Applies a 2D adaptive max pooling over an input signal composed of
several input planes.
Usage
nnf_adaptive_max_pool2d(input, output_size, return_indices = FALSE)
Arguments
- input
input tensor (minibatch, in_channels , iH , iW)
- output_size
the target output size (single integer or double-integer tuple)
- return_indices
whether to return pooling indices. Default: FALSE