search

dotnet / machinelearning

ML.NET is an open source and cross-platform machine learning framework for .NET.
https://dot.net/ml
MIT License
9k stars 1.88k forks source link

Native .NET alternatives of Python's data science and AI libraries #5713

Open neoffer opened 3 years ago

neoffer commented 3 years ago

I don't know how to develop on-premise chatbot app using .NET. Because .NET does not provide native data science and ai libraries.

Please develop native libraries for data science, machine learning and deep learning in .NET like TensorFlow, NumPy, SciPy, Matplotlab, Keras, NLTK, Pandas so we can develop on-premise ChatBot application in .NET Core using Microsoft's NuGet Packages.

jwood803 commented 3 years ago

Hey @neoffer. For those check out the SciSharp stack. ML.NET is already utilizing Tensorflow.NET but they have quite a few .NET bindings for those other Python libraries you mentioned.

luisquintanilla commented 3 years ago

@neoffer to echo what @jwood803 pointed out. Take a look at the interoperability section in FsLab.

https://fslab.org/

If you're looking for more Python-like and custom DS solutions, SciSharp is a good option. For statistical tools, take a look at FsLab. For classical ML algorithms and interop with existing TF / ONNX models ML.NET would help you there.

saint4eva commented 3 years ago

Hey @neoffer. For those check out the SciSharp stack. ML.NET is already utilizing Tensorflow.NET but they have quite a few .NET bindings for those other Python libraries you mentioned.

One difficult challenge with SciSharp libraries is the naming convention - it is not idiomatic to .NET/ C#. They put in so much energy trying to please and attract already python developers, instead making it more .NET/ C# naming and idiosyncrasies so .NET Developers can just use it to develop solutions.