d3rlpy - A data-driven deep reinforcement learning library as an out-of-the-box tool.

d3rlpy is a easy-to-use data-driven deep reinforcement learning library.

d3rlpy provides state-of-the-art data-driven deep reinforcement learning algorithms through out-of-the-box scikit-learn-style APIs. Unlike other RL libraries, the provided algorithms can achieve extremely powerful performance beyond the paper via several tweaks.

These are the design principles of d3rlpy:

• d3rlpy is desined for practical projects.
• d3rlpy provides state-of-the-art techniques across RL researches to let everyone easily achieve an incredible performance.
• d3rlpy provides scikit-learn-compatible APIs.

Tutorials

• Jupyter Notebooks

References

• API Reference
  • Algorithms
  • Q Functions
  • MDPDataset
  • Datasets
  • Preprocessing
  • Data Augmentation
  • Metrics
  • Save and Load
  • Logging
  • scikit-learn compatibility
  • Online Training
  • Model-based Data Augmentation
• Installation
  • Recommended Platforms
  • from pip
  • from source

Other

• License

Indices and tables

• Index
• Module Index
• Search Page