https://github.com/jesperdramsch/complex-cnn-seismic

Code accompanying the paper "Complex-valued neural networks for machine learning on non-stationary physical data".
https://github.com/jesperdramsch/complex-cnn-seismic

complex complex-numbers neural-network seismic

Last synced: about 2 months ago
JSON representation

Code accompanying the paper "Complex-valued neural networks for machine learning on non-stationary physical data".

• Host: GitHub
• URL: https://github.com/jesperdramsch/complex-cnn-seismic
• Owner: JesperDramsch
• License: apache-2.0
• Created: 2019-12-05T14:14:24.000Z (about 5 years ago)
• Default Branch: master
• Last Pushed: 2020-11-30T22:31:20.000Z (about 4 years ago)
• Last Synced: 2024-06-11T17:28:52.931Z (7 months ago)
• Topics: complex, complex-numbers, neural-network, seismic
• Language: Python
• Homepage: https://www.sciencedirect.com/science/article/pii/S0098300420306208
• Size: 3.74 MB
• Stars: 27
• Watchers: 6
• Forks: 11
• Open Issues: 3
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Complex-CNN-Seismic

This repository reproduces "Complex-valued neural networks for machine learning on non-stationary physical data".

## Data

Obtained from https://github.com/olivesgatech/facies_classification_benchmark via

```

# download the files: 

wget https://zenodo.org/record/3755060/files/data.zip

# check that the md5 checksum matches: 

openssl dgst -md5 data.zip # Make sure the result looks like this: MD5(data.zip)= bc5932279831a95c0b244fd765376d85, otherwise the downloaded data.zip is corrupted. 

```

Preparation for training via `src/data_prep.py`.

## Training

Training done on GPU cluster using `src/mass_train.py`.

## Prediction

Use trained models to generate predictions `src/save_predictions.py`.

## Analysis

Numerical and qualitative analysis generated via `src/explore.py`.

## Citation

Please cite the according paper as

```

@article{dramsch2020complex,

  title={Complex-valued neural networks for machine learning on non-stationary physical data},

  author={Dramsch, Jesper S{\"o}ren and L{\"u}thje, Mikael and Christensen, Anders Nymark},

  journal={Computers \& Geosciences},

  pages={104643},

  year={2020},

  publisher={Elsevier}

}

```