ICQ-MA ：Why is the greedy strategy not used in the testing phase?

YiqinYang / ICQ

Codes accompanying the paper "Believe What You See: Implicit Constraint Approach for Offline Multi-Agent Reinforcement Learning" (NeurIPS 2021 Spotlight https://arxiv.org/abs/2106.03400)

70 stars 8 forks source link

ICQ-MA ：Why is the greedy strategy not used in the testing phase? #8

Open fanchunpeng opened 2 years ago

fanchunpeng commented 2 years ago

**'test_greedy': False,**
'test_interval': 20000,
'test_nepisode': 20,
'use_cuda': True,
'use_tensorboard': True}