This repository contains code and data relevant to the Mathews et al. (2023) article submitted to the Astrophysical Journal.
The /data directory contains the emulator used in the paper in addition and documentation for reading the emulator file. In addition, the /src directory contains code that allows one to run Prospector fits using the emulator in place of FSPS. Finally, the /notebooks directory contains a notebook demonstrating a sample fit to a mock galaxy using the emulator and Prospector.
You will need to have Git Large File Storage installed on your system in order to properly clone this repository, or else your downloads of the emulator files may be corrupted. To check if the emulator files are correctly installed, use the CHECKSUM included in the /data directory to verify the integrity of the files (i.e. sha256sum -c CHECKSUM.asc).
The code in this repository does not (currently) contain code for generating emulator training sets and training new emulators. To train a new emulator, consider using:
- Speculator: A Python library (an extension of the Tensorflow machine learning library) containing code useful for training spectral population synthesis emulators. See Alsing et al. (2020).
- Parrot.jl: A Julia library (an extension of the Flux.jl machine learning library) containing code useful for training spectral population synthesis emulators, including the activation function used in Speculator. This is the library used in the paper.