Integrations of OpenML in PyTorch¶
Along with this PyTorch API, OpenML is also integrated in PyTorch through the following modules.
Reinforcement Learning¶
- The RL library TorchRL supports loading OpenML datasets as part of inbuilt modules.
TorchRL - OpenMLExperienceReplay¶
- Experience replay is a technique used in reinforcement learning to improve the stability and performance of deep reinforcement learning algorithms by storing and reusing experience tuples.
- This module provides a direct interface to OpenML datasets to be used in experience replay buffers.
TorchRL - OpenMLEnv¶
- Bandits are a class of RL problems where the agent has to choose between multiple actions and receives a reward based on the action chosen.
- This module provides an environment interface to OpenML data to be used in bandits contexts.
- Given a dataset name (obtained from openml datasets), it returns a PyTorch environment that can be used in PyTorch training loops.