ernestum's picture
Initial Commit
9e5b409
metadata
library_name: stable-baselines3
tags:
  - seals/Walker2d-v0
  - deep-reinforcement-learning
  - reinforcement-learning
  - stable-baselines3
model-index:
  - name: PPO
    results:
      - metrics:
          - type: mean_reward
            value: 1429.13 +/- 411.75
            name: mean_reward
        task:
          type: reinforcement-learning
          name: reinforcement-learning
        dataset:
          name: seals/Walker2d-v0
          type: seals/Walker2d-v0

PPO Agent playing seals/Walker2d-v0

This is a trained model of a PPO agent playing seals/Walker2d-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 utils.load_from_hub --algo ppo --env seals/Walker2d-v0 -orga ernestumorga -f logs/
python enjoy.py --algo ppo --env seals/Walker2d-v0  -f logs/

Training (with the RL Zoo)

python train.py --algo ppo --env seals/Walker2d-v0 -f logs/
# Upload the model and generate video (when possible)
python -m utils.push_to_hub --algo ppo --env seals/Walker2d-v0 -f logs/ -orga ernestumorga

Hyperparameters

OrderedDict([('batch_size', 8),
             ('clip_range', 0.4),
             ('ent_coef', 0.00013057334805552262),
             ('gae_lambda', 0.92),
             ('gamma', 0.98),
             ('learning_rate', 3.791707778339674e-05),
             ('max_grad_norm', 0.6),
             ('n_envs', 1),
             ('n_epochs', 5),
             ('n_steps', 2048),
             ('n_timesteps', 1000000.0),
             ('normalize', True),
             ('policy', 'MlpPolicy'),
             ('policy_kwargs',
              'dict(activation_fn=nn.ReLU, net_arch=[dict(pi=[256, 256], '
              'vf=[256, 256])])'),
             ('vf_coef', 0.6167177795726859),
             ('normalize_kwargs', {'norm_obs': True, 'norm_reward': False})])