Edit model card

(CleanRL) C51 Agent Playing BeamRiderNoFrameskip-v4

This is a trained model of a C51 agent playing BeamRiderNoFrameskip-v4. The model was trained by using CleanRL and the most up-to-date training code can be found here.

Get Started

To use this model, please install the cleanrl package with the following command:

pip install "cleanrl[c51_atari_jax]"
python -m cleanrl_utils.enjoy --exp-name c51_atari_jax --env-id BeamRiderNoFrameskip-v4

Please refer to the documentation for more detail.

Command to reproduce the training

curl -OL https://huggingface.co/kinalmehta/BeamRiderNoFrameskip-v4-c51_atari_jax-seed1/raw/main/c51_atari_jax.py
curl -OL https://huggingface.co/kinalmehta/BeamRiderNoFrameskip-v4-c51_atari_jax-seed1/raw/main/pyproject.toml
curl -OL https://huggingface.co/kinalmehta/BeamRiderNoFrameskip-v4-c51_atari_jax-seed1/raw/main/poetry.lock
poetry install --all-extras
python c51_atari_jax.py --save-model --upload-model --hf-entity kinalmehta --env-id BeamRiderNoFrameskip-v4

Hyperparameters

{'batch_size': 32,
 'buffer_size': 1000000,
 'capture_video': False,
 'end_e': 0.01,
 'env_id': 'BeamRiderNoFrameskip-v4',
 'exp_name': 'c51_atari_jax',
 'exploration_fraction': 0.1,
 'gamma': 0.99,
 'hf_entity': 'kinalmehta',
 'learning_rate': 0.00025,
 'learning_starts': 80000,
 'n_atoms': 51,
 'save_model': True,
 'seed': 1,
 'start_e': 1,
 'target_network_frequency': 10000,
 'total_timesteps': 10000000,
 'track': False,
 'train_frequency': 4,
 'upload_model': True,
 'v_max': 10,
 'v_min': -10,
 'wandb_entity': None,
 'wandb_project_name': 'cleanRL'}
Downloads last month

-

Downloads are not tracked for this model. How to track
Video Preview
loading

Evaluation results

  • mean_reward on BeamRiderNoFrameskip-v4
    self-reported
    5873.40 +/- 1897.78