# Keras

> neural network library

**Wikidata**: [Q28470421](https://www.wikidata.org/wiki/Q28470421)  
**Wikipedia**: [English](https://en.wikipedia.org/wiki/Keras)  
**Source**: https://4ort.xyz/entity/keras

## Summary

Keras is a software application . It is classified specifically as a type of software .

## FAQs
### Q: Who created Keras?
A: Keras was created by François Chollet, a French machine learning researcher and software developer. He launched the project to facilitate fast experimentation with deep neural networks.

### Q: What programming language is Keras written in?
A: Keras is written in Python, a general-purpose programming language. It relies on Python libraries such as NumPy and SciPy for its backend operations.

### Q: What license is Keras released under?
A: Keras is released under a dual-license structure. Users can utilize the software under either the MIT License or the Apache Software License 2.0.

## Why It Matters
Keras plays a significant role in the field of machine learning and software development by lowering the barrier to entry for working with artificial neural networks. As an open-source library written in Python, it provides a high-level interface that allows developers and researchers to design, configure, and train complex deep learning models with minimal code. This accessibility facilitates rapid prototyping and experimentation, which is essential for advancing research and applying deep learning solutions across various industries.

The library's structure supports modularity and extensibility, allowing users to define and connect neural network layers, optimizers, and activation functions with ease. By standardizing the implementation of common neural network components—such as those found in Deep Residual Learning (ResNet) and MobileNets—Keras ensures that state-of-the-art architectures are reproducible and accessible. Its widespread adoption is evidenced by its extensive language support on Wikipedia (including English, French, and German) and its integration into various software repositories, such as PyPI, Arch User Repository, and Gentoo packages. The inclusion of standard datasets like CIFAR-10 and MNIST directly within the library further streamlines the workflow for training and testing models, making it a foundational tool in modern computational science.

## Notable For
- **High-Level API**: Provides a user-friendly interface for building and training neural networks compared to lower-level libraries.
- **Modularity**: Comprises configurable modules (layers, optimizers, activation functions) that can be combined freely.
- **Dataset Integration**: Includes direct access to widely-used machine learning datasets such as CIFAR-10, CIFAR-100, and MNIST.
- **Implementation of Research**: Integrates implementations of significant academic works, such as "Deep Residual Learning for Image Recognition" and "MobileNets."
- **Wide Distribution**: Available across multiple package management systems, including PyPI, AUR, FreeBSD ports, and openSUSE.

## Body

### Development and Origin
Keras was developed by François Chollet, a French machine learning researcher and computer scientist born on October 20, 1989. The project is identified as a "software" instance specifically designed for the use of "artificial neural network" creation.

### Technical Specifications
The library is implemented in Python, a general-purpose programming language that was conceived in 1991. To function, Keras depends on several external software libraries:
- NumPy
- SciPy
- PyYAML
- h5py

The project is dual-licensed, allowing for use under either the MIT License or the Apache Software License 2.0.

### Repository and Availability
The source code for Keras is maintained publicly on GitHub under the username `keras-team` at the repository URL `https://github.com/keras-team/keras`. The official project website is hosted at `https://keras.io/`.

Keras is widely available across various operating system repositories and package managers:
- **Python Package Index (PyPI)**: `Keras`
- **Arch User Repository (AUR)**: `python-keras`, `python2-keras-git`, etc.
- **Gentoo**: `sci-libs/keras`
- **FreeBSD**: `math/py-keras`
- **openSUSE**: `python-Keras`

### Features and Components
Keras offers pre-built modules that cite and implement findings from prominent research papers. Notable cited works include:
- "Self-Normalizing Neural Networks"
- "Deep Residual Learning for Image Recognition"
- "MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications"
- "A Theoretically Grounded Application of Dropout in Recurrent Neural Networks"

The library also includes utility functions to load standard benchmark datasets, specifically:
- CIFAR-10
- CIFAR-100
- MNIST database

### Version History
Historical data indicates a frequent release cycle for stable versions in 2017 and 2018. Specific recorded releases include:
- **Version 2.1.5**: Released March 6, 2018
- **Version 2.1.4**: Released February 13, 2018
- **Version 2.1.3**: Released January 16, 2018
- **Version 2.1.2**: Released December 1, 2017
- **Version 2.1.1**: Released November 14, 2017
- **Version 2.1.0**: Released November 13, 2017
- **Version 2.0.9**: Released November 1, 2017
- **Version 2.0.8**: Released August 25, 2017
- **Version 2.0.7**: Released August 21, 2017
- **Version 2.0.6**: Released July 7, 2017

