A deep convolutional net, with one convo/subsamp set than LeNet5.
8.1.0.7.15.0. (new net-cscscscf iniinjki0kj0tbl0si0sj0ki1kj1tbl1si1sj1ki2kj2tbl2si2sj2ki3kj3tbl3outthickprm)
[CLASS] (packages/gblearn2/net-cscscscf.lsh)
makes a new net-cscscf module. ini
inj : expected max size of input for preallocation of
internal states ki0
kj0 : kernel size for first convolutional layer
tbl0 : table of connections between input and feature maps
for first layer si0
sj0 : subsampling for first layer ki1
kj1 tbl1
si1 sj1 : same for next 2
layers... ki2
kj2 tbl2
si2 sj2 : ...and the next 2
ki3 kj3
tbl3 : same yet again for last convolution layer
outthick : number of outputs. prm
an idx1-ddparam in which the parameters will be allocated.