sanjay126 / violencedetection Goto Github PK

Thank you for sharing your amazing modification, I am beginner in deep learning :)
when i am turned cross validation to false, my data are divided into training, validation and testing. say 100 training samples and 50 validation sample. after each 5 epoch, 50 validation sample is used to validate my model performance. my question is in every epoch we use same training and validation data?? or they changed randomly in every epoch? if yes then we use cross validation or not? i am confused in this point.

Thanks in advance :)

I would like to extract all frames instead of 20 frames only. Could you let me know how to do this in the dataset that has videos with a different number of frames?