AI for every developer
summary
Claims that .NET developers can build and run neural models locally using ML.NET and exportable ONNX models, demonstrates training on the handwritten digit dataset (MNIST — quoted here as ~10,000 training items and ~60 test items), and shows a WPF-style canvas app that recognizes drawn digits successfully. Uses Visual Studio, TensorFlow and CNTK for comparison, notes GPU training is faster but his demo falls back to CPU, and emphasizes data volume and error-rate variability.
AI-written summary · about
chapters (11)
- 0:00Start
- 2:10Computational Power
- 6:04Machine Learning
- 8:53Preparing Data
- 12:59Train a Model
- 13:39Reinforcement Learning
- 17:53Deep Neural Network
- 19:16Simple Artificial Intelligence as a Service
- 29:07Tensorflow
- 31:27Visual Representation of the Training Data
- 34:49Resources
related talks
01h 00mAI for Every Developer
41mAI + Machine Learning for .NET Developers
57mEasier AI allows everyone to build smart apps
53mBuilding Intelligent .NET MAUI Apps with ML.NET
42mStep into Machine Learning .NET
58mExploring Machine Learning for .Net Developers
08mInterview Speaker
58mDeep Learning in Microsoft Azure: CNTK, CaffeOnSpark and Tensorflow