[ci skip] Tune RecurrentPPO Swimmer-v3 (#360)

DLR-RM · Feb 21, 2023 · 309ad8c · 309ad8c
1 parent fb9383a
commit 309ad8c
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 1 deletion.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -4,6 +4,7 @@
 - Upgraded to SB3 >= 1.8.0
 
 ### New Features
+- Tuned hyperparameters for RecurrentPPO on Swimmer
 
 ### Bug fixes
 - Set ``highway-env`` version to 1.5 and ``setuptools to`` v65.5 for the CI

diff --git a/hyperparams/ppo_lstm.yml b/hyperparams/ppo_lstm.yml
@@ -339,7 +339,11 @@ Ant-v3: &mujoco-defaults
 Swimmer-v3:
   <<: *mujoco-defaults
   gamma: 0.9999
-
+  n_envs: 4
+  n_steps: 1024
+  batch_size: 256
+  learning_rate: !!float 6e-4
+  gae_lambda: 0.98
 
 # 10 mujoco envs