Keras is an open-source software library that provides a Python interface for artificial neural networks. It acts as an interface for the TensorFlow library. Keras was first released in 2015 and was designed to enable fast experimentation with deep neural networks, focusing on being user-friendly, modular, and extensible. The name Keras derives from the Ancient Greek word κέρας (keras), meaning 'horn'.
History
Keras was developed as part of the research effort of project ONEIROS (Open-ended Neuro-Electronic Intelligent Robot Operating System). Its primary author and maintainer is François Chollet, a Google engineer until 2024. Initially, Keras was independent software, but it was later integrated into the TensorFlow library. Up until version 2.3, Keras supported multiple backends, including TensorFlow, Microsoft Cognitive Toolkit, Theano, and PlaidML. From version 2.4 up until version 3.0, only TensorFlow was supported. Starting with version 3.0 (and its preview version, Keras Core), Keras became multi-backend again, supporting TensorFlow, JAX, and PyTorch. Keras 3 became the default version for TensorFlow 2.16 onwards, but Keras 2 can still be used.
Features
Keras contains numerous implementations of commonly used neural-network building blocks, such as layers, objectives, activation functions, optimizers, and a host of tools for working with image and text data. This simplifies programming for deep neural networks. In addition to standard neural networks, Keras supports convolutional and recurrent neural networks. It also supports other common utility layers like dropout, batch normalization, and pooling. Keras allows users to produce deep models on smartphones (iOS and Android), on the web, or on the Java Virtual Machine. It also allows the use of distributed training of deep-learning models on clusters of graphics processing units (GPU) and tensor processing units (TPU).
Adoption and Impact
Keras has become one of the most widely used high-level neural network APIs, particularly in the Machine learning community. Its integration with TensorFlow has made it a standard tool for many Deep learning applications. The library is known for its ease of use, which has contributed to its popularity among both beginners and researchers. Keras is also used in various Artificial intelligence projects, including those involving Generative AI and Large language models. The library's flexibility and support for multiple backends have made it a versatile choice for developing and deploying neural networks.
See Also
- Comparison of deep-learning software
- Neural network
- TensorFlow
- PyTorch
- JAX
References
- Keras documentation and official website
- TensorFlow release notes
- François Chollet's publications and talks