Conversation
|
Hi @jsuarez5341, sorry for being out for the last 3 months, I got a new job and stuff. However, I was able to improve the policy and solve Boids successfully with all of the rewards. P.S. When I get situated in my new job(couple weeks hopefully) I'll come back to PufferLib and contribute in my free time. |
|
@jsuarez5341 I see that this PR has not been look at( Probably because I forgot to leave it as a draft). But I think it will need to be rebased to Puffer 4.0. If you would like me to do that, then let me know, BOIDS is not relevant longer and isn't worth working on anymore please let me know as well. |
|
Closing PR as I just made a new PR migrating the improvements of this PR to 4.0 |
Description
This PRs goal is to improve the reward calculation of the boids env and train a policy on it
Todo
Improve reward calculations
Train policy on two factors
Train policy successfully on all factors