/PPO_Lagrangian_PyTorch

Implementation of PPO Lagrangian in PyTorch

Primary LanguagePythonMIT LicenseMIT

PPO Lagrangian Reproduction in Pytorch

Implementation of PPO Lagrangian from Benchmarking Safe Exploration in Deep Reinforcement Learning Paper (Ray et al, 2019) in PyTorch

python ppo.py

Results

  1. Reward Returns
    reward
  2. Cost Returns (Cost limit=25)
    cost