自己的数据集，内存一直在涨，请问什么原因呢？

Pascalson / Conditional-Seq-GANs

GANs for Conditional Sequence Generation. Tensorflow. Including the code of paper "Improving Conditional Sequence Generative Adversarial Networks by Stepwise Evaluation" IEEE/ACM TASLP, 2019.

MIT License

34 stars 6 forks source link

自己的数据集，内存一直在涨，请问什么原因呢？ #3

Open gumanchang opened 4 years ago

gumanchang commented 4 years ago

使用自己的数据集训练MLE，内存一直在涨，请问什么原因呢？

Pascalson commented 4 years ago

Hi,

Sorry for my late reply. I have not met that problem. A similar problem is that the usage of RAM will first be only a small part. After constructing all the computation graph, the usage of RAM will be approximately fixed and larger than the initial. But I'm not absolutely sure about the reason. Hope this would help.