Author(s): Yann LeCun

neural net with one convolutional layer, one (replicable) fully-connected layer, and one RBF (euclidiean distance) layer. (new net-cfe n e thick si sj)
[CLASS] (packages/gblearn2/net-cfe.lsh)

create a new net-cfe. n is a net-cf, e is a e-layer, thick is the number of outputs, and si , sj are the number of output replications.