N Ways to Do Machine Learning on .NET

on YouTube ↗27m

" summary

Compares concrete .NET options for building models: coding linear regression from scratch (about 100 lines of F# or ~250 lines of C#), using automatic-differentiation via DiffSharp, or ready libraries like Accord.NET (stable but hard to extend). Covers low-level frameworks (CNTK with new C# API, TensorFlow bindings), point‑and‑click Azure ML Studio, experimental TensorFlowSharp, and pretrained Cognitive Services. Highlights a newcomer ML.NET (v0.7) as a high‑level, cross‑platform pipeline layer backed by TensorFlow/CNTK, useful for prototyping though not yet production‑mature.

AI-written summary · about

" chapters (6)

1. 0:00Start
2. 7:23CODING IT ALL FROM SCRATCH
3. 10:17#1.5 - LITTLE HELP FROM LIBRARIES
4. 13:30CNTK- MICROSOFT COGNITIVE TOOLKIT
5. 15:58#3.5-F# BINDINGS FOR CNTK
6. 21:11HONORABLE MENTIONS

" related talks

• 57mMachine Learning for .NET Developers with ML.NETOlia Havrysh · .NET Fest 2018
• 57mMachine Learning for .Net DevelopersOlia Havrysh · .NET Fest 2019
• 52mML.NET - Machine Learning made for .NETPeter Drougge · Update Conference Prague 2018
• 59mML.NET: Using Machine Learning in Ordinary .NET ProjectsSergey Korzh · .NET Fest 2018
• 01h 03mMachine learning for C# developers: Introducing MLJeff Prosise · DotNext 2019 Moscow
• 59mMachine Learning in .NET (ML.NET)Agatha Stena · .NET Conf 2018
• 42mMachine Learning for .NET DevelopersOleksandr Krakovetskiy · .NET fwdays'21 conference
• 58mExploring Machine Learning for .Net DevelopersNiloshima Srivastava · AzConf 2020