A toolkit for reproducible reinforcement learning research.
Role in this project:
ML Engineer Contributions:1 release, 96 commits, 178 PRs in 1 year 9 months
Contributions summary:Anson implemented and tested Multi-Layer Perceptron (MLP) and Convolutional Neural Network (CNN) models, along with the relevant unit tests. They added functionality for layer normalization within MLPs and expanded the test suite. The user also added gym.Env wrappers, particularly for image-based environments, including resizing, grayscale conversion, and frame stacking. The user made modifications to address code style issues and added RL2 support.
reinforcement-learningrl-algorithmsreproducibilitypytorchtensorflow
Contributions:51 pushes, 1 branch in 9 years 2 months