Home > Categories > Machine learning and neural networks > DL Frameworks > Keras

Keras

Rating:

4.8 / 5.0

Tags

Machine Learning, Deep Learning, Neural Networks, Open Source, Library, AI Framework, Python

Categories:
Machine learning and neural networks
Creator François Chollet
Date 2015-03-01
Platforms Python
Status Active
Website keras.io
Price Model Free (library)
Sections:
DL FrameworksModel Training

Pricing Details

Free (open source).

Features

High-level API, modularity, fast prototyping, support for various network types (CNNs, RNNs), runs on CPU/GPU/TPU, integrates with backend frameworks.

Integrations

Integration with TensorFlow (primary backend), PyTorch (via Keras 3), JAX (via Keras 3), and other data science libraries (NumPy, Pandas, Scikit-learn).

Preview

Keras is a high-level API for building and training neural networks, written in Python. It is designed for fast experimentation and is built on top of lower-level deep learning frameworks like TensorFlow, Theano, or CNTK (though primarily used with TensorFlow now). Keras focuses on being modular, extensible, and easy to use. It provides a simple and consistent interface for building various types of neural networks, making it popular for researchers and developers. It's widely used for rapid prototyping and building production-ready models.