python - how many epoch for training 1k images

Question

am doing training for detecting the objects using yolov3 but i face some problem when i set batch_size > 1 it causes me cuda out of memory so i searched in google to see another solution found it depends on my GPU (GTX 1070 8G) .

may be the number of epoch is high and it require to be optimized .

maybe the epoch number should be decreased? and for training 1k images and 200 pic for validations .

what is best epoch should i set to avoid overfitting?

question from:https://stackoverflow.com/questions/65672181/how-many-epoch-for-training-1k-images

Answer 1

Your model's overfitting wont depend on the no. of epochs you set.....
since you hav made a val split in your data, make sure that your train loss - val loss OR train acc - val acc is nearly the same.This will assure that your model is not overfitting