Make a simple Graph Neural Network and plot some images of training in frequency domain
Budget: $30 – $250 USD
I have a code here which runs a DNN with 1D data and produces a fourier transform plot of the training: https://github.com/xuzhiqin1990/F-Principle
I want the exact same output as this code! But it needs to work for 2D simulated data. e.g. make a code which simulates random points and connects edges through an adjacency matrix. Find attached an image of training at the 8,500th epoch, comparing the GNN output and the true data in the fourier domain. Note that BasicFunc.py produces all plots, so maybe you don't need to change this code?? but maybe you do as it needs to work for 2d data not 1d.
I want the exact same output as this code! But it needs to work for 2D simulated data. e.g. make a code which simulates random points and connects edges through an adjacency matrix. Find attached an image of training at the 8,500th epoch, comparing the GNN output and the true data in the fourier domain. Note that BasicFunc.py produces all plots, so maybe you don't need to change this code?? but maybe you do as it needs to work for 2d data not 1d.