PaddlePaddle/PARL
A flexible and high-efficient reinforcement learning framework for distributed training of RL agents.
Not currently ranked — collecting fresh signals.
star history
PARL is a high-performance distributed training framework for reinforcement learning developed by PaddlePaddle. It provides abstractions for Model, Algorithm, and Agent to build RL agents that perform complex tasks. The framework supports large-scale parallelization across thousands of CPUs and multiple GPUs, and includes implementations of influential RL algorithms that can be reproduced and adapted to new tasks.
Frequently asked
- What is PaddlePaddle/PARL?
- A flexible and high-efficient reinforcement learning framework for distributed training of RL agents.
- Is PARL open source?
- Yes — PaddlePaddle/PARL is open source, released under the Apache-2.0 license.
- What language is PARL written in?
- PaddlePaddle/PARL is primarily written in Python.
- How popular is PARL?
- PaddlePaddle/PARL has 3.5k stars on GitHub.
- Where can I find PARL?
- PaddlePaddle/PARL is on GitHub at https://github.com/PaddlePaddle/PARL.