Huixxi / TensorFlow2.0-for-Deep-Reinforcement-Learning

TensorFlow 2.0 for Deep Reinforcement Learning. :octopus:

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

NoisyDense incorrect sigma init?

pedrohpf opened this issue · comments

commented

You have this code when init sigma:

sigma_initializer = tf.constant_initializer(self.std_init / np.sqrt(self.units))

Going into the original paper (section 3.2) I would assume the init to be like this:

sigma_initializer = tf.constant_initializer(self.std_init / np.sqrt(input_dim))

Was this change intended? Or maybe I'm misunderstanding the paper?