The code now supports dynamic loss weighting for the Fourier loss function. By setting dynamic_weight=True during initialization, the model can adaptively adjust the importance of the Fourier loss during training using a learnable parameter weight. This allows for more flexible optimization and potentially improved model performance.
The code now supports dynamic loss weighting for the Fourier loss function. By setting dynamic_weight=True during initialization, the model can adaptively adjust the importance of the Fourier loss during training using a learnable parameter weight. This allows for more flexible optimization and potentially improved model performance.