PyTorch-Tutorial

PyTorch-Tutorial copied to clipboard

About Memory Storage

1

莫烦老师，在记忆存储模块，我运行代码使其输出index的值，发现他的数值一值是在1-200之间循环显示，这是否代表存储的记忆只是每个回合的，并没有将记忆进行累计存储。或者说记忆库没有发挥作用，即时在记忆库没存储满之前，每次存储新的回合都会将老的记忆覆盖

zhangjianguo-max

你好!我在使用DQN算法时，无法确定智能体的状态，我的模型 是将队列里的任务分发给其他几个服务器，以期望获得最小时延，我开始是以队列里的任务个数来确定不行，请问我这个state如何确定？谢谢

wutianhaha

Modify argument in TensorDataset Error code: Data.TensorDataset(data_tensor=x, target_tensor=y) Fixed code: Data.TensorDataset(x, y)

YumiGo

retain_graph=True 相关问题

2

运行406_conditional_GAN时，出现以下错误： RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.FloatTensor [128, 1]], which is output 0 of TBackward, is at version 2;...

Kirito-E2D

Added another example of how to easily create a neural net with custom layer names - using OrderedDict from collections module Added another example of how to save and load...

Chief-Buka

DQN code chose_action

1

我发现老师在训练DQN的时候使用的是self.eval_net.forward(input),为什么这里不用self.eval_net(input)？

heyfavour

 I cope the teacher's code and run it in my pycharm, but there is no real time show about decoded_data. Hope to solve it. Thanks~ 运行老师代码并没有得到实时演示，非常想知道原因，万分感谢~

Youzi-ciki

动态显示图片失效

1

实时画动态图那部分 **114** 行 `plt.draw(); plt.pause(0.05)` 并没有实现动态显示解码结果的图片，第二行的画图部分应该是只运行过一次，结果如下所示 > python 3.8 > pytorch 1.4.0

ablueday

应是 return outs,h_state 感谢MorvanZhou老师分享！

mingyuezhi

Why reain_graph is needed here ?

1

Hi, I read the GAN code and I have the confusion that why the retain_graph is needed here ? https://github.com/MorvanZhou/PyTorch-Tutorial/blob/4e5122c3734dcdc9ace164c2a87fbd412ca1d431/tutorial-contents/406_GAN.py#L71 The computation graph of discriminator should not be released ?

LeoniusChen