The loss curve is normal？

[Passenger12138]

I am now training on the celebA dataset. I am not sure whether the training results are not good because of the unreasonable settings of the training hyperparameters or the insufficient number of iterations. I would like to know the correct loss curve style to help me further understand.