actions_v = torch.tensor(actions).to(device) rewards_v ... def _get_conv_out(self, shape): o = self.conv(torch.zeros(1, *shape)) return ...
確定! 回上一頁