Keras is a machine-learning framework written in Python. It’s more like a meta-framework, it wraps multiple back-ends like PyTorch, TensorFlow, and JAX.
Picking a backend
The default backend of Keras is Tensorflow. This can be changed using the KERAS_BACKEND environment variable.
import os
os.environ["KERAS_BACKEND"] = "torch"
import keras
This snippet will use the torch backend. When I’m working on ML projects, I tend to use the torch backend, but I’ve heard good things about the JAX backend too.
Citation
If you find this work useful, please cite it as:
@article{yaltirakli,
title = "Keras",
author = "Yaltirakli, Gokberk",
journal = "gkbrk.com",
year = "2025",
url = "https://www.gkbrk.com/keras"
} Not using BibTeX? Click here for more citation styles.
IEEE Citation Gokberk Yaltirakli, "Keras", August, 2025. [Online]. Available: https://www.gkbrk.com/keras. [Accessed Aug. 16, 2025].
APA Style Yaltirakli, G. (2025, August 16). Keras. https://www.gkbrk.com/keras
Bluebook Style Gokberk Yaltirakli, Keras, GKBRK.COM (Aug. 16, 2025), https://www.gkbrk.com/keras
