The following tables compare notable software frameworks, libraries, and computer programs for deep learning applications.
Deep learning software by name
Software Creator Initial release Software license Platform Written in Interface OpenMP support OpenCL support CUDA support Automatic differentiation Has pretrained models BigDLJason Dai (Intel)2016Apache SparkScalaScala, Python Caffe Berkeley Vision and Learning Center 2013 Linux, macOS, Windows C++ Python, MATLAB, C++ Chainer Preferred Networks 2015 Linux, macOS Python Python Deeplearning4j Skymind engineering team; Deeplearning4j community; originally Adam Gibson 2014 Linux, macOS, Windows, Android (Cross-platform) C++, Java Java, Scala, Clojure, Python (Keras), Kotlin Deep Learning Studio Deep Cognition 2017 Windows, Ubuntu Python GUI DeepSpeed Microsoft 2019 Linux, macOS, Windows Python, C++, CUDA Python DlibDavis King2002 Cross-platformC++C++, Python Fastai fast.ai 2018 Linux, macOS, Windows Python, CUDA Python FluxMike Innes2017 Linux, macOS, Windows (Cross-platform)JuliaJulia Horovod Uber Technologies 2017 Linux, macOS, Windows Python, C++, CUDA Python Hugging Face Transformers Hugging Face 2018 Linux, macOS, Windows Python Python Intel Data Analytics Acceleration Library Intel 2015 Linux, macOS, Windows on Intel CPU C++, Python, Java C++, Python, Java Intel Math Kernel Library 2017 and later Intel 2017 Linux, macOS, Windows on Intel CPU C/C++, DPC++, Fortran C Google JAX Google 2018 Linux, macOS, Windows Python Python Keras François Chollet 2015 Linux, macOS, Windows Python Python, R MATLAB + Deep Learning Toolbox (formerly Neural Network Toolbox) MathWorks 1992 Linux, macOS, Windows C, C++, Java, MATLAB MATLAB Microsoft Cognitive Toolkit (CNTK) Microsoft Research 2016 Windows, Linux (macOS via Docker on roadmap) C++ Python (Keras), C++, Command line, BrainScript (.NET on roadmap) MindSpore Huawei 2020 Linux, Windows, macOS, EulerOS, openEuler, OpenHarmony, Oniro OS, HarmonyOS, Android C++, Rust, Julia, Python, ArkTS, Cangjie, Java (Lite)ML.NETMicrosoft2018Windows, Linux, macOSC#, C++C#, F# Apache MXNet Apache Software Foundation 2015 Linux, macOS, Windows, AWS, Android, iOS, JavaScript Small C++ core library C++, Python, Julia, MATLAB, JavaScript, Go, R, Scala, Perl, Clojure Neural Designer Artelnics 2014 Linux, macOS, Windows C++ Graphical user interface OpenNN Artelnics 2003 Cross-platform C++ C++ OpenVINO Intel 2018 Linux, macOS, Windows C++, Python C++, C, Python, JavaScript PlaidML Vertex.AI, Intel 2017 Linux, macOS, Windows Python, C++, OpenCL Python, C++ PyTorch Meta AI 2016 Linux, macOS, Windows, Android Python, C, C++, CUDA Python, C++, Julia, R PyTorch Lightning Lightning-AI (originally William Falcon)https://github.com/williamFalcon 2019 Linux, macOS, Windows Python Python https://lightning.ai/docs/litserve/how-tos/parallelism-types Apache SINGA Apache Software Foundation 2015 Linux, macOS, Windows C++ Python, C++, Java TensorFlow Google Brain 2015 Linux, macOS, Windows, Android C++, Python, CUDA Python (Keras), C/C++, Java, Go, JavaScript, R, Julia, Swift but already with SYCL support TensorRT Nvidia 2017 Linux, Windows on Nvidia GPUs C++, Python, CUDA C++, Python Theano Université de Montréal 2007 Cross-platform Python Python (Keras) Torch Ronan Collobert, Koray Kavukcuoglu, Clement Farabet 2002 Linux, macOS, Windows, Android, iOS C, Lua Lua, LuaJIT, C, utility library for C++/OpenCL RANT (Real-time Artificial Neural Tool) Douglas Santry 2023 Linux, macOS, Windows C++, Python C++, Python Wolfram Mathematica 10 and later Wolfram Research 2014 Windows, macOS, Linux, Cloud computing C++, Wolfram Language, CUDA Wolfram Language Software Creator Initial release Software license Platform Written in Interface OpenMP support OpenCL support CUDA support Automatic differentiation Has pretrained models
Comparison of machine learning model compatibility
+Format nameDesign goalCompatible with other formatsSelf-contained DNN ModelPre-processing and Post-processingRun-time configuration for tuning & calibrationDNN model interconnectCommon platformTensorFlow, Keras, Caffe, TorchAlgorithm training / Separate files in most formats ONNXAlgorithm training / Separate files in most formats
See also Comparison of machine learning software Comparison of statistical packages Comparison of cognitive architectures Lists of open-source artificial intelligence software List of datasets for machine-learning research List of numerical-analysis software Mamba – deep learning architecture based on selective state-space models MLIR compiler — sub-project of the LLVM designed for machine learning, hardware acceleration, and high-level synthesis. tinygrad — being developed by George Hotz
References