Deep Learning Frameworks logopeterindia.net
Knowledge Hub

Deep Learning Frameworks

A directory of frameworks for building and training neural networks — mainstream frameworks, specialised and research libraries, distributed training, inference runtimes and archived projects.

24 topics•6 categories•Updated October 2026

Mainstream Frameworks

7 entries

The frameworks most used in research and production.

01PyTorchAn open-source machine learning framework that accelerates the path from research prototyping to production deployment, backed by Meta AI. 02TensorFlowAn ecosystem of tools, libraries, and community resources that lets developers easily build and deploy ML-powered applications at scale. 03KerasExascale machine learning built on TensorFlow 2 — an industry-strength framework that scales to large clusters of GPUs or an entire TPU pod. 04JAXComposable transformations of Python and NumPy programs for high-performance machine learning. 05PaddlePaddlePArallel Distributed Deep LEarning — an industrial-practice machine learning framework from Baidu supporting large-scale distributed training. 06MATLAB for Deep LearningComplete tooling for data preparation, network design, simulation, and deployment of deep neural networks — integrated within the MATLAB environment. 07PyTorch LightningHigh-level PyTorch wrapper that organises training code.

Specialised & Research Libraries

4 entries

Libraries with a particular focus or language.

08Eclipse Deeplearning4jThe first commercial-grade, open-source, distributed deep-learning library written for Java and Scala — integrated with Hadoop and Apache Spark. 09ChainerA powerful, flexible, and intuitive framework for neural networks — pioneering the define-by-run approach for dynamic computation graphs. 10CaffeA deep learning framework made with expression, speed, and modularity in mind — developed at Berkeley Vision and Learning Center. 11PyTensorLibrary for defining and optimising tensor expressions; successor to the Aesara and Theano line.

Distributed Training

1 entries

Scaling training across clusters.

12TensorFlowOnSparkBrings scalable deep learning to Apache Hadoop and Apache Spark clusters — enabling TensorFlow model training within existing big data pipelines.

Inference & Deployment Runtimes

3 entries

Running trained models on devices and servers.

13ONNX RuntimeCross-platform engine for running ONNX models. 14TensorFlow.jsTrain and run models in the browser or Node.js. 15MLXApple's array framework for machine learning on Apple silicon.

Archived & Legacy Projects

3 entries

Projects that are archived, retired or no longer current; listed for reference.

16Apache MXNet (retired to the Apache Attic)A truly open-source deep learning framework suited for flexible research prototyping and scalable production deployments across multiple GPUs. 17Aesara (repository archived)A Python library for defining, optimizing, and efficiently evaluating mathematical expressions involving multi-dimensional arrays, built for deep learning research. 18DeepLearningKit (2015 project)An open-source deep learning framework designed specifically for Apple iOS and OS X — enabling on-device neural network inference on Apple hardware.

Related Pages

6 entries

Related directories on this site.

19Deep LearningDeep learning hub. 20GNN FrameworksFrameworks for graph neural networks. 21Python Frameworks for AIPython frameworks for AI. 22Machine Learning FrameworksMachine learning frameworks. 23Data Sets for ML/DLDatasets for training. 24Machine LearningMachine learning hub.

Frequently Asked Questions

What is a deep learning framework?

A software library that provides tensors, automatic differentiation, neural network layers and training tools so you can define and train models efficiently on CPUs and GPUs.

PyTorch or TensorFlow?

PyTorch dominates research and is widely used in production; TensorFlow and Keras remain common, especially for mobile and browser deployment. JAX suits high-performance research.

What happened to Apache MXNet?

Apache MXNet has been retired to the Apache Attic and its repository is archived, so it is listed for reference only.

What is an inference runtime?

Software such as ONNX Runtime that runs trained models efficiently across hardware, often converting from a training framework.

Where are graph neural network libraries?

See the GNN Frameworks page.