Open moyi7712 opened 4 years ago
I think you don't have to use a GPU. The actor-critic algorithm is trained on a single multi-core CPU instead of a GPU. Please refer to the paper [1]. [1] MNIH, Volodymyr, et al. Asynchronous methods for deep reinforcement learning. In ICML. 2016.
thanks for your nice work but Chainer is the first time use it