To know more about TorchRL philosophy, the problem it is trying to solve and get some idea about its general capabilities, refer to the TorchRL paper.
We also have some introductory videos for you to get to know the library better, check them out:
See here
See here
- Spinning Up in Deep RL
- Hugging face syllabus
- RL lectures from Deepmind
- Best benchmarks
- Awesome RL: Reinforcement learning resources curated
- A Succinct Summary of Reinforcement Learning
For completeness, we provide a list of RL libraries. Some are not actively maintained but they still provide good examples of RL solutions with and without PyTorch:
- ReAgent: RL in production oriented settings
- Stable-baselines3
- TF-Agent: Bandits and RL algorithms in tensorflow
- keras-rl: Deep Reinforcement Learning for Keras
- Acme: a research framework for reinforcement learning
- dopamine: research framework for fast prototyping of reinforcement learning algorithms
- salina: A Flexible and Simple Library for Learning Sequential Agents (including Reinforcement Learning)
- tianshou: An elegant PyTorch deep reinforcement learning library.
- rlpyt: Reinforcement Learning in PyTorch
- RLlib: Industry-Grade Reinforcement Learning with TF and Torch
- sample-factory: High throughput asynchronous reinforcement learning
- cherry: A PyTorch Library for Reinforcement Learning Research
- JaxRL: implementation of algorithms for Deep Reinforcement Learning with continuous action spaces
- MBRL-lib: Library for Model Based RL in pytorch
- RLMeta: light-weight flexible framework for Distributed Reinforcement Learning Research.
- ElegantRL: Cloud-native Deep Reinforcement Learning
- MTRL: Multi Task RL Baselines