Commit graph

18 commits

Author SHA1 Message Date
Vasilis Valatsos
8e2664533f Removed grad clip 2023-12-10 11:59:27 +01:00
Vasilis Valatsos
939a90dd0f Updated grad clips 2023-12-10 06:55:07 +01:00
Vasilis Valatsos
85c1532920 Updated critic to use leaky ReLu 2023-12-09 13:48:16 +01:00
Vasilis Valatsos
948ae9af4f Once again player with rewards, and added clipping to the params 2023-12-08 22:08:25 +01:00
Vasilis Valatsos
3b9b25441e removed redundant if is_dead() check 2023-12-06 14:14:43 +01:00
Vasilis Valatsos
6d316834d3 Updated reward and state features 2023-12-06 13:58:00 +01:00
Vasilis Valatsos
0aea150454 exp per player now resets for each episode 2023-12-06 13:03:59 +01:00
Vasilis Valatsos
ca29a0e6dc Implemented Maxwell distribution for rewards 2023-12-04 05:08:41 +01:00
Vasilis Valatsos
bebc660060 Updated enemy detection 2023-11-30 18:28:17 +01:00
Vasilis Valatsos
c7ddaa630c fixed three bugs 2023-11-29 12:10:04 +01:00
Vasilis Valatsos
84000dd28b Major update, made single main file for both multi and single agent, added argsparse, polished everything 2023-11-29 11:53:30 +01:00
Vasilis Valatsos
0bdcc8ca6f Found bug again, fixed bug again 2023-11-25 00:26:47 +01:00
Vasilis Valatsos
a9868e6c1a Weird bug showed, added diagnostics 2023-11-24 21:15:50 +01:00
Vasilis Valatsos
1f91ec9d5d Training done 2023-11-24 14:23:12 +01:00
Vasilis Valatsos
8809c1b06c Fixed errors for MARL 2023-11-23 16:37:02 +01:00
Vasilis Valatsos
3fb147afff Updated some stuff 2023-11-20 01:51:54 +01:00
Vasilis Valatsos
da649ccca8 Added more rewards 2023-11-19 04:27:47 +01:00
Vasilis Valatsos
115b2e4151 Hopefully implemented PPO 2023-11-17 03:19:03 +01:00