def begin_episode(self, agent_indices):
"""Reset the recurrent states and stored episode.
Args:
agent_indices: 1D tensor of batch indices for agents starting an episode.
Returns:
Summary tensor.
"""
with tf.name_scope('begin_episode/'):
reset_state = utility.reinit_nested_vars(self._last_state, agent_indices)
reset_buffer = self._episodes.clear(agent_indices)
with tf.control_dependencies([reset_state, reset_buffer]):
return tf.constant('')
def begin_episode(self, agent_indices):
"""Reset the recurrent states and stored episode.
Args:
agent_indices: 1D tensor of batch indices for agents starting an episode.
Returns:
Summary tensor.
"""
with tf.name_scope('begin_episode/'):
reset_state = utility.reinit_nested_vars(self._last_state,
agent_indices)
# At the beginning of episodes, first empty the self._episodes on the rows
# specified by agent_indices, because self._episodes can store only one episode
# per environments.
reset_buffer = self._episodes.clear(agent_indices)
with tf.control_dependencies([reset_state, reset_buffer]):
return tf.constant('')
def begin_episode(self, agent_indices):
"""Reset the recurrent states and stored episode.
Args:
agent_indices: Tensor containing current batch indices.
Returns:
Summary tensor.
"""
with tf.name_scope('begin_episode/'):
if self._last_state is None:
reset_state = tf.no_op()
else:
reset_state = utility.reinit_nested_vars(
self._last_state, agent_indices)
reset_buffer = self._episodes.clear(agent_indices)
with tf.control_dependencies([reset_state, reset_buffer]):
return tf.constant('')