Manually destroying AMDSI workspace Debug by ND4J AsyncMultiDataSetIterator

Hello Everyone, I’m new to the forum here so I’m sorry if my question is out of place :pray:

Recently when using TinyYolo example given by the deeplearning4j.examples repository I found that somehow I got this debug massage when training my dataset.

It’s was not there before and I was able to train my data splendidly (Most of them is 90% accuracy just at 500 epoch) but now after I got this debug message it ruin all the trained model and it will not detect anything anymore (even if it did the bounding box is just too big and out of the image, the width or height of detected objects can go to 10k mark but the training and test data is just 416x416 resolution)

And after that the message will never go, even if I remake my project from scratch using the deeplearning4j TinyYolo example again. Here is some screenshot that shows the debug message and right after that the training just bounce like crazy and it eventually will make the model unusable again like when it corrupt my trained model before. Hopefully anyone know why the debug occurs.

After reading similar topic that someone else posted here, I found that it just a debug message for Preloading the training images for the Network to use by the .fit function and It should not bring any harm for the training itself.

And since I don’t know how to close this thread I’ll just leave it here and will do the debugging process once again to find the real culprit of my training problem. Thank You Very Much :pray: