VMPO-PyTorch

Minimal PyTorch implementation of V-MPO: On-Policy Maximum a Posteriori Policy Optimization for OpenAI gym environments.

Modified from nikhilbarhate99/PPO-PyTorch

Usage

• To train a new network : run VMPO.py

Dependencies

Trained and tested on:

Python 3.6
PyTorch 1.0
NumPy 1.15.3
gym 0.10.8
Pillow 5.3.0

References

• VMPO paper
• OpenAI Spinning up