Author: Chris Yoon
Implementations of important policy gradient algorithms in deep reinforcement learning.
-
Advantage Actor-Critic (A2C)
Paper: "Asynchronous Methods for Deep Reinforcement Learning" (Mnih et al., 2016)
-
Asynchronous Advantage Actor-Critic (A3C)
Paper: "Asynchronous Methods for Deep Reinforcement Learning" (Mnih et al., 2016)
-
Deep Deterministic Policy Gradients (DDPG)
Paper: "Continuous control with deep reinforcement learning" (Lillicrap et al., 2015)
-
Twin Dueling Deep Deterministic Policy Gradients (TD3)
Paper: "Addressing Function Approximation Error in Actor-Critic Methods" (Fujimoto et al., 2018)
-
Soft Actor Critic (SAC)
- Paper (
sac2018.py
): "Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor" (Haarnoja et al., 2018) - Paper (
sac2019.py
): "Soft Actor-Critic Algorithms and Applications" (Haarnoja et al., 2019) - Algorithm in 2018 paper uses value network, double Q networks, and Gaussian policy. Algorithm in 2019 paper uses double Q networks and Gaussian policy, and adds automatic entropy tuning.
- TODO: SAC for discrete action space
- Paper (
More implementations will be added soon.
- Python 3.6
- PyTorch 1.0.2
- gym 0.12.5
Install package
git clone [email protected]:cyoon1729/Policy-Gradient-Methods.git
cd Policy-Gradient-Methods
pip install .
Example:
import gym
from policygradients.common.utils import mini_batch_train # import training function
from policygradients.td3.td3 import TD3Agent # import agent from algorithm of interest
# Create Gym environment
env = gym.make("Pendulum-v0")
# check agent class for initialization parameters and initialize agent
gamma = 0.99
tau = 1e-2
noise_std = 0.2
bound = 0.5
delay_step = 2
buffer_maxlen = 100000
critic_lr = 1e-3
actor_lr = 1e-3
agent = TD3Agent(env, gamma, tau, buffer_maxlen, delay_step, noise_std, bound, critic_lr, actor_lr)
# define training parameters
max_episodes = 100
max_steps = 500
batch_size = 32
# train agent with mini_batch_train function
episode_rewards = mini_batch_train(env, agent, max_episodes, max_steps, batch_size)