def __init__(self,
sess,
ob_space,
ac_space,
n_env,
n_steps,
n_batch,
reuse=False,
scale=False,
obs_phs=None,
add_action_ph=False):
self.n_env = n_env
self.n_steps = n_steps
self.n_batch = n_batch
with tf.variable_scope("input", reuse=False):
if obs_phs is None:
self._obs_ph, self._processed_obs = observation_input(
ob_space, n_batch, scale=scale)
else:
self._obs_ph, self._processed_obs = obs_phs
self._action_ph = None
if add_action_ph:
self._action_ph = tf.placeholder(dtype=ac_space.dtype,
shape=(n_batch, ) +
ac_space.shape,
name="action_ph")
self.sess = sess
self.reuse = reuse
self.ob_space = ob_space
self.ac_space = ac_space
def __init__(self, observation_space, name=None):
"""Creates an input placeholder tailored to a specific observation space
Parameters
----------
observation_space:
observation space of the environment. Should be one of the gym.spaces types
name: str
tensorflow name of the underlying placeholder
"""
inpt, self.processed_inpt = observation_input(observation_space, name=name)
super().__init__(inpt)