CODEWITHSUNDEEP
pythondeep-learningcaffe
Mak
Mak

Reputation: 3

Caffe python layer backword pass implementation

I am writing a caffe python layer that resale the input between [0 255] along specific axis (code attached) and the forward pass is working fine. Is the backward pass required for such layer? if so, how can i implement it?

caffe_root = 'caffe_root'           
import sys
sys.path.insert(0, caffe_root + 'python')
import caffe
import numpy as np

class scale_layer(caffe.Layer):

  def setup(self, bottom, top):
    assert len(bottom)==1 and len(top)==1, "scale_layer expects a single input and a single output"

  def reshape(self, bottom, top):
    top[0].reshape(*bottom[0].data.shape)

  def forward(self, bottom, top):
    in_ = np.array(bottom[0].data)
    x_min = in_.min(axis=(0, 1), keepdims=True) 
    x_max = in_.max(axis=(0, 1), keepdims=True)
    top[0].data[...] = np.around(255*((in_-x_min)/(x_max-x_min)))

  def backward(self, top, propagate_down, bottom):
    # backward pass is not implemented!
    ???????????????????????????
    pass

Upvotes: 0

Views: 67

Answers (1)

Shai
Shai

Reputation: 114816

Your function is quite simple, if you are willing to ignore the np.around:

enter image description here

For x=x_min and for x=x_max the derivative is zero, for all other x the derivative is 255/(x_max-x_min).

This can be implemented by

def forward(self, bottom, top):
  in_ = bottom[0].data
  self.x_min = in_.min(axis=(0, 1), keepdims=True)  # cache min/max for backward
  self.x_max = in_.max(axis=(0, 1), keepdims=True)
  top[0].data[...] = 255*((in_-self.x_min)/(self.x_max-self.x_min)))

def backward(self, top, propagate_down, bottom):
  in_ = bottom[0].data
  b, c = in_.shape[:2]
  diff = np.tile( 255/(self.x_max-self.x_min), (b, c, 1, 1) )
  diff[ in_ == self.x_min ] = 0
  diff[ in_ == self.x_max ] = 0
  bottom[0].diff[...] = diff * top[0].diff

Do not forget to test this numberically. This can be done, e.g., using test_gradient_for_python_layer.

Upvotes: 1

Related Questions