## References

1. [The keras Open Source Project on Open Hub: Licenses Page. Open Hub](https://www.openhub.net/p/keras/licenses)
2. [Source](https://github.com/keras-team/keras/blob/master/LICENSE)
3. [The keras Open Source Project on Open Hub: Languages Page. Open Hub](https://www.openhub.net/p/keras/analyses/latest/languages_summary)
4. [Release 2.1.5. 2018](https://github.com/keras-team/keras/releases/tag/2.1.5)
5. [Release 2.1.4. 2018](https://github.com/keras-team/keras/releases/tag/2.1.4)
6. [Release 2.1.3. 2018](https://github.com/keras-team/keras/releases/tag/2.1.3)
7. [Release 2.1.2. 2017](https://github.com/keras-team/keras/releases/tag/2.1.2)
8. [Release 2.1.1. 2017](https://github.com/keras-team/keras/releases/tag/2.1.1)
9. [Release 2.1.0. 2017](https://github.com/keras-team/keras/releases/tag/2.1.0)
10. [Release 2.0.9. 2017](https://github.com/keras-team/keras/releases/tag/2.0.9)
11. [Release 2.0.8. 2017](https://github.com/keras-team/keras/releases/tag/2.0.8)
12. [Release 2.0.7. 2017](https://github.com/keras-team/keras/releases/tag/2.0.7)
13. [Release 2.0.6. 2017](https://github.com/keras-team/keras/releases/tag/2.0.6)
14. [Release 2.0.5. 2017](https://github.com/keras-team/keras/releases/tag/2.0.5)
15. [Release 2.0.4. 2017](https://github.com/keras-team/keras/releases/tag/2.0.4)
16. [Release 2.0.0. 2017](https://github.com/keras-team/keras/releases/tag/2.0.0)
17. [Release 2.1.6. 2018](https://github.com/keras-team/keras/releases/tag/2.1.6)
18. [Release 2.2.0. 2018](https://github.com/keras-team/keras/releases/tag/2.2.0)
19. [Release 2.2.1. 2018](https://github.com/keras-team/keras/releases/tag/2.2.1)
20. [Release 2.2.2. 2018](https://github.com/keras-team/keras/releases/tag/2.2.2)
21. [Release 2.2.3. 2018](https://github.com/keras-team/keras/releases/tag/2.2.3)
22. [Release 2.2.4. 2018](https://github.com/keras-team/keras/releases/tag/2.2.4)
23. [Release 2.2.5. 2019](https://github.com/keras-team/keras/releases/tag/2.2.5)
24. [Release 2.3.0. 2019](https://github.com/keras-team/keras/releases/tag/2.3.0)
25. [Release 2.3.1. 2019](https://github.com/keras-team/keras/releases/tag/2.3.1)
26. [Release 2.4.0. 2020](https://github.com/keras-team/keras/releases/tag/2.4.0)
27. [Release 2.6.0. 2021](https://github.com/keras-team/keras/releases/tag/v2.6.0)
28. [Release 2.7.0. 2021](https://github.com/keras-team/keras/releases/tag/v2.7.0)
29. [Release 2.8.0. 2022](https://github.com/keras-team/keras/releases/tag/v2.8.0)
30. [Release 2.10.0. 2022](https://github.com/keras-team/keras/releases/tag/v2.10.0)
31. [Release 2.11.0. 2022](https://github.com/keras-team/keras/releases/tag/v2.11.0)
32. [Release 2.12.0. 2023](https://github.com/keras-team/keras/releases/tag/v2.12.0)
33. [Release 2.13.1. 2023](https://github.com/keras-team/keras/releases/tag/v2.13.1)
34. [Release 2.14.0. 2023](https://github.com/keras-team/keras/releases/tag/v2.14.0)
35. [Release 2.15.0. 2023](https://github.com/keras-team/keras/releases/tag/v2.15.0)
36. [Release 3.0.0. 2023](https://github.com/keras-team/keras/releases/tag/v3.0.0)
37. [Release 3.0.1. 2023](https://github.com/keras-team/keras/releases/tag/v3.0.1)
38. [Release 3.0.2. 2023](https://github.com/keras-team/keras/releases/tag/v3.0.2)
39. [Release 3.0.3. 2024](https://github.com/keras-team/keras/releases/tag/v3.0.3)
40. [Release 3.0.4. 2024](https://github.com/keras-team/keras/releases/tag/v3.0.4)
41. [Release 3.0.5. 2024](https://github.com/keras-team/keras/releases/tag/v3.0.5)
42. [Release 3.1.0. 2024](https://github.com/keras-team/keras/releases/tag/v3.1.0)
43. [Release 3.1.1. 2024](https://github.com/keras-team/keras/releases/tag/v3.1.1)
44. [Release 3.2.0. 2024](https://github.com/keras-team/keras/releases/tag/v3.2.0)
45. [Release 3.2.1. 2024](https://github.com/keras-team/keras/releases/tag/v3.2.1)
46. [Release 3.3.0. 2024](https://github.com/keras-team/keras/releases/tag/v3.3.0)
47. [Release 3.3.1. 2024](https://github.com/keras-team/keras/releases/tag/v3.3.1)
48. [Release 3.3.2. 2024](https://github.com/keras-team/keras/releases/tag/v3.3.2)
49. [Release 3.3.3. 2024](https://github.com/keras-team/keras/releases/tag/v3.3.3)
50. [Release 3.4.0. 2024](https://github.com/keras-team/keras/releases/tag/v3.4.0)