sb3/sac-HopperBulletEnv-v0

SAC Agent playing HopperBulletEnv-v0

This is a trained model of a SAC agent playing HopperBulletEnv-v0 using the stable-baselines3 library and the RL Zoo.

The RL Zoo is a training framework for Stable Baselines3 reinforcement learning agents, with hyperparameter optimization and pre-trained agents included.

Usage (with SB3 RL Zoo)

RL Zoo: https://github.com/DLR-RM/rl-baselines3-zoo
SB3: https://github.com/DLR-RM/stable-baselines3
SB3 Contrib: https://github.com/Stable-Baselines-Team/stable-baselines3-contrib

# Download model and save it into the logs/ folder
python -m rl_zoo3.load_from_hub --algo sac --env HopperBulletEnv-v0 -orga sb3 -f logs/
python enjoy.py --algo sac --env HopperBulletEnv-v0  -f logs/

Training (with the RL Zoo)

python train.py --algo sac --env HopperBulletEnv-v0 -f logs/
# Upload the model and generate video (when possible)
python -m rl_zoo3.push_to_hub --algo sac --env HopperBulletEnv-v0 -f logs/ -orga sb3

Hyperparameters

OrderedDict([('batch_size', 256),
             ('buffer_size', 300000),
             ('ent_coef', 'auto'),
             ('env_wrapper', 'sb3_contrib.common.wrappers.TimeFeatureWrapper'),
             ('gamma', 0.98),
             ('gradient_steps', 64),
             ('learning_rate', 'lin_7.3e-4'),
             ('learning_starts', 10000),
             ('n_timesteps', 1000000.0),
             ('policy', 'MlpPolicy'),
             ('policy_kwargs', 'dict(log_std_init=-3, net_arch=[400, 300])'),
             ('tau', 0.02),
             ('train_freq', 64),
             ('use_sde', True),
             ('normalize', False)])