I'm planning an RL project and I have to decide which RL framework do I use if any at all. The project has a highly custom environment, and testing different algorithms will be required to obtain optimal results. Furthermore, it will use a custom neural network, not implemented in the popular TensorFlow/PyTorch ML frameworks. Therefore, the framework should allow for customization with regard to approximation function (1) and the environment (2). The problem is that to my current knowledge, most of the framework allows only to work with a built-in environment. Does anybody know a framework that meets the two conditions (1) and (2)? Or anybody knows a review that contains information about framework in the context of those conditions?

  • $\begingroup$ Although I asked you to clarify your actual question, I don't think that asking "which library/framework is most suited for my task" is on-topic here, as our on-topic page (ai.stackexchange.com/help/on-topic, please, read our on-topic page) says. Given that this is related to reinforcement learning, I will leave it open for a while, but questions that ask for libraries, datasets or GPUs are probably more suited for Data Science SE, but I don't ensure you that they are on-topic there, given I am not familiar with the details of that site. $\endgroup$
    – nbro
    Sep 18, 2020 at 13:16
  • $\begingroup$ You can modify the environments and create custom models using Ray RLlib: docs.ray.io/en/latest/rllib.html $\endgroup$
    – Tom Dörr
    Oct 18, 2020 at 16:40


Browse other questions tagged .