M
Modular14mo ago
sryu1

Has anyone tried implementing Reinforcement Learning in mojo yet?

^
3 Replies
Maximum Limelihood Estimator
No; the language is in the extremely early stages. It's still working on basic features for a programming language, like objects, classes, methods, traits, etc. If you want to do RL right now, I'd suggest trying JAX or Torch.compile
roboquant
roboquant14mo ago
If you want to learn more about RL, you could build some RL algo using mojo, as long as it is a small space. Good exercise I would say. For example TicTacToe can be done with a simple table for the Q function. Anything more serious requires typically deep learning networks to approximate Q (and V) functions. And this is not yet possible in native Mojo.
ModularBot
ModularBot14mo ago
Congrats @roboquant, you just advanced to level 1!
Want results from more Discord servers?
Add your server