Neto is a module that provide helper functions to store Artificial Neural Networks (ANN) and compiled function objects built with lasagne and theano.
store_compiled_fn(fns, filename, old_dir='old_file') -> None
store the functions object compiled with theano.function.
fns : list or function object
filename : str
the filename(or path) to store the functions, since it uses pickle to store the objects, suggested file extension is '.pic'.
old_dir : str
if the filename exist, move the existing file into this directory, in order to preserve previously compiled functions
load_compiled_fn(filename) -> tuple
return the tuple that contains the functions stored in the file
fileanme : str
store_network(net, filename, old_dir='old_file') -> None
store the parameters of network into the file, note that next time you still have to build the network with lasagne, than you can use load_network function to load the parameter into the network
net : lasagne layer object
filename : str
the filename(or path) to store the parameters.
old_dir : str
if the filename exist, move the existing file into this directory, in order to preserve previously compiled functions
load_network(net, filename) -> None
load the network with the parameters stored in file
net : lasagne layer object
filename : str
the filename(or path) to store the parameters.