Neural networks for feature extraction for gravitational waves.
Project description
dingo-enets
This reposity contains the embedding networks from [1], which are trained for the purpose of gravitational wave parameter estimation. If you find this code useful please cite [1].
Note: This is only a partial release of the code used in [1]. A more comprehensive package will be released in the near future. The present repository will not be maintained once the full package is publicly available.
Usage
Build a virtual environment and install dingo-enets
.
$ python3 -m venv venv
$ source venv/bin/activate
$ pip install dingo-enets
Use dingo-enets
to download and build a trained embedding network for a particular observing run. The model will be saved in </path/to/model_directory>
.
> from dingo_enets import build_enet
> enet = build_enet(run="O1", detectors=["H1", "L1"], model_dir="</path/to/model_directory>")
> import torch
> input = torch.rand(10, 2, 3, 8033)
> output = enet(input)
> print(output.shape)
The function build_enet
recognises whether a suitable model is present in the model directory, in which case it is not downloaded but instead loaded directly from disk.
References
[1] M. Dax, S.R. Green, J. Gair, J.H. Macke, A. Buonanno, B. Schölkops, Real-Time Gravitational Wave Science with Neural Posterior Estimation, Phys.Rev.Lett. 127 (2021) 24, 241103. [arXiv] [inspirehep]
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for dingo_enets-0.1.0-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | ddd548ceaa5f073bf511e4426450906cdbdd5175703ef107ef1548b84e844616 |
|
MD5 | 810ceb81c5c836042419236c7cc67317 |
|
BLAKE2b-256 | e3c1ba269c91ad3815b54b51cf24b67486b377c44098d864085040a4d7799c08 |