Fno

continuiti.operators.fno

The Fourier Neural Operator

FourierNeuralOperator(shapes, depth=3, width=3, act=None, device=None, **kwargs)

Bases: NeuralOperator

Fourier Neural Operator (FNO) architecture

Reference: Z. Li et al. Fourier Neural Operator for Parametric Partial Differential Equations arXiv:2010.08895 (2020)

PARAMETER	DESCRIPTION
shapes	Shapes of the input and output data. TYPE: OperatorShapes
depth	Number of Fourier layers. TYPE: int DEFAULT: 3
width	Latent dimension of the Fourier layers. TYPE: int DEFAULT: 3
act	Activation function. TYPE: Optional[Module] DEFAULT: None
device	Device. TYPE: Optional[device] DEFAULT: None
**kwargs	Additional arguments for the Fourier layers. DEFAULT: {}

Source code in src/continuiti/operators/fno.py

def __init__(
    self,
    shapes: OperatorShapes,
    depth: int = 3,
    width: int = 3,
    act: Optional[torch.nn.Module] = None,
    device: Optional[torch.device] = None,
    **kwargs,
):
    latent_shapes = OperatorShapes(
        x=shapes.x,
        u=TensorShape(width, shapes.u.size),
        y=shapes.x,
        v=TensorShape(width, shapes.u.size),
    )
    output_shapes = OperatorShapes(
        x=shapes.x,
        u=TensorShape(width, shapes.u.size),
        y=shapes.y,
        v=TensorShape(width, shapes.v.size),
    )

    layers = []
    for _ in range(depth - 1):
        layers += [FourierLayer(latent_shapes, device=device, **kwargs)]
    layers += [FourierLayer(output_shapes, device=device, **kwargs)]

    layers = torch.nn.ModuleList(layers)

    super().__init__(shapes, layers, act, device)

---

Last update: 2024-08-20
Created: 2024-08-20