Metadata-Version: 2.1
Name: deepspace
Version: 0.0.5
Summary: A collection of search space for the DeepHyper package.
Home-page: https://github.com/deephyper/deepspace
Author: Prasanna Balaprakash <pbalapra@anl.gov>, Romain Egele <romain.egele@polytechnique.edu>, Romit Maulik <rmaulik@anl.gov>
License: ANL
Project-URL: Source, https://github.com/deephyper/deepspace
Project-URL: Tracker, https://github.com/deephyper/deepspace/issues
Platform: UNKNOWN
Classifier: Development Status :: 3 - Alpha
Classifier: Programming Language :: Python
Classifier: Programming Language :: Python :: 3
Requires-Python: >=3.6, <3.9
Description-Content-Type: text/markdown
Provides-Extra: tests
Provides-Extra: dev
Provides-Extra: docs
License-File: LICENSE


# DeepSpace

A collection of search space for the DeepHyper package.

## Requirements

Graphviz.

## Quick Start

Generate a neural architecture space for fully connected networks with residual connections:

```python
from deepspace.tabular import DenseSkipCoSpace

def create_search_space(input_shape=(54,), output_shape=(7,), **kwargs)
    return DenseSkipCoSpace()(input_shape, output_shape, num_layers=10, dropout=0.0)
```

Generate a neural architecture space for AutoEncoder guided by an estimator:

```python
from deepspace.tabular import SupervisedRegAutoEncoderSpace

factory = SupervisedRegAutoEncoderSpace()(
     input_shape=(100,), output_shape=[(100), (10,)]
)
```


