Full Deep Reinforcement Example

All that's left to do is bring together everything from the previous sections!

from hula.DeepR import Net, FeedForwardLayer
from hula.rlutils import sigmoid, softplus, tanh

ExampleNet = Net(
  FeedForwardLayer(2, 16, tanh),
  FeedForwardLayer(16, 6, softplus),
  FeedForwardLayer(6, 1, sigmoid)
)

for epoch in range(1000):
  
  ExampleNet.randomAct(0.001)
  error = ExampleNet.activate([1, 1])[0]
  ExampleNet.score(error)
  
  ExampleNet.train(0.8)
  
  if epoch % 50 == 0:
    ExampleNet.simplify(2, 0.5)
  
  print(epoch, error)

The Network will be trained to approach 1.0 for 1000 epochs.