While training, any sort of screen interruption (rotation, popup, back button, etc.) will cause the app to stop and attempting to return execution of the app will result in a black screen. I have debugging information logging when I am reading the training/test data files and after the screen has been interrupted I can see data files still being read, but no tensorflow operations are running, so this is also a huge memory leak.
While training, any sort of screen interruption (rotation, popup, back button, etc.) will cause the app to stop and attempting to return execution of the app will result in a black screen. I have debugging information logging when I am reading the training/test data files and after the screen has been interrupted I can see data files still being read, but no tensorflow operations are running, so this is also a huge memory leak.