thanks for explaining

but how do i tell the model that train until accuracy improves n number of times