validation loss looks like overfitting

[Maryamzk]

Hi,
I have applied this model to other data sets.
I want to forecast the energy consumption using a history of energy consumption and 23 weather features over 4 years.

The validation loss is increasing, which I assume it is a sign of overfitting. and the training loss does not go under 0.2.
I have tried decreasing the learning rate, adding decay rate, and reducing the lstm layers, but I still have overfitting.

How can I modify my model to prevent overfitting (the increasing trend in the validation loss?

Thanks for any advice.