Metadata-Version: 2.1
Name: vacuum-cleaner
Version: 0.3
Summary: Deep Vacuum Cleaner
Home-page: https://github.com/gijzelaerr/vacuum-cleaner
Author: Gijs Molenaar
Author-email: gijs@pythonic.nl
License: MIT
Keywords: radio astronomy deep learning cleaning science
Platform: UNKNOWN
Classifier: License :: OSI Approved :: MIT License
Classifier: Development Status :: 3 - Alpha
Classifier: Programming Language :: Python
Classifier: Programming Language :: Python :: 3.6
Classifier: Programming Language :: Python :: 3.7
Classifier: Intended Audience :: Science/Research
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Classifier: Topic :: Scientific/Engineering :: Astronomy
Description-Content-Type: text/markdown
Provides-Extra: gpu
Requires-Dist: tensorflow
Requires-Dist: astropy
Requires-Dist: scipy
Requires-Dist: pathlib
Requires-Dist: future
Requires-Dist: six
Requires-Dist: typing
Requires-Dist: repoze.lru
Provides-Extra: gpu
Requires-Dist: tensorflow-gpu; extra == 'gpu'

# Deep Vacuum Cleaner

Radio telescope deconvolution based using a Conditional Generative Adversarial Deep Network.

Based on [pix2pix-tensorflow](https://github.com/affinelayer/pix2pix-tensorflow)

Whch is based on [pix2pix](https://phillipi.github.io/pix2pix/) by Isola et al.

[Article about this implemention](https://affinelayer.com/pix2pix/)


## preparations

You probably want to download a pretrained model.

download:

http://repo.kernsuite.info/vacuum/model.tar.xz

And extract to `share/vacuum/model`.


## Setup

```
$ pip install vacuum-cleaner

```

or if you want to try the GPU accelerated version:

```
$ pip install "vacuum-cleaner[gpu]"

```
But the tensorflow-gpu package is not the most portable package available.

## Usage

```
$ vacuum-clean dirty.fits psf.fits
```       



## Training

Have a look at `vacuum-train --help` or at the source. Intended to be trained with
[spiel](https://github.com/gijzelaerr/spiel/) as training data generator.




