Using CNTKs Python Interface for Deep Learning

on YouTube ↗01h 57m

" summary

Hands-on use of Microsoft's Cognitive Toolkit via its Python/Keras interfaces with concrete notebooks: installation steps, Azure NC6 GPU setup (Azure NC6 (K80) ~ $0.90/hr), data pipelines and code for logistic regression, MLPs, CNNs, LSTMs, and word embeddings. Measured outcomes include logistic regression ~92% on MNIST, MLP ~94–95%, CNN ~99% on MNIST, CIFAR‑10 ~75% early, and LSTM IMDB ~88%, showing what was built, what libs were used, and which results actually improve with dropout, RMSprop and pretrained features. CNTK + Keras back-end

AI-written summary · about

" chapters (16)

1. 0:00CNTK and deep learning intro
2. 1:06Machine learning basics
3. 14:26Tensors and dot products
4. 18:27GPUs for deep learning
5. 26:47Azure VM setup for CNTK
6. 30:20Logistic regression in CNTK
7. 49:59Gradient descent explained
8. 59:53Multi-layer perceptron and backpropagation
9. 1:02:22CNTK and Keras install
10. 1:13:45Keras logistic regression
11. 1:21:49MLP with Keras and dropout
12. 1:28:19Convolutional neural networks with Keras
13. 1:36:28CNN on CIFAR-10
14. 1:43:16Text classification with Keras
15. 1:50:04RNNs and LSTM explained
16. 1:55:32Summary of examples

" related talks

• 26mIntroduction to the Microsoft Cognitive Toolkit (Deep Learning Framework)Ben Sadeghi · FOSSASIA Summit 2017
• 58mDeep Learning in Microsoft Azure: CNTK, CaffeOnSpark and TensorflowJen Stirrup · NDC Oslo 2019
• 55mDeep Learning with CNTK and F#Mathias Brandewinder · NDC Sydney 2019
• 55mA Tour of Deep Learning With C++Peter Goldsborough · CppCon 2017
• 57mDeep Learning with CNTK and F#Mathias Brandewinder · NDC Oslo 2018
• 57mDeep Learning with CNTK and F#Mathias Brandewinder · NDC Minnesota 2018
• 33mBuilding end-to-end computer vision solutions from deep learning modelsVanja Paunic · Data Science Conference 4.0
• 01h 29mHands-on introduction to Deep Learning with Keras and TensorflowRodrigo Agundez · PyData Amsterdam 2018