rg-dataset contains data for training and testing ML-based source finder applications. This dataset contains images (FITS format), regions (DS9 format) and masks (FITS format) for these classes of radio objects:
- Galaxies
- Sidelobes (or artefacts)
- Point-sources
NB: These data include private observatory data that are to be kept private and not shared to the public unless explicitly agreed before.
This repository contains only the Data Version Control (DVC) files tracking the image data. Actual data are instead kept in a remote gDrive shared folder. To get access to the data and work on them do the following:
-
Clone the repo:
git clone https://github.com/SKA-INAF/rg-dataset.git -
Download remote data:
dvc pull -
Add or modify a file (e.g. img.fits) and commit the change:
dvc add img.fits(generate file .dvc)
git add .gitignore img.fits.dvc
git commit -m ‘Added file’
git push origin master
dvc push(upload new/updated data in the remote storage)