DarkDawn233/SeCA

Codes of SeCA accompanying the paper "Sequential Cooperative Multi-Agent Reinforcement Learning"(AAMAS 2023). SeCA is a sequential credit assignment method that factorizes and simplifies the complex interaction analysis of multi-agent systems into a sequential evaluation process for more efficient learning.

PythonMIT

Issues

FileNotFoundError: [Errno 2] No such file or directory: '/SeCA/3rdparty/StarCraftII/Versions'
#5 opened 7 months ago by haytamelouarrat00
0
unable to run replay - Error in torch.size
#4 opened a year ago by shahaamirbader
0
How to generate graphs similar to Figure 6?
#3 opened a year ago by shahaamirbader
0
How to distinguish the results in tb_logs as per each map
#2 opened a year ago by shahaamirbader
1
Understanding the output on terminal while training
#1 opened a year ago by shahaamirbader
5