

Where should I change to create a new argument similar to output which returns a tensor in depth wise convolution?


I’m not sure to understand your question.
You want to create a Tensor that is the same size as the output of a given convolution?
Could you elaborate on why you want to be able to do that?

Thanks for replying ,yeah that’s right I would like to create a tensor of same size as output .

There is no existing method to do this. But you can use the formula from the convolution doc where the output shape is specified as a function of the parameters. Then create a new Tensor of that size.

I would like to do that in C source code in SpatialDepthwiseConv.cpp in Aten
How can I do that.


Why would you want that in the cpp code? It will be much simpler to do it in a python function in your code no?