iOS Core ML On Device Training

iOSCoreMLOnDeviceTraining/model + scripts/coremlupdatablemodel.py

@@ -0,0 +1,68 @@
+import coremltools
+
+output_labels = ['Cat', 'Dog']
+coreml_model = coremltools.converters.keras.convert('model.h5', input_names=['image'], output_names=['output'],
+ class_labels=output_labels,
+ image_input_names='image')
+
+coreml_model.author = 'Anupam Chugh'
+coreml_model.short_description = 'Cat Dog Classifier converted from a Keras model'
+coreml_model.input_description['image'] = 'Takes as input an image'
+coreml_model.output_description['output'] = 'Prediction as cat or dog'
+coreml_model.output_description['classLabel'] = 'Returