(CleanRL) PPO Agent Playing tactile_mnist:TactileMNIST-v0

This is a trained model of a PPO agent playing tactile_mnist:TactileMNIST-v0. 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[ppo_tactile_mnist_new_baseline_clip_action_softplus]"
python -m cleanrl_utils.enjoy --exp-name ppo_tactile_mnist_new_baseline_clip_action_softplus --env-id tactile_mnist:TactileMNIST-v0

Please refer to the documentation for more detail.

Command to reproduce the training

curl -OL https://huggingface.co/Yunzhe/TactileMNIST-v0-ppo_tactile_mnist_new_baseline_clip_action_softplus-seed1/raw/main/ppo_tactile_mnist_new_baseline_clip_action_softplus.py
curl -OL https://huggingface.co/Yunzhe/TactileMNIST-v0-ppo_tactile_mnist_new_baseline_clip_action_softplus-seed1/raw/main/pyproject.toml
curl -OL https://huggingface.co/Yunzhe/TactileMNIST-v0-ppo_tactile_mnist_new_baseline_clip_action_softplus-seed1/raw/main/poetry.lock
poetry install --all-extras
python ppo_tactile_mnist_new_baseline_clip_action_softplus.py --capture_video --save_model --upload_model --comment clip_action_softplus --num_envs 128 --total_timesteps 10000000

Hyperparameters

{'anneal_lr': True,
 'batch_size': 2048,
 'capture_video': True,
 'clip_coef': 0.2,
 'clip_vloss': True,
 'comment': 'clip_action_softplus',
 'cuda': True,
 'ent_coef': 0.01,
 'env_id': 'tactile_mnist:TactileMNIST-v0',
 'exp_name': 'ppo_tactile_mnist_new_baseline_clip_action_softplus',
 'gae_lambda': 0.95,
 'gamma': 0.99,
 'hf_entity': 'Yunzhe',
 'learning_rate': 0.0003,
 'max_grad_norm': 0.5,
 'minibatch_size': 512,
 'norm_adv': True,
 'num_envs': 128,
 'num_iterations': 4882,
 'num_minibatches': 4,
 'num_steps': 16,
 'save_model': True,
 'seed': 1,
 'target_kl': None,
 'torch_deterministic': True,
 'total_timesteps': 10000000,
 'track': True,
 'update_epochs': 4,
 'upload_model': True,
 'vf_coef': 0.5,
 'wandb_entity': 'huyz7830',
 'wandb_project_name': 'TactileMNIST-v0'}
Downloads last month

-

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

Evaluation results