Skip to content

Latest commit

 

History

History
83 lines (53 loc) · 2.38 KB

README.md

File metadata and controls

83 lines (53 loc) · 2.38 KB

Pix2Prof: deep learning for the extraction of useful sequential information from galaxy imagery

This repository is the official implementation of Pix2Prof.

The model as trained produces a surface brightness profile from an unprocessed galaxy image from the SDSS in either the g, r, or i bands. With a throughput speed of ~ 1 galaxy/second on an Intel Xeon E5-2650 v3 CPU, Pix2Prof improves on the speed of the method it approximates by over two orders of magnitude!

Also, given suitable training data, Pix2Prof can be retrained to produce any galaxy profile from any galaxy image.

Requirements

To install requirements:

pip install -r requirements.txt

Training

To train Pix2Prof, run this command:

python train.py --gals ./gals.txt

Where gals.txt is a textfile containing a list of galaxy names.

Evaluation

To evaluate Pix2Prof, run:

python eval.py ./gals.txt <checkpoint_location>

Where gals.txt is a textfile listing a list of galaxy names.

Pre-trained Models

You can download the pretrained model used in the paper here:

SDSS Pix2Prof

If you want to use the pretrained model to infer your dataset, download the checkpoint to your working directory and run:

python eval.py ./gals.txt pix2prof_2020-10-08.pth

Where gals.txt is a textfile listing a list of galaxy names.

Results

Here is a random selection of predictions from the test set:

Here is the summary statistics plot from the paper:

Contributing

Pix2Prof is free software: you can redistribute it and/or modify it under the terms of the GNU Affero General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Pix2Prof is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Affero General Public License for more details.

You should have received a copy of the GNU Affero General Public License along with Pix2Prof. If not, see https://www.gnu.org/licenses/.