Spatial models derived from spatially constrained calibration are automatically displayed using .ppm files. See for example:
Home | Download | Code | Tutorial
SAGECal is a fast, distributed and GPU accelerated radio astronomial calibration package. The many optimization algorithms in SAGECal are implemented in a computationally efficient way and can be used in many other applications.
What can you do with SAGECal? Here is a list of publications that have used SAGECal to produce scientific results. Have a look to see if SAGECal can help you.
For more information, visit program description.
Tue 14 May, 2024
SAGECal version 0.8.3 is released. Support for lunar coordinate transforms using the CSPICE library is added. Large scale diffuse sky models can be used in calibration using large shapelet models.
Tue 12 Sep, 2023
Spatial models derived from spatially constrained calibration are automatically displayed using .ppm files. See for example:
Tue 27 Dec, 2022
SAGECal version 0.7.9 is released. CPU version can be significantly sped-up by using vectorized math operations provided by libmvec. See installation instructions.
Wed 15 Dec, 2021
Spatially constraind calibration: paper published!
Sun 03 Oct, 2021
SAGECal version 0.7.6 is released, with spatial regularization in distributed, direction dependent calibration. Also see this document.
Thu 12 Aug, 2021
SAGECal version 0.7.5 is released, with support for LOFAR dipole beam model.
Tue 15 June, 2021
SAGECal becomes the fist and only calibration suite with AI support! Read the paper here. This is also the first ever use of reinforcement learning in radio astronomy.
Wed 04 Mar, 2020
The paper describing distributed stochastic calibration has been accepted.
Wed 18 Dec, 2019
Here is a simple demo of radio interferometric calibration with PyTorch. The main intention is to compare the performance of various optimizers of PyTorch (LBFGS, Adam, etc.) in calibration.
Tue 12 Nov, 2019
SAGECal version 0.7.0 is released! Major addition is stochastic calibration. With stochastic calibration, it is possible to work with very large volumes of data, without running out of memory. Direct applications are bandpass calibration and RFI mitigation with much better accuracy than ever before.
Tue 13 Aug, 2019
A new use of our improved LBFGS in deep learning: Federated learning by federated averaging and consensus optimization. Here is the demo code.
Mon 15 Jul, 2019
SAGECal runs on the NVIDIA Jetson Nano embedded system! Here is how we do it.
Wed 12 Jun, 2019
SAGECal 0.6.1 is released! With bug fixes, better build support and speed improvements! Also included is libdirac.a, a nonlinear optimization library for general use, especially the stochastic LBFGS algorithm. See ./test/Dirac/ directory for an example.
Wed 12 Dec, 2018
A short introduction to using our improved LBFGS in PyTorch.
Tue 24 Apr, 2018
Moved the repository to GitHub. A mirror is always maintained at the usual place (but might be out of date).
Thu 15 Dec, 2016
A major boost for SAGECal! We just got funded from the Netherlands eScience Center to do a major overhaul and improve user friendliness. This is under the Accelerating Scienctific Discovery (ASDI) project calls. The DIRAC (Distributed Radio Astronomical Computing) project starts in 2017.
Tue 25 Oct, 2016
A script to create a cluster file from a sky model is added.
Wed 19 Oct, 2016
Added the Tutorial.
Wed 06 Jul, 2016
New web page.