sciai.architecture
Cover functions such as activation functions, basic blocks, multi-task loss, etc.
API Name |
Description |
Supported Platforms |
Adaptive activation function with trainable Parameter and fixed scale. |
|
|
Network that return the first output of given network. |
|
|
The 1-dimensional Fourier Neural Operator (FNO1D) contains a lifting layer, multiple Fourier layers and a decoder layer. |
|
|
The 2-dimensional Fourier Neural Operator (FNO2D) contains a lifting layer, multiple Fourier layers and a decoder layer. |
|
|
The 3-dimensional Fourier Neural Operator (FNO3D) contains a lifting layer, multiple Fourier layers and a decoder layer. |
|
|
Get the activation function according to its name. |
|
|
Multi-layer perceptron. |
|
|
Multi-layer perceptron with adaptive activation function. |
|
|
Multi-layer perceptron with shortcuts. |
|
|
Mean square error with 0. |
|
|
Convert a net with inputs into a net without inputs in construct. |
|
|
Normalize inputs with given lower bound and upper bound. |
|
|
The PDE-Net model. |
|
|
Sin rectified Linear Unit activation function. |
|
|
Sum square error with 0. |
|
|
Swish(Silu) activation function with backward propagation. |
|
|
This module based on ViT backbone which including encoder, decoding_embedding, decoder and dense layer. |
|