Machine Learning in .NET (ML.NET)
summary
ML.NET demos show concrete results: a GitHub issue classifier built with text transforms, key encoding and a pipeline that saves a model to disk; a movie recommender using factorization machines on MovieLens (≈20M ratings) turned into a binary like/dislike and achieving about 72% accuracy in this demo; and image scoring via TensorFlow model scoring (Inception labels). The preview 0.6 API exposes typed estimators/transformers, supports on-disk models, has samples in the repo (≈3,500 stars) and a coming UI to generate training and consumption code.
AI-written summary · about
chapters (13)
- 0:00Intro
- 0:19Your platform for building anything
- 4:31Pre-built ML Models (Azure Cognitive Services)
- 6:02Build your own ML models
- 6:13ML.NET is a framework
- 11:06ML.NET - Machine Learning framework made for NET developers
- 15:19GitHub Issue Classification
- 15:29GitHub Issue Example
- 15:56Label (Output)
- 16:21Machine Learning Workflow
- 24:24Approaches for Movie Recommendation
- 41:35Recommender Systems
- 42:52Good start, but obvious limitations
related talks
52mML.NET - Machine Learning made for .NET
01h 04mIntroduction to ML.NET
01h 03mMachine learning for C# developers: Introducing ML
58mExploring Machine Learning for .Net Developers
57mMachine Learning Made Easy – An Introduction to ML.NET
45mML: ML.NET: deep dive into dev AI
57mMachine Learning for .NET Developers with ML.NET
46mMachine Learning for .NET Developers