sksksk94/distributed_sac

Pytorch Implementation of the Distributed SAC. Test environment is LunarLanderContinuous-v2 and Metaworld MT1, MT10

Distributed SAC

Getting Started

You can run this repository as follows:

[Important] Before running, you should set your own ray parameters that fit your computer in respective main.py.

Choose one of these.

1. To run LunarLander_Distributed_SAC, python3 LunarLander_Distributed_SAC/src/main.py
2. To run MT1_Distributed_VSAC, python3 MT1_Distributed_VSAC/src/main.py
3. To run MT1_Distributed_CARE, python3 MT1_Distributed_CARE/src/main.py
4. To run MT10_Distributed_MTSAC, python3 MT10_Distributed_MTSAC/src/main.py
5. To run MT10_Distributed_CARE, python3 MT10_Distributed_CARE/src/main.py

Test environment

• CPU : i7-6700k
• GPU : GTX 1080
• OS : Ubuntu 18.04

Reference papers

1. Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor
2. Soft Actor-Critic Algorithms and Applications
3. Meta-World: A Benchmark and Evaluation for Multi-Task and Meta Reinforcement Learning
4. Multi-Task Reinforcement Learning with Soft Modularization
5. Multi-Task Reinforcement Learning with Context-based Representations

distributed_sac's Issues

environment

Can you tell me the specific version of each package used by my environment?

Running error

Hi, I was using your code on windows. The error information is as follows:
redis.exceptions.AuthenticationError: Authentication required. Could you please help me, thanks.