python - Slice Layer in Caffe prototxt -
the last layer right in neural network convolutional layer, output of 4 dimensional feature map of size [n, m, p , q].
i want know how can split 3 feature map shown in example below:
[n, :m+5, :m+5, :m+5] [n, m+5:m+20, m+5:m+20, m+5:m+20] [n, m+20:, m+20:, m+20:]
how can using slice layer in prototxt file? if not prototxt, how can in code? need slice before use softmax function.
please help!
Comments
Post a Comment