在caffe中理解net_surgery

Question

我正在关注 caffe 的 net_surgery.ipynb 示例，它解释了如何修改已保存 .caffemodel 的权重。但是，由于我是 python 的新手，所以我无法理解某些语法。

谁能解释一下下面给出的代码示例中以 conv_params = {pr: ... 开头的第 7 行是什么意思？（示例来自 net_surgery.ipynb - 步骤 8）。特别是什么是pr:？它是 (key,value) 类结构中的键吗？

# Load the fully convolutional network to transplant the parameters.
net_full_conv = caffe.Net('net_surgery/bvlc_caffenet_full_conv.prototxt', 
                          '../models/bvlc_reference_caffenet/bvlc_reference_caffenet.caffemodel',
                          caffe.TEST)
params_full_conv = ['fc6-conv', 'fc7-conv', 'fc8-conv']
# conv_params = {name: (weights, biases)}
conv_params = {pr: (net_full_conv.params[pr][0].data, net_full_conv.params[pr][1].data) for pr in params_full_conv}

for conv in params_full_conv:
    print '{} weights are {} dimensional and biases are {} dimensional'.format(conv, conv_params[conv][0].shape, conv_params[conv][1].shape)

Answer 1

您正在纠结的线路：

conv_params = {pr: (net_full_conv.params[pr][0].data, net_full_conv.params[pr][1].data) for pr in params_full_conv}

定义一个字典 conv_params，键为 'fc6-conv'、'fc7-conv' 和 'fc8-conv'。
使用 for 语句 (... for pr in ...) 构建字典称为“字典理解”，您可以找到有关此构造的更多信息 here.

在caffe中理解net_surgery

Understanding net_surgery in caffe

python

neural-network

deep-learning

caffe

pycaffe