caffe学习笔记2：net forward与backward

caffe学习笔记2：Forward and Backward

原网页:http://caffe.berkeleyvision.org/tutorial/forward_backward.html

forward 和backward是网络必要的功能。

net具有forward和backward两个重要功能，是由layer的backward和forward叠加而成的

• The Net::Forward() and Net::Backward() methods carry out the respective passes while Layer::Forward() and Layer::Backward() compute each step.

• layer forward backward ：Every layer type has forward_{cpu,gpu}() and backward_{cpu,gpu}() methods to compute its steps according to the mode of computation. A layer may only implement CPU or GPU mode due to constraints or convenience.

• Solver : 1.first calling forward to yield the output and loss
  2. then calling backward to generate the gradient of the model
  3. and then incorporating the gradient into a weight update that attempts to minimize the loss.
  4. Division of labor between the Solver, Net, and Layer keep Caffe modular and open to development

Solver net layer之间的关系保证了caffe的模块性和开放性，
更多细节请看caffe 层的分类