Has anyone used Ray and RLLib to implement any reinforcement learning? I've taken the Deep Reinforcement Learning in Trading course and I'm trying to conver that to train faster. If you have any resources, would love to learn more.
Hi Claud,
I am not too familiar with Ray and RLLib library. If you have any specific questions about these libraries you can also ask them on Stack Overflow.
Since Deep Reinforcement Learning models are quite data-hungry making them train faster would not be as straightforward. But here are some things that can be done to make these models train faster you can upgrade your hardware like GPU. You can also try to optimise your algorithm. You can change parameters and try different models and see which one works best for you.
Here are some articles you will find helpful.
How to speed up Deep Reinforcement Learning with PPMP and RLHF? | Xomnia
Fast reinforcement learning through the composition of behaviours - Google DeepMind
https://medium.com/distributed-computing-with-ray/intro-to-rllib-example-environments-3a113f532c70
Thank you, this will help. The last article you linked is Ray and RLLib that I mentioned. It can help speed things up and could be a good addition to the course as an update.