This quake does not exist

Generative modelling of seismic waveforms using denoising diffusion.

About

This repository contains the experimental code of the manuscript High Resolution Seismic Waveform Generation using Denoising Diffusion. It can be used to generate seismic waveforms, replicate the results from the manuscript, and for training custom models from scratch.

Installation

To set up the environment and install dependencies:

1. Clone the repository:

   git clone --depth 1 https://github.com/highfem/tqdne.git
   cd tqdne

   Omit --depth 1 if you want to access the full commit history. Alternatively, download the latest release if you do not require commmit history.

2. Create and activate a conda environment:

   conda env create -f environment.yaml
   conda activate tqdne

   If conda is not installed, download it from Miniconda.

3. (Optional) If you prefer to install the environment in a custom path, run:

   conda env create -f environment.yaml -p <PATH>
   conda activate <PATH>

   Replace <PATH> with your desired installation directory.

Weight files

You can find the weight files for the neural networks on Zenodo, under the weights folder.

Experiments

To reproduce the experiments from the manuscript, including data preprocessing, training, and evaluation, navigate to the experiments folder. Refer to the corresponding README files for step-by-step guidance.

Sampling waveforms

You can generate your own waveforms using the scripts in scripts. See the corresponding README files for more information.

Acknowledgements

Some Python code has been adapted from the following repositories:

• EDM
• Consistency Models