FourierFlow

KO | EN

gitlite — search

#javascript #python #ai #golang #android #linux #docker #nodejs #windows #awesome #awesome-list #pytorch #security #openai

(★ 10)

🤖 A streamlined, easy‑to‑install implementation of Fourier Neural Operators for time‑dependent and cloud‑point learning tasks.

FourierFlow 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

<img src="assets/header_fourierflow.gif" />
<hr>

Fourier Flow

<hr>

In this repository, you will find the code for the:
- Fourier Neural Operators (**FNO**) : Base (V1/V2), Time Dependant & Cloud point

- Factorized Fourier Neural Operators (**FFNO**) : Base, Time Dependant  (V1/V2), & Cloud point

The reason behind the creation of this repository is to have something simpler and more convenient to install than what is available online. It is important to notice that I do not own the credit of most of the code !

<hr>

Fourier Neural Operator

<hr>

```python
import fourierflow.FourierNeuralOperator._2D_.Base as FNO_2D

# Loading a spectral convolutional layer
layer = FNO_2D.SpectralConv2d(in_channels = 3,
			out_channels = 1,
			modes1 = 12,
			modes2 = 12)

# Loading FNO
model = FNO_2D.FNO(input_channels = 3,
			output_channels = 1,
			modes1 = 12,
			modes2 = 12,
			width = 32,
			n_layers = 4)
```
<img src="assets/fno.png" />
<hr>

Factorized Fourier Neural Operator

<hr>

```python
import fourierflow.FactorizedFourierNeuralOperator._2D_.Base as FFNO_2D

# Loading a spectral convolutional layer
layer = FFNO_2D.SpectralConv2d(in_dim = 3,
			out_dim = 1,
			modes_x = 12,
			modes_y = 12,
			mode = "full",
			n_ff_layers = 2,
			factor = 4,
			dropout = 0.1,
			ff_weight_norm = True,
			use_fork = False,
			layer_norm  = False,
			fourier_weight = None,
			forecast_ff = None,
			backcast_ff = None)

# Loading FNO
model = FFNO_2D.FFNO(input_dim = 3,
			output_dim = 1,
			modes_x = 12,
			modes_y = 12,
			width = 32,
			n_layers = 4,
			factor = 4,
			n_ff_layers = 2,
			share_weight = True,
			ff_weight_norm = True,
			layer_norm  = False)
```

<img src="assets/ffno.png" />
<hr>

Installation

<hr>

```
pip install git+https://github.com/VikVador/FourierFlow
```

<hr>

Original Papers & GitHubs

<hr>
The authors of this remastered code are Zongyi Li and Daniel Zhengyu Huang ! This code is based on the papers :

Factorized Fourier Neural Operators
                          (https://arxiv.org/abs/2111.13802)

Fourier Neural Operator for Parametric Partial Differential Equations
                              (https://arxiv.org/abs/2010.08895)

and comes from:

https://github.com/alasdairtran/fourierflow

https://github.com/neuraloperator/neuraloperator

FourierFlow

FourierFlow 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

FourierFlow

FourierFlow 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README