tf-->caffe input shape convert error

[zouxinhua]

Platform (like ubuntu 16.04/win10): ubuntu 16.04

Python version: 2.7

Source framework with version (like Tensorflow 1.4.1 with GPU): tensorflow 1.8

Destination framework with version (like CNTK 2.3 with GPU): caffe 1.0

in my case, my tensorflow input type is NCWH,
input = tf.placeholder(tf.float32,shape=[-1,4,32,32],name='data_image')

but after convert to caffe model, input shape is [1,32,4,32] ,I think this maybe not right.

in caffe_emitter.emit_DataInput(line 336),there is a shape convert like this:

shape = [shape[0], shape[-1]] + shape[1:-1]

I think this means convert NWHC to NCWH,but my shape is NCWH, this convert maybe unnecessary.

please check it. thx