def make_policy(name, ob_space, ac_space):
return MlpPolicy(name=name,
ob_space=ob_space,
ac_space=ac_space,
hid_size=hid_size,
num_hid_layers=num_hid_layers,
gaussian_fixed_var=True,
use_bias=use_bias,
use_critic=False,
hidden_W_init=policy_initializer,
output_W_init=policy_initializer)
def make_policy(name, ob_space, ac_space):
return MlpPolicy(name=name,
ob_space=ob_space,
ac_space=ac_space,
hid_size=hid_size,
num_hid_layers=num_hid_layers,
gaussian_fixed_var=True,
use_bias=False,
use_critic=False,
hidden_W_init=tf.contrib.layers.xavier_initializer(),
output_W_init=tf.contrib.layers.xavier_initializer())