Closed dpblnt closed 1 year ago
Without initializing the weights the net produces results. After initializing it like this:
@net.init_weights( RubyFann::TrainData.new( inputs: [Array.new( @input_size){0.1}], desired_outputs: [Array.new( @input_size){0.1}] ))
the result fom @net.run is [NaN, NaN, NaN] Tried different values for initialization. Any ideas?
[NaN, NaN, NaN]
if i call init_weights with my first set of real data then it's fine
Without initializing the weights the net produces results. After initializing it like this:
the result fom @net.run is
[NaN, NaN, NaN]
Tried different values for initialization. Any ideas?