Added CrossQ

CHANGELOG.md

@@ -4,6 +4,7 @@
 - Updated defaults hyperparameters for TQC/SAC for Swimmer-v4 (decrease gamma for more consistent results) (@JacobHA) [W&B report](https://wandb.ai/openrlbenchmark/sbx/reports/SAC-MuJoCo-Swimmer-v4--Vmlldzo3NzM5OTk2)
 
 ### New Features
+- Added `CrossQ` hyperparameters for SB3-contrib
 
 ### Bug fixes
 

hyperparams/crossq.yml

@@ -0,0 +1,25 @@
+# === Mujoco Envs ===
+
+HalfCheetah-v4: &mujoco-defaults
+  buffer_size: 1_000_000
+  learning_rate: !!float 1e-3
+  learning_starts: 5000
+  n_timesteps: !!float 5e6
+  policy: 'MlpPolicy'
+  policy_delay: 3
+  policy_kwargs: "dict(net_arch=dict(pi=[256, 256], qf=[2048, 2048]))"
+
+Ant-v4:
+  <<: *mujoco-defaults
+
+Hopper-v4:
+  <<: *mujoco-defaults
+
+Walker2d-v4:
+  <<: *mujoco-defaults
+
+Humanoid-v4:
+  <<: *mujoco-defaults
+
+HumanoidStandup-v4:
+  <<: *mujoco-defaults

rl_zoo3/utils.py

@@ -12,7 +12,7 @@
 from gymnasium import spaces
 from huggingface_hub import HfApi
 from huggingface_sb3 import EnvironmentName, ModelName
-from sb3_contrib import ARS, QRDQN, TQC, TRPO, RecurrentPPO
+from sb3_contrib import ARS, QRDQN, TQC, TRPO, CrossQ, RecurrentPPO
 from stable_baselines3 import A2C, DDPG, DQN, PPO, SAC, TD3
 from stable_baselines3.common.base_class import BaseAlgorithm
 from stable_baselines3.common.callbacks import BaseCallback
@@ -32,6 +32,7 @@
     "td3": TD3,
     # SB3 Contrib,
     "ars": ARS,
+    "crossq": CrossQ,
     "qrdqn": QRDQN,
     "tqc": TQC,
     "trpo": TRPO,