randomly sample a batch of data for training :int batch_size: the number of data in a batch for Stochastic Gradient Descent :return torch.Tensor reward: reward.shape==(now_len, 1) :return torch.Tensor mask: mask.shape ==(now_len, 1), mask = 0.0 if done else gamma
(self, batch_size)
source not stored for this graph (policy: none)
no test coverage detected