Hierarchical Reinforcement envs in pybullet

This package was created because all of the HRL locomotion envs are only available in mujoco. This is an implementation of as many as possible in pybullet.

Install

pip install pybullet hrl_pybullet_envs
This project requires pybullet-gym which must be installed along side this package.

Envs:

• AntGatherBulletEnv-v0
• AntMazeBulletEnv-v0
• AntMjBulletEnv-0
• AntFlagrunBulletEnv-v0
• PointGatherBulletEnv-v0

Example

Also see this notebook

import hrl_pybullet_envs
import gym
import numpy as np

env = gym.make('AntGatherBulletEnv-v0')
env.render()
ob = env.reset()
tot_rew = 0

for i in range(1000):
  # Take random actions
  ob, rew, done, _ = env.step(np.random.uniform(-1, 1, env.action_space.shape))
  tot_rew += rew

  if done: break

print(f'Achieved total reward of: {tot_rew}')