https://github.com/fkodom/ddpg-her-pytorch

Implementation of the Hindsight Experience Replay (HER) algorithm using PyTorch. Utilizes Deep Deterministic Policy Gradients for off-policy optimization of RL agents in continuous action spaces.
https://github.com/fkodom/ddpg-her-pytorch

Last synced: 7 days ago
JSON representation

Implementation of the Hindsight Experience Replay (HER) algorithm using PyTorch. Utilizes Deep Deterministic Policy Gradients for off-policy optimization of RL agents in continuous action spaces.

• Host: GitHub
• URL: https://github.com/fkodom/ddpg-her-pytorch
• Owner: fkodom
• License: gpl-3.0
• Created: 2019-07-09T20:59:53.000Z (over 5 years ago)
• Default Branch: master
• Last Pushed: 2019-07-10T01:18:04.000Z (over 5 years ago)
• Last Synced: 2024-10-28T06:30:00.081Z (about 2 months ago)
• Language: Python
• Size: 2.36 MB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# DDPG+HER-PyTorch

Implementation of the [Hindsight Experience Replay](https://arxiv.org/abs/1707.01495) algorithm using PyTorch.  Utilizes [Deep Deterministic Policy Gradients](https://arxiv.org/pdf/1509.02971.pdf) for off-policy optimization of the RL agent -- hence, DDPG+HER.  Includes (for now) just one pre-trained example agent (FetchReach-v1), as well as a training script for creating new agents.

### FetchReach-v1 Agent

![DDPG+HER FetchReach-v1 Animation](https://raw.githubusercontent.com/fkodom/ddpg-her-pytorch/master/figures/fetch-reach.gif)