Science Score: 67.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
✓CITATION.cff file
Found CITATION.cff file -
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
✓DOI references
Found 1 DOI reference(s) in README -
✓Academic publication links
Links to: zenodo.org -
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (8.5%) to scientific vocabulary
Repository
codes for SKA sdc3a data challenge
Basic Info
- Host: GitHub
- Owner: CEA-jiangming
- License: gpl-3.0
- Language: Python
- Default Branch: main
- Size: 39.1 KB
Statistics
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
- Releases: 1
Metadata Files
README.md
Hausos-sdc3a
This is the repository containing codes for SKA sdc3a data challenge, our team is named Hausos, which means The Original Dawn Goddess, a decent name for the detection of EoR signal of the Universe. For more details about the data challenge, please refer to the website (https://sdc3.skao.int/overview)
Contents
Introduction
The method is a deep learning method for the EoR estimation, which is based on the transformer model. The author admits that the method is not optimized at the current stage due to time limitations. The repository includes preprocessing, methodology, postprocessing, and evaluation, one can execute them separately. preprocessing: - cutImage.py: file used to cut testing images into patches - cutImageSDC3.py: file used to cut training images into patches methodology: - main.py: script to run the neural network model, one can choose 'train', 'test' or 'all' mode for the training, testing or training and testing. - myDatasets.py: setup for the training and testing datasets - utils.py: some utilities - myModels/setr: neural network model, which is based on transformer. postprocessing: - combineImage.py: combine predicted patches into images for the final evaluation evaluation: - ps2d.py: function to calculate the 2d power spectrum, which is called in ps2dscript.py - ps2d_script.py: script to evaluate the 2d power spectrum
Exectution
4 steps to produce the final 2d power spectrum 1. prepare the data using cutImageSDC3.py to produce format-correct training patches, and cut the test data into patches using cutImage.py to feed the neural network model 1. train the neural network by seting 'train' in main.py, and 'test' to predict the EoR patches after training 1. combine the patches into images by using combineImage.py 1. evaluate the 2d power spectrum by using ps2dscript.py. Note that the format of power spectrum is saved according to the requirements of SDC3a data challenge.
Authors
On behalf of Hausos team, one can contact the team leader Dr. Ming Jiang via * Ming Jiang mingjiang@xidian.edu.cn
Owner
- Name: Ming Jiang
- Login: CEA-jiangming
- Kind: user
- Location: Xi'an, China
- Company: Xidian University
- Repositories: 1
- Profile: https://github.com/CEA-jiangming
Citation (CITATION.cff)
@misc{Hausos_sdc3a,
author = {M. Jiang, J.-L. Starck, H. Shan},
title = {Hausos SDC3a software},
year = {2023},
publisher = {Zenodo},
journal = {Zenodo},
doi = {10.5281/zenodo.10230697},
howpublished = {\url{https://zenodo.org/records/10230697}}
}