Horovod is an open-source framework for distributed training of deep learning models across many GPUs or machines. Originally developed at Uber, it supports TensorFlow, Keras, PyTorch and MXNet and synchronises the gradients computed by the participating machines using ring-allreduce, without a central parameter server. The project is hosted by the LF AI & Data Foundation; the latest stable release, 0.28.1, dates from 2023 (as of October 2026).
TensorFlow is an open-source machine learning framework developed by Google and released in 2015. It is used to build, train and deploy neural networks in applications, on servers or on mobile devices. Keras provides an accessible high-level interface; typical uses include image classification, forecasting and recommendation systems.
PyTorch is an open-source framework for machine learning and deep learning in the Python programming language. It was originally developed at Meta (then Facebook) and has been stewarded since 2022 by the PyTorch Foundation under the umbrella of the Linux Foundation. PyTorch is widely used in research and industry because it lets developers build models flexibly, train them on GPUs and test them step by step; many well-known language and image models are built with it.
Keras is an open-source Python library for building and training neural networks, originally developed by François Chollet. It offers an accessible, high-level interface for assembling models from layers. Since version 3, Keras can run on JAX, TensorFlow or PyTorch as its computational backend. Example: a team builds a prototype image classification model in just a few lines of code.
Kubeflow is an open-source platform for running machine learning workflows on Kubernetes, from data preparation through training and hyperparameter search to model serving. Originally initiated by Google, the project belongs to the Cloud Native Computing Foundation, where it reached “graduated” status in August 2026 (as of October 2026). It is aimed mainly at teams that train their own models at scale on their own or cloud infrastructure.
