araffin's picture
Upload with huggingface_hub
library_name: stable-baselines3
  - donkey-avc-sparkfun-v0
  - deep-reinforcement-learning
  - reinforcement-learning
  - stable-baselines3
  - name: TQC
      - metrics:
          - type: mean_reward
            value: 552.57 +/- 285.35
            name: mean_reward
          type: reinforcement-learning
          name: reinforcement-learning
          name: donkey-avc-sparkfun-v0
          type: donkey-avc-sparkfun-v0

TQC Agent playing donkey-avc-sparkfun-v0

This is a trained model of a TQC agent playing donkey-avc-sparkfun-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:
SB3 Contrib:

Autoencoder: branch: feat/race_june
Gym env: branch: feat/race_june
RL Zoo branch: feat/gym-donkeycar

Pretrained autoencoder can be downloaded here:

# Export path to autoencoder
export AE_PATH=/path/to/ae-32_avc.pkl
# Download model and save it into the logs/ folder
python -m rl_zoo3.load_from_hub --algo tqc --env donkey-avc-sparkfun-v0 -orga araffin -f logs/
python --algo tqc --env donkey-avc-sparkfun-v0  -f logs/

Training (with the RL Zoo)

python --algo tqc --env donkey-avc-sparkfun-v0 -f logs/
# Upload the model and generate video (when possible)
python -m rl_zoo3.push_to_hub --algo tqc --env donkey-avc-sparkfun-v0 -f logs/ -orga araffin


OrderedDict([('batch_size', 256),
             ('buffer_size', 200000),
              [{'rl_zoo3.callbacks.ParallelTrainCallback': {'gradient_steps': 200}},
             ('ent_coef', 'auto'),
               {'rl_zoo3.wrappers.HistoryWrapper': {'horizon': 2}}]),
             ('gamma', 0.99),
             ('gradient_steps', 256),
             ('learning_rate', 0.00073),
             ('learning_starts', 500),
             ('n_timesteps', 2000000.0),
             ('normalize', "{'norm_obs': True, 'norm_reward': False}"),
             ('policy', 'MlpPolicy'),
              'dict(log_std_init=-3, net_arch=[256, 256], n_critics=2, '
             ('sde_sample_freq', 16),
             ('tau', 0.02),
             ('train_freq', 200),
             ('use_sde', True),
             ('use_sde_at_warmup', True),
             ('normalize_kwargs', {'norm_obs': True, 'norm_reward': False})])

Environment Arguments

{'conf': {'cam_resolution': (120, 160, 3),
          'car_config': {'body_rgb': (226, 112, 18),
                         'body_style': 'donkey',
                         'car_name': 'Toni',
                         'font_size': 40},
          'frame_skip': 1,
          'host': 'localhost',
          'level': 'sparkfun_avc',
          'log_level': 20,
          'max_cte': 16,
          'port': 9091,
          'start_delay': 5.0},
 'min_throttle': -0.2,
 'steer': 0.3}