photcalib 0.0

calibration tool for photometric data

PhotCalib

Calibration tool for Prisitine CaHK data. For more details see Martin, Starkenburg & Yuan 2023

https://github.com/zyuan-astro/PhotCalib/assets/20896020/07018137-40fb-4e04-b96f-2ba3f767fade

https://github.com/zyuan-astro/PhotCalib/assets/20896020/26806571-87b4-4624-b44c-068567b22d61

Installation

Before installtion, make sure you have pytorch or install essential dependencies :

pip install -r requirements.txt

• Installation from PyPI

pip install photcalib

• Manual installation

download the repository and run the setup

python setup.py install

Getting started

cd examples

To Calibrate CaHK data using the run model trained by the main survey: .. code:

python calib_raw.py 17Am05 combined_catalogue_17Am05_sel.fits -D device

Substitute 17Am05 to the run you need and make sure the input data for calibration has the same format as that in the examples/data/.fits

• D (str) is the device (default:cpu, mps, gpu) to run pytorch. mps is not faster than cpu on M1. If gpu is chosen, it will use the first graphic card cuda:0 by default.

To train calibration run model: .. code:

python calib_mod.py 17Am05

Substitute 17Am05 to the run for calibration with the input file as examples/data/inputs_run.npy

Specify the device and training parameters (the followings are the default settings) .. code:

python calib_mod.py 17Am05 -D cpu -lr 1e-6 -n 400 -mom 0.9 -thr 1e-2

• D (str) is the device (default:cpu, mps, gpu) to run pytorch.

• n (int) is the number of training epochs.

• lr (float) is the initial learning rate, which will decrease using scheduler (see details at https://pytorch.org/docs/stable/generated/torch.optim.lr_scheduler.ReduceLROnPlateau).

• thr (float) is the threshold of the scheduler.

• mom (float) is the momentum of gradient descent.

Citing this work

coming soon