IsaacGymEnvs-Ant-PPO

<!-- --- torch: 4996.72 +/- 273.16 jax: 5094.76 +/- 310.06 numpy: 4542.73 +/- 467.69 --- -->

IsaacGymEnvs-Ant-PPO

Trained agent for NVIDIA Isaac Gym Preview environments.

• Task: Ant
• Agent: PPO

Usage (with skrl)

Note: Visit the skrl Examples section to access the scripts.

• PyTorch

  from skrl.utils.huggingface import download_model_from_huggingface
  
  # assuming that there is an agent named `agent`
  path = download_model_from_huggingface("skrl/IsaacGymEnvs-Ant-PPO", filename="agent.pt")
  agent.load(path)
  

• JAX

  from skrl.utils.huggingface import download_model_from_huggingface
  
  # assuming that there is an agent named `agent`
  path = download_model_from_huggingface("skrl/IsaacGymEnvs-Ant-PPO", filename="agent.pickle")
  agent.load(path)
  

Hyperparameters

Note: Undefined parameters keep their values by default.

# https://skrl.readthedocs.io/en/latest/api/agents/ppo.html#configuration-and-hyperparameters
cfg = PPO_DEFAULT_CONFIG.copy()
cfg["rollouts"] = 16  # memory_size
cfg["learning_epochs"] = 4
cfg["mini_batches"] = 2  # 16 * 4096 / 32768
cfg["discount_factor"] = 0.99
cfg["lambda"] = 0.95
cfg["learning_rate"] = 3e-4
cfg["learning_rate_scheduler"] = KLAdaptiveRL
cfg["learning_rate_scheduler_kwargs"] = {"kl_threshold": 0.008}
cfg["random_timesteps"] = 0
cfg["learning_starts"] = 0
cfg["grad_norm_clip"] = 0
cfg["ratio_clip"] = 0.2
cfg["value_clip"] = 0.2
cfg["clip_predicted_values"] = True
cfg["entropy_loss_scale"] = 0.0
cfg["value_loss_scale"] = 1.0
cfg["kl_threshold"] = 0
cfg["rewards_shaper"] = lambda rewards, timestep, timesteps: rewards * 0.01
cfg["state_preprocessor"] = RunningStandardScaler
cfg["state_preprocessor_kwargs"] = {"size": env.observation_space, "device": device}
cfg["value_preprocessor"] = RunningStandardScaler
cfg["value_preprocessor_kwargs"] = {"size": 1, "device": device}