fhswf / mlpro-int-sb3

MLPro: Integration StableBaselines3

MLPro-Int-SB3 - Integration of Stable-Baselines3 into MLPro

Welcome to MLPro-Int-SB3, an extension to MLPro to integrate the Stable-Baselines3 package. MLPro is a middleware framework for standardized machine learning in Python. It is developed by the South Westphalia University of Applied Sciences, Germany, and provides standards, templates, and processes for hybrid machine learning applications. Stable-Baselines3, in turn, provides a diverse suite of reliable implementations of reinforcement learning algorithms in PyTorch.

MLPro-Int-SB3 offers wrapper classes that allow the reuse of environments from Stable-Baselines3 in MLPro, or the other way around.

Learn more

MLPro - Machine Learning Professional
MLPro-Int-SB3 - Integration of Stable-Baselines3 into MLPro
Stable-Baselines3 - Reliable Reinforcement Learning Implementations

mlpro-int-sb3's Issues

Bug: MuJoCo-based Howtos don't perform any more

1 Describe the bug
Some of the (MuJoCo-based) howtos don't perform any more. They terminate every training episode after just 1 cycle. Affected howtos are:

• howto_rl_agent_021_train_and_reload_single_agent_mujoco_cartpole_discrete.py

Positive reference: howto_rl_agent_011_train_and_reload_single_agent_gym.py

2 To Reproduce
Steps to reproduce the behavior:

1. Go to '...'
2. Click on '....'
3. Scroll down to '....'
4. See error

3 Expected behavior
A clear and concise description of what you expected to happen.

4 Screenshots
If applicable, add screenshots to help explain your problem.

5 Additional Informations

5.1 Operating System

• Linux
• Windows
• Mac OS

Bug: Unit Test for MacOS has failed

1 Describe the bug
A clear and concise description of what the bug is.

2 To Reproduce
Steps to reproduce the behavior:

1. Go to '...'
2. Click on '....'
3. Scroll down to '....'
4. See error

3 Expected behavior
A clear and concise description of what you expected to happen.

4 Screenshots
If applicable, add screenshots to help explain your problem.

5 Additional Informations

5.1 Operating System

• Linux
• Windows
• Mac OS

Bug: Howtos that store the scenario run into a pickle/dill error

1 Describe the bug
See title. All howtos that store the results on the file system are affected.

2 To Reproduce
Steps to reproduce the behavior:

1. Run howtos rl_agent_, rl_att_ (temporarily reduce number of training cycles)

3 Expected behavior
A clear and concise description of what you expected to happen.

4 Screenshots
If applicable, add screenshots to help explain your problem.

5 Additional Informations

5.1 Operating System

• Linux
• Windows
• Mac OS