Gap between train and valid loss

1

my dataset training is 30.000 and for testing 5000. I got a gap between training and valid

how can I solve this problem, please ? or it is not a problem ?

1