Caltech-geoquake / GM-GANO

Code for "Broadband Ground Motion Synthesis via Generative Adversarial Neural Operators: Development and Validation"

cGM-GANO

The repository contains code for Broadband Ground Motion Synthesis via Generative Adversarial Neural Operators: Development and Validation , for more information about GANO implementation, please refer to GANO

Installation

create conda environment and install necessary libraries

conda create --name gano

conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia

conda install -c anaconda ipykernel

pip install ipykernel

python -m ipykernel install --user --name=gano

conda install pandas

conda install matplotlib
 
pip install scipy

pip install tqdm

Load the pretrained model

please download the trained model through following link, and store the model under kik_net_trained_model folder https://drive.google.com/file/d/18k366Y4UmaGoYxepwzZaGo_nw6Kup0cW/view?usp=sharing

Quick start of using trained GANO

run tutorials_for_cGmGANO.ipynb file for generating 3C waveforms

Train your own cGM-GANO

run train_GANO.ipynb