- This repository will be filled with codes reproducing some DRL algos I'm interested in.
- Most algorithms remain untested because my PC is occupied by my current research experiments.
- Language: Python-3.6
- Main library: PyTorch-1.3.0, Mujoco-py-2.0.2.8, Gym-0.15.3
- DQN - Deterministic, Discrete (LSTM network for Atari)
- DDPG - Deterministic, Continuous
- PPO - Stochastic, Continuous
- TD3 -Deterministic, Discrete
- SAC (Adaptive Temperature) - Stochastic, Continuous
- D4PG - Deterministic, Continuous
- R2D2 - Deterministic, Discrete
- Option DQN - Hindsight, Deterministic, Discrete
- Option Critic - Hindsight, Stochastic, Discrete & Continuous
- HIRO - Hindsight, Deterministic, Continuous
- HAC - Hindsight, Deterministic, Continuous
- Hindsight
- Prioritised
- GridWorld_MultiRoomKeyDoor (Discrete, Multi-goal, Customized)
- OpenAI Gym Mujoco Robotics Multigoal Environment (Continuous, Official)
- Pybullet Multigoal Gym (OpenAI Robotics Multigoal Pybullet Migration) (Continuous, Official)
- OpenAI Gym Mujoco Robotic Multi-goal/task/stage Environment (Continuous, Customized)
- DQN
- DoubleDQN
- LSTM network on raw Atari pixel observation
- DDPG
- TD3
- SAC (Adaptive Temperature)
- PER
- HER
- HIRO
- HAC
- OptionFramework
- OptionCritic
- D4PG
- R2D2
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent