CORL: Research-oriented Deep Offline Reinforcement Learning Library

Denis Tarasov, Alexander Nikulin, Dmitry Akimov, Vladislav Kurenkov, Sergey Kolesnikov
10/13/2022
cs.LGcs.AI
rl

Official code: tinkoff-ai/CORL (1,376 stars, Apache-2.0, last commit 2023-08-03). 2 more implementations below.

Abstract

CORL is an open-source library that provides thoroughly benchmarked single-file implementations of both deep offline and offline-to-online reinforcement learning algorithms. It emphasizes a simple developing experience with a straightforward codebase and a modern analysis tracking tool. In CORL, we isolate methods implementation into separate single files, making performance-relevant details easier to recognize. Additionally, an experiment tracking feature is available to help log metrics, hyperparameters, dependencies, and more to the cloud. Finally, we have ensured the reliability of the implementations by benchmarking commonly employed D4RL datasets providing a transparent source of results that can be reused for robust evaluation tools such as performance profiles, probability of improvement, or expected online performance.

View on arXivView PDF

Code Implementations(3)

tinkoff-ai/CORLOfficial100%
Apache-2.0

High-quality single-file implementations of SOTA Offline and Offline-to-Online RL algorithms: AWAC, BC, CQL, DT, EDAC, IQL, SAC-N, TD3+BC, LB-SAC, SPOT, Cal-QL, ReBRAC

1,376177Python, DockerfileSep 23, 20223 years agoApache-2.0
d4rlgymoffline-reinforcement-learningreinforcement-learning
corl-team/CORLOfficial100%
Apache-2.0

High-quality single-file implementations of SOTA Offline and Offline-to-Online RL algorithms: AWAC, BC, CQL, DT, EDAC, IQL, SAC-N, TD3+BC, LB-SAC, SPOT, Cal-QL, ReBRAC

66339Python, DockerfileAug 10, 20232 years agoApache-2.0
hanjuku-kaso/awesome-offline-rlCommunity100%

An index of algorithms for offline reinforcement learning (offline-rl)

1,08192Dec 4, 20202 years ago
awesomeawesome-listoff-policy-evaluationoffline-rlreinforcement-learning+1 more

Similar Papers

D4RL: Datasets for Deep Data-Driven Reinforcement Learning

86% match1 repo1.7k2020

NeoRL: A Near Real-World Benchmark for Offline Reinforcement Learning

86% match2 repos1512021

SCOPE-RL: A Python Library for Offline Reinforcement Learning and Off-Policy Evaluation

85% match1 repo1432023

Active Offline-to-Online Reinforcement Learning

85% match2026

SAMG: Offline-to-Online Reinforcement Learning via State-Action-Conditional Offline Model Guidance

84% match2024

A Minimalist Approach to Offline Reinforcement Learning

84% match1 repo4132021

Cite this paper

@article{tarasov2022corl,
  title  = {CORL: Research-oriented Deep Offline Reinforcement Learning Library},
  author = {Denis Tarasov and Alexander Nikulin and Dmitry Akimov and Vladislav Kurenkov and Sergey Kolesnikov},
  year   = {2022},
  eprint = {2210.07105},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/2210.07105}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2210.07105.svg)](https://codeofpaper.com/paper/2210.07105)

Discussion