How severe does this issue affect your experience of using Ray?
- None: Just asking a question out of curiosity
Hello. I was curious if there’s any official plans to add an algorithm that makes use of Implicit Quantile Networks to compute Q-value distributions, instead of just the standard Distributional-Q learning? (An extra reading source, if you’re curious, though I know QR-DQN is also not currently implemented in RLLib: How does Implicit Quantile-Regression Network (IQN) differ from QR-DQN?)
If not, I’m curious about opinions on how difficult it could be if I tried to do this myself? I don’t have any issues with the theory of DQNs and IQNs and standard Distributional DQNs, but I do not know RLLib under the hood that well.