This repository contains codes for producing synthetic data, training the model and producing all figures in the article https://arxiv.org/abs/2210.09602
Process 1: produce the simulation data The codes for generating synthetic data with SFM and ORCA are in the folder "sfm_data" and "orca_data" respectively.
Process 2: train the ODE-Net Training data from the SFM and ORCA are provided in the folder "result/sfm" and "result/orca" respectively. Parameters of the neural network are saved as the trained model.
Process 3: test the ODE-Net Run "draw_sfm.py" or "draw_orca.py" to draw figures and animations of the test result of the trained ODE-Net.
2022/10/20