Situation:
Tensorflow2.0.0 GPU + win10 + configured corresponding environment
Situation: training my own model in infinite loop is unable to stop (no problem oh my code); If you don't believe that my code, then run the minist data set an example of the club's official website, is also unable to stop, after training the epoch automatic start from scratch again, can only think that stop!
Origin: look at the online posts, want to have a look at the changing situation of the utilization ratio of GPU nvidia and look at the below this post wrote - smi - 500 command LMS, post in which there are about 500 ms refresh once, and then look after stop running program for the first time is normal, the second after the start will never stop,,,,
https://blog.csdn.net/ProLayman/article/details/104758331
Pray god to solve!