Caffe Python Layer

Python layer in Caffe can speed up development process Issue1703

Compile WITH_PYTHON_LAYER option

First, you have to build Caffe with WITH_PYTHON_LAYER option 1. Run make clean to delete all the compiled binaries. Then,

WITH_PYTHON_LAYER=1 make && make pycaffe

If you skip this, caffe will complain that layer factory function can’t find Python layer.

layer_factory.hpp:77] Check failed: registry.count(type) == 1 (0 vs. 1) Unknown layer type: Python

Python Layer

A gist from Evan Shelhamer summarizes the basics of the python layer.

...
layer {
  type: 'Python'
  name: 'loss'
  top: 'loss'
  bottom: 'ipx'
  bottom: 'ipy'
  python_param {
    # the module name -- usually the filename -- that needs to be in $PYTHONPATH
    module: 'pyloss'
    # the layer name -- the class name in the module
    layer: 'EuclideanLossLayer'
  }
  # set loss weight so Caffe knows this is a loss layer
  loss_weight: 1
}

You have to define a python layer that is defined in your $PYTHONPATH. In the prototxt, the module is pyloss, which means that the file that contains the EuclideanLossLayer should be named pyloss.py.

import caffe
import numpy as np

class EuclideanLossLayer(caffe.Layer):

    def setup(self, bottom, top):
        # check input pair
        if len(bottom) != 2:
            raise Exception("Need two inputs to compute distance.")

    def reshape(self, bottom, top):
        # check input dimensions match
        if bottom[0].count != bottom[1].count:
            raise Exception("Inputs must have the same dimension.")
        # difference is shape of inputs
        self.diff = np.zeros_like(bottom[0].data, dtype=np.float32)
        # loss output is scalar
        top[0].reshape(1)

    def forward(self, bottom, top):
        self.diff[...] = bottom[0].data - bottom[1].data
        top[0].data[...] = np.sum(self.diff**2) / bottom[0].num / 2.

    def backward(self, top, propagate_down, bottom):
        for i in range(2):
            if not propagate_down[i]:
                continue
            if i == 0:
                sign = 1
            else:
                sign = -1
            bottom[i].diff[...] = sign * self.diff / bottom[i].num
  1. https://github.com/BVLC/caffe/issues/2093 

Categories:

Updated:

Leave a Comment