Deep Learning with CNTK and F#

on YouTube ↗57m

" summary

Shows using the CNTK .NET API from F# with a small wrapper library and a VS Code scripting workflow, building sequential models (convolution + activation + pooling) and mini‑batch readers. Demonstrates a synthetic 8x8 test (64 images, 32 ones / 32 twos) and a MNIST subset, exports learned 8x8 weight filters, and saves models via NuGet package CNTK.FSharp 0.1.5. Training runs on CPU (CUDA GPU support added) and yields about 2% error on 500 test images.

AI-written summary · about

" chapters (32)

1. 0:00Introduction
2. 0:30Who am I
3. 1:18Context
4. 2:34Plan
5. 3:38Deployment
6. 4:09Rendering Numbers on Screen
7. 4:38Creating a System
8. 6:44Volume Controller
9. 7:20Weights
10. 8:25Filters
11. 9:09Machine Learning
12. 16:41Deep Learning
13. 21:09CNTK
14. 25:34What is a tensor
15. 27:08How does this plugin work
16. 28:34Fully trained model
17. 29:38Variables
18. 29:58Code
19. 32:56Dataset
20. 34:42Model
21. 39:21Running the model
22. 40:47Results
23. 42:45API
24. 43:02Raw API
25. 43:44Models
26. 45:10Demo
27. 47:25Data Set
28. 50:36change model
29. 52:40model file
30. 54:00model predictions
31. 54:59conclusion
32. 56:12thank you

" related talks

• 55mDeep Learning with CNTK and F#Mathias Brandewinder · NDC Sydney 2019
• 57mDeep Learning with CNTK and F#Mathias Brandewinder · NDC Minnesota 2018
• 26mIntroduction to the Microsoft Cognitive Toolkit (Deep Learning Framework)Ben Sadeghi · FOSSASIA Summit 2017
• 01h 57mUsing CNTKs Python Interface for Deep LearningDave DeBarr · PyData Seattle 2017
• 55mA Tour of Deep Learning With C++Peter Goldsborough · CppCon 2017
• 01h 05mDeep Learning with PyTorchSeth Juarez · NDC London 2019
• 44mAI for every developerGian Paolo Santopaolo · Future Tech 2019
• 57mDeep Learning with C++Peter Goldsborough · Meeting C++ 2017