HW5c fix (nit): make build_rnn more clear, remove unneeded arg

Author: katerakelly

hw5/meta/train_policy.py

@@ -57,11 +57,14 @@ def build_mlp(x, output_size, scope, n_layers, size, activation=tf.tanh, output_
     x = tf.layers.dense(inputs=x, units=output_size, activation=output_activation, name='fc{}'.format(i + 1), kernel_regularizer=regularizer, bias_regularizer=regularizer)
     return x
 
-def build_rnn(x, h, output_size, scope, n_layers, size, gru_size, activation=tf.tanh, output_activation=None, regularizer=None):
+def build_rnn(x, h, output_size, scope, n_layers, size, activation=tf.tanh, output_activation=None, regularizer=None):
     """
     builds a gated recurrent neural network
     inputs are first embedded by an MLP then passed to a GRU cell
 
+    make MLP layers with `size` number of units
+    make the GRU with `output_size` number of units
+
     arguments:
         (see `build_policy()`)
 
@@ -96,7 +99,7 @@ def build_policy(x, h, output_size, scope, n_layers, size, gru_size, recurrent=T
     """
     with tf.variable_scope(scope, reuse=tf.AUTO_REUSE):
         if recurrent:
-            x, h = build_rnn(x, h, gru_size, scope, n_layers, size, gru_size, activation=activation, output_activation=output_activation)
+            x, h = build_rnn(x, h, gru_size, scope, n_layers, size, activation=activation, output_activation=output_activation)
         else:
             x = tf.reshape(x, (-1, x.get_shape()[1]*x.get_shape()[2]))
             x = build_mlp(x, gru_size, scope, n_layers + 1, size, activation=activation, output_activation=activation)
@@ -115,7 +118,7 @@ def build_critic(x, h, output_size, scope, n_layers, size, gru_size, recurrent=T
     """
     with tf.variable_scope(scope, reuse=tf.AUTO_REUSE):
         if recurrent:
-            x, h = build_rnn(x, h, gru_size, scope, n_layers, size, gru_size, activation=activation, output_activation=output_activation, regularizer=regularizer)
+            x, h = build_rnn(x, h, gru_size, scope, n_layers, size, activation=activation, output_activation=output_activation, regularizer=regularizer)
         else:
             x = tf.reshape(x, (-1, x.get_shape()[1]*x.get_shape()[2]))
             x = build_mlp(x, gru_size, scope, n_layers + 1, size, activation=activation, output_activation=activation, regularizer=regularizer)