How to futher tune hyper-parameters with the following trend lines of traning loss and validation loss?

Question

I am training a variational encoder (VAE) with MSE loss for 100 epochs, and I got the following training and validation loss. My questions are: