Reinforcement Learning Algorithms in PyTorch
Based on https://github.com/ShangtongZhang/DeepRL hash 5d0ad07
- Use Marathon Envs instead of Majoco
- removed dependency on OpenAI Baselines
Implemented algorithms:
- Twined Delayed DDPG (TD3)
- PyTorch v1.4.0
- Python 3.6, 3.5
- Core dependencies: pip install -e .
- Download MarathonEnvs binary
- xxxxx show how ****
conda env create -f environment.yml
conda activate baby_rl
# manually install marathon-envs
pip install -r requirements.txt
pip install -e .
for tensorboard
tensorboard --logdir=tf_log
examples.py contains examples for all the implemented algorithms