Skip to content

Latest commit

 

History

History
30 lines (25 loc) · 545 Bytes

README.md

File metadata and controls

30 lines (25 loc) · 545 Bytes

irlax

a simple and customizable library for training RL and imitation learning models using flax !!! this library is still in development and is not yet ready for use !!!

Installation

pip install irlax

roadmap

RL algorithms

  • DQN
    • DQN
    • Double DQN
    • QR-DQN
  • PPO
  • TD3

Imitation learning algorithms

  • BC
  • DAgger

Replay Buffers

  • Uniform Replay Buffer
  • Prioritized Replay Buffer
    • python
    • concurrent rust implementation

callbacks

  • learning curve plotter