https://github.com/dotnet/machinelearning

ML.NET is an open source and cross-platform machine learning framework for .NET.
https://github.com/dotnet/machinelearning

algorithms dotnet machine-learning ml

Last synced: 27 days ago
JSON representation

ML.NET is an open source and cross-platform machine learning framework for .NET.

• Host: GitHub
• URL: https://github.com/dotnet/machinelearning
• Owner: dotnet
• License: mit
• Created: 2018-05-03T16:20:42.000Z (over 6 years ago)
• Default Branch: main
• Last Pushed: 2024-04-10T17:55:17.000Z (7 months ago)
• Last Synced: 2024-04-10T21:31:13.809Z (7 months ago)
• Topics: algorithms, dotnet, machine-learning, ml
• Language: C#
• Homepage: https://dot.net/ml
• Size: 93.4 MB
• Stars: 8,812
• Watchers: 576
• Forks: 1,838
• Open Issues: 910
• Metadata Files:
  • Readme: README-oneDAL.md
  • Contributing: CONTRIBUTING.md
  • License: LICENSE
  • Code of conduct: CODE_OF_CONDUCT.md
  • Security: SECURITY.md
  • Roadmap: ROADMAP.md

Awesome Lists containing this project

• awesome-dotnet-core - ML.NET - Cross-platform open-source machine learning framework which makes machine learning accessible to .NET developers [http://dot.net/ml](http://dot.net/ml). (Frameworks, Libraries and Tools / Machine Learning and Data Science)
• awesome-dotnet-datascience - ML.NET - Cross-platform open-source machine learning framework which makes machine learning accessible to .NET developers. This framework wraps several other libraries and is actively being developed. So a great place to start. (Machine Learning and Differential Programming)
• awesome-dotnet-core - ML.NET - ML.NET是.NET的开源和跨平台机器学习框架。 (框架, 库和工具 / 机器学习和科学研究)
• awesome-dotnet-machine-learning - ML.NET - platform machine learning framework for .NET. (Uncategorized / Uncategorized)
• fucking-awesome-dotnet-core - ML.NET - Cross-platform open-source machine learning framework which makes machine learning accessible to .NET developers [http://dot.net/ml](http://dot.net/ml). (Frameworks, Libraries and Tools / Machine Learning and Data Science)
• awesome-dotnet - ML.NET - ML.NET is an open source and cross-platform machine learning framework for .NET. (Libraries, Frameworks and Tools / Machine Learning and Data Science)
• awesome-dotnet-core - ML.NET - Cross-platform open-source machine learning framework which makes machine learning accessible to .NET developers [http://dot.net/ml](http://dot.net/ml). (Frameworks, Libraries and Tools / Machine Learning and Data Science)
• awesome-yolo-object-detection - ML.NET - platform machine learning framework for .NET. (Other Versions of YOLO)
• awesome-yolo-object-detection - ML.NET - platform machine learning framework for .NET. (Other Versions of YOLO)

README

        
# oneDAL supported algorithms

oneAPI Data Analytics Library (oneDAL) is a library providing highly optimized machine learning and data analytics kernels. Some of these kernels is integrated into ML.NET via C++/C# interoperability.

[oneDAL Documentation](http://oneapi-src.github.io/oneDAL/) | [oneDAL Repository](https://github.com/oneapi-src/oneDAL)

> Please note that oneDAL acceleration paths are only available in x64 architectures

Integration consists of:

* A "native" component (under `src/Native/Microsoft.ML.OneDal`) implementing wrapper to pass data and parameters to oneDAL;

* Dispatching to oneDAL kernels inside relevant learners: `OLS` (`src/Microsoft.ML.Mkl.Components`), `Logistic Regression` (`src/Microsoft.ML.StandardTrainers`), `Random Forest` (`src/Microsoft.ML.FastTree`);

## Running ML.NET trainers with dispatching to oneDAL kernels

Currently, dispatching to oneDAL inside ML.NET is regulated by `MLNET_BACKEND` environment variable. If it's set to `ONEDAL`, oneDAL kernel will be used, otherwise - default ML.NET.