Tutorials¶
For a quick tour if you are familiar with another deep learning toolkit please fast forward to CNTK 200 (A guided tour) for a range of constructs to train and evaluate models using CNTK.
- Classify cancer using simulated data (Logistic Regression)
- CNTK 101:Logistic Regression with NumPy (source)
- Classify cancer using simulated data (Feed Forward, FFN)
- CNTK 102: Feed Forward network with NumPy (source)
- Recognize hand written digits (OCR) with MNIST data
- CNTK 103 Part A: MNIST data preparation (source), Part B: Multi-class logistic regression classifier (source) Part C: Multi-layer perceptron classifier (source) Part D: Convolutional neural network classifier (source)
- Learn how to predict the stock market
- CNTK 104: Time Series basics with finance data (source with finance data)
- Compress (using autoencoder) hand written digits from MNIST data with no human input (unsupervised learning, FFN)
- CNTK 105 Part A: MNIST data preparation (source), Part B: Feed Forward autoencoder (source)
- Forecasting using data from an IOT device
- CNTK 106: LSTM based forecasting - Part A: with simulated data (source), Part B: with real IOT data (source)
- Quick tour for those familiar with other deep learning toolkits
- CNTK 200: Guided Tour (source)
- Recognize objects in images from CIFAR-10 data (Convolutional Network, CNN)
- CNTK 201 Part A: CIFAR data preparation (source), Part B: VGG and ResNet classifiers (source)
- Infer meaning from text snippets using LSTMs and word embeddings
- CNTK 202: Language understanding (source)
- Train a computer to perform tasks optimally (e.g., win games) in a simulated environment
- CNTK 203: Reinforcement learning basics with OpenAI Gym data (source)
- Translate text from one domain (grapheme) to other (phoneme)
- CNTK 204: Sequence to sequence basics with CMU pronouncing dictionary (source)
- Teach a computer to paint like Picasso or van Gogh
- CNTK 205: Artistic Style Transfer (source)
- Produce realistic images with no human input (unsupervised learning)
- CNTK 206 Part A: MNIST data preparation (source), Part B: Basic Generative Adversarial Networks (GAN) (source), Part C: Deep Convolutional GAN (source) Part D: Wasserstein GAN and Loss Sensitive GAN (source)
- Training with Sampled Softmax
- CNTK 207: Training with Sampled Softmax (source)
- Training with Connectionist Temporal Classification
- CNTK 208: Training with Connectionist Temporal Classification (source)
- Recognize flowers and animals in natural scene images using deep transfer learning
- CNTK 301: Deep transfer learning with pre-trained ResNet model (source)
- Generate higher resolution images from low resolution ones
- CNTK 302 Part A : Use pre-trained models for generating super-resolution images (source), Part B: Train super resolution models using CNNs and GANs (source)
- Compare the similarity between a pair of documents
- CNTK 303: Deep structured semantic modeling with LSTM (source)
Try these notebooks pre-installed on CNTK Azure Notebooks for free.
For our Japanese users, you can find some of the tutorials in Japanese (unsupported).