Name		Name	Last commit message	Last commit date
parent directory ..
images		images
models		models
OnDeviceTraining_API_Usage.md		OnDeviceTraining_API_Usage.md
README.md		README.md
updatable_linked_model.ipynb		updatable_linked_model.ipynb
updatable_mnist.ipynb		updatable_mnist.ipynb
updatable_nearest_neighbor_classifier.ipynb		updatable_nearest_neighbor_classifier.ipynb
updatable_tiny_drawing_classifier.ipynb		updatable_tiny_drawing_classifier.ipynb

README.md

Updatable Models Examples

In this set of notebooks examples, we show how to create different types of updatable models using coremltools.

Updatable Neural Network Classifier on MNIST Dataset
This notebook demonstrates the process of creating a simple convolutional model on the MNIST dataset with Keras, converting it to a Core ML model, and making it updatable. The updatable model has 2 updatable layers and uses Categorical Cross Entropy Loss and SGD Optimizer.
Updatable Tiny Drawing Classifier - Pipeline Model
This notebook creates a model which can be used to train a simple drawing / sketch classifier based on user examples.
The model is a pipeline composed of a drawing embedding model and an updatable nearest neighbor classifier.
Updatable Tiny Drawing Classifier - Linked Pipeline Model
This notebook creates a model which can be used to train a simple drawing / sketch classifier based on user examples. The model is a 'linked' pipeline composed of a 'linked' drawing embedding model and an updatable nearest neighbor classifier.
Updatable Nearest Neighbor Classifier
This notebook makes an empty updatable nearest neighbor classifier. Before updating with training examples it predicts 'defaultLabel' for all input.

In addition of the above examples, a short document on CoreML 3.0 Update Task API usage is provided here.