/kuka_rl

Reinforcement Learning Experiments using PyBullet

Primary LanguageJupyter NotebookApache License 2.0Apache-2.0

Robotic Grasping

Machine learning implementations using OpenAI Gym and PyBullet for robotic grasping.

Deep Q Learning (DQN) Open In Colab

Proximal Policy Optimization (PPO) Open In Colab

  • Continuous control problem is considered.
  • kuka_rl_2

A2C Parallel Implementation

  • Twice the number of available CPU threads is considered for the KUKA environment to run in parallel.
  • kuka_rl_2_parallel

Dependencies

pip install pybullet
pip install gym