Learning Attentional Communication with a Common Network for Multiagent Reinforcement Learning

Table 3

Experimental results of GPU memory cost for different algorithms.

Scenes	MAACCN-VDN	TarMAC-VDN	CommNet-VDN	VDN
2s_vs_1sc	1490M	1120M	690M	680M
5m_vs_6m	1550M	1150M	710M	680M
3s_vs_5z	2010M	1480M	750M	730M