conv neural network - Which part of the deploy.prototxt file in caffe is absolutely necessary for testing? -
in recent discussion, found out parts of deploy.prototxt exist because have been directly copied train_test.prototxt , ignored during testing. example:
layer { name: "conv1" type: "convolution" bottom: "data" top: "conv1" param { #starting here lr_mult: 1 } param { lr_mult: 2 } #to here convolution_param { #is section useful? num_output: 20 kernel_size: 5 stride: 1 weight_filler { type: "xavier" } bias_filler { type: "constant" } } }
i told section containing lr weights biases useless in deploy files , deleted. got me thinking, convolution_param portion absolutely required? if yes, still have define weight , bias fillers testing using file , fillers initialized when need train network. there other detail unnecessary?
the convolution_param
portion required can remove weight_filler
, bias_filler
if want.
Comments
Post a Comment