Statements (70)
Predicate | Object |
---|---|
gptkbp:instanceOf |
open-source library
|
gptkbp:developedBy |
gptkb:Anyscale
|
gptkbp:documentation |
https://docs.ray.io/en/latest/rllib/index.html
|
gptkbp:feature |
hyperparameter tuning
fault tolerance visualization tools policy evaluation cloud support policy gradient methods self-play custom callbacks custom environments model serving model-free RL multi-agent support curriculum learning imitation learning custom models custom metrics custom preprocessors multi-GPU support checkpointing policy optimization production-ready actor-critic methods model-based RL multi-node support batch RL high scalability custom policies customizable algorithms custom loss functions Q-learning methods algorithm registry asynchronous training custom exploration custom optimizers custom reward functions distributed rollout workers exploration strategies flexible APIs metrics logging multi-agent communication offline dataset ingestion parameter server architecture population-based training synchronous training |
gptkbp:firstReleased |
2018
|
gptkbp:focusesOn |
gptkb:reinforcement_learning
|
https://www.w3.org/2000/01/rdf-schema#label |
Ray RLlib
|
gptkbp:integratesWith |
gptkb:OpenAI_Gym
gptkb:TensorFlow gptkb:PyTorch gptkb:PettingZoo gptkb:RLlib_Tune |
gptkbp:latestReleaseVersion |
Ray 2.9 (as of 2024)
|
gptkbp:license |
gptkb:Apache_License_2.0
|
gptkbp:partOf |
right
|
gptkbp:programmingLanguage |
gptkb:Python
|
gptkbp:repository |
https://github.com/ray-project/ray
|
gptkbp:supports |
deep reinforcement learning
multi-agent reinforcement learning distributed training offline reinforcement learning |
gptkbp:usedBy |
gptkb:researchers
industry practitioners |
gptkbp:usedFor |
scalable RL workloads
|
gptkbp:bfsParent |
gptkb:Proximal_Policy_Optimization_(PPO)
gptkb:Ray_(distributed_computing) |
gptkbp:bfsLayer |
7
|