wassname/ray
1
0
Fork 0
mirror of https://github.com/wassname/ray.git synced 2026-07-13 17:45:08 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix SAC bug (twin Q not used for min'ing over both Q-nets in loss func). (#7354)

Browse Source
This commit is contained in:
Sven Mika
2020-02-27 21:49:08 +01:00
committed by GitHub
parent ced062319d
commit 0c9e5db9cb
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
rllib/agents/sac/sac_policy.py
+1 -1
View File
@@ -160,7 +160,7 @@ def actor_critic_loss(policy, model, _, train_batch):
# Q-values for current policy (no noise) in given current state
q_t_det_policy = model.get_q_values(model_out_t, policy_t)
if policy.config["twin_q"]:
twin_q_t_det_policy = model.get_q_values(model_out_t, policy_t)
twin_q_t_det_policy = model.get_twin_q_values(model_out_t, policy_t)
q_t_det_policy = tf.reduce_min(
(q_t_det_policy, twin_q_t_det_policy), axis=0)