Caffe PythonLayer

编译

重新编译之前，注释掉Makefile.config 中的下面这句话：

WITH_PYTHON_LAYER:=1

重新进行编译：

make clean
make -j
make pycaffe

设置路径

将生成的pycaffe添加到你的路径$PYTHONPATH中，以便能够import caffe

export PYTHONPATH=$CAFFE_ROOT/python:$PYTHONPATH

其中$CAFFE_ROOT 为你caffe的安装根目录，根据自己的实际情况进行修改。
当然你也可以在你的python layer 中导入caffe时使用相对导入，这样也可以避免你在安装了多个caffe时，导入caffe包混乱的问题

写好你自己定义的Python Layer 之后，你可以将该*.py文件放在$PYTHONPATH 中；也可以放在任意路径（例如：$Layer ）中，随后再将该路径添加到$PYTHONPATH 中即可。

编写你的Python Layer

参考Caffe官方样例：$CAFFE_ROOT/examples/pycaffe/layers/pyloss.py

import caffe
import numpy as np


class EuclideanLossLayer(caffe.Layer):
    """
    Compute the Euclidean Loss in the same manner as the C++ EuclideanLossLayer
    to demonstrate the class interface for developing layers in Python.
    """

    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

使用你的Python Layer

参考$CAFFE_ROOT/examples/pycaffe/linreg.prototxt

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
    # 文件名
    module: 'pyloss'
    # the layer name -- the class name in the module
    # 该模块中的类名
    layer: 'EuclideanLossLayer'
  }
  # set loss weight so Caffe knows this is a loss layer.
  # since PythonLayer inherits directly from Layer, this isn't automatically
  # known to Caffe
  loss_weight: 1
}

杂七杂八

python_param还可以有param_str，用来传递该层参数，是字符串。
若以命令行方式而不是python接口使用Python Layer 时，往往语句print 没有效果。可以使用，如下方式：

import logging
logging.basicConfig(level=logging.INFO)

logging.info('Info you want to print')