hiro/ray
1
0
Fork 0
mirror of https://github.com/vale981/ray synced 2025-03-06 10:31:39 -05:00
Code Issues Projects Releases Packages Wiki Activity Actions

Checkpoint Adam momenta for DDPG (#7449)

Browse source
This commit is contained in:
Eric Liang 2020-03-04 10:03:41 -08:00 • committed by GitHub
parent fe7820fec9
commit aa4861c2a0
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 2 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

3
rllib/agents/ddpg/ddpg_policy.py
View file

@ -295,7 +295,8 @@ class DDPGTFPolicy(DDPGPostprocessing, TFPolicy):
# Note that this encompasses both the policy and Q-value networks and
# their corresponding target networks
self.variables = ray.experimental.tf_utils.TensorFlowVariables(
tf.group(q_t_det_policy, q_tp1), self.sess)
tf.group(q_t_det_policy, q_tp1, self._actor_optimizer.variables(),
self._critic_optimizer.variables()), self.sess)
# Hard initial update
self.update_target(tau=1.0)