neural network - How does Caffe determine the number of neurons in each layer? -
recently, i've been trying use caffe of deep learning work i'm doing. although writing model in caffe easy, i've not been able know answer question. how caffe determine number of neurons in hidden layer? know determination of number of neurons in layer , number of hidden layers problems cannot determined analytically , use of 'thumb rules' imperative in regard. there way define or know number of neurons in each layer in caffe? , default, how caffe inherently determine this?
any appreciated!
caffe doesn't determine number of neurons--the user does.
pulled straight caffe's website, here: http://caffe.berkeleyvision.org/tutorial/layers.html
for example, convolution layer of 96 nodes (or neurons):
layer { name: "conv1" type: "convolution" bottom: "data" top: "conv1" # learning rate , decay multipliers filters param { lr_mult: 1 decay_mult: 1 } # learning rate , decay multipliers biases param { lr_mult: 2 decay_mult: 0 } convolution_param { num_output: 96 # learn 96 filters kernel_size: 11 # each filter 11x11 stride: 4 # step 4 pixels between each filter application weight_filler { type: "gaussian" # initialize filters gaussian std: 0.01 # distribution stdev 0.01 (default mean: 0) } bias_filler { type: "constant" # initialize biases 0 (0) value: 0 } } }
Comments
Post a Comment