https://github.com/nikhilbarhate99/td3-pytorch-bipedalwalker-v2
Twin Delayed DDPG (TD3) PyTorch solution for Roboschool and Box2d environment
https://github.com/nikhilbarhate99/td3-pytorch-bipedalwalker-v2
bipedalwalker ddpg deep-reinforcement-learning lunar-lander openai-gym openai-gym-environments pytorch pytorch-implmention reinforcement-learning td3
Last synced: 5 months ago
JSON representation
Twin Delayed DDPG (TD3) PyTorch solution for Roboschool and Box2d environment
- Host: GitHub
- URL: https://github.com/nikhilbarhate99/td3-pytorch-bipedalwalker-v2
- Owner: nikhilbarhate99
- License: mit
- Created: 2018-11-29T04:28:48.000Z (almost 7 years ago)
- Default Branch: master
- Last Pushed: 2019-06-07T12:30:53.000Z (over 6 years ago)
- Last Synced: 2025-04-08T07:42:58.704Z (6 months ago)
- Topics: bipedalwalker, ddpg, deep-reinforcement-learning, lunar-lander, openai-gym, openai-gym-environments, pytorch, pytorch-implmention, reinforcement-learning, td3
- Language: Python
- Homepage:
- Size: 28.8 MB
- Stars: 105
- Watchers: 3
- Forks: 30
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# TD3-BipedalWalker-v2-PyTorch
PyTorch implementation of [Twin Delayed DDPG](https://arxiv.org/abs/1802.09477) (TD3) tested on the following environments:
- [Bipedal Walker v2](http://gym.openai.com/envs/BipedalWalker-v2/)
- [Lunar Lander Continuous v2](http://gym.openai.com/envs/LunarLanderContinuous-v2/)
- [Walker2d v1 (Roboschool)](https://github.com/openai/roboschool)
- [HalfCheetah v1 (Roboschool)](https://github.com/openai/roboschool)
### Usage
- To test a preTrained network : run `test.py`
- To train a new network : run `train.py`
## Dependencies
Trained and tested on:
```
Python 3.6
PyTorch 0.4.1
NumPy 1.15.3
gym 0.10.8
Roboschool 1.0.46
Pillow 5.3.0
```
## Results
BipedalWalker-v2 (800 episodes) | LunarLanderContinuous-v2 (1500 episodes)
:-------------------------:|:-------------------------:
 |  |
RoboschoolWalker2d-v1 (lr=0.002, 1400 episodes) | HalfCheetah-v1 (lr=0.002, 1400 episodes)
:-------------------------:|:-------------------------:
 |  |
*The results are not consistent for BipedalWalker-v2 env
## References
- Official [TD3 paper](https://arxiv.org/abs/1802.09477) and [code](https://github.com/sfujim/TD3)
- [OpenAI SpinningUp in Deep RL](https://spinningup.openai.com/en/latest/algorithms/td3.html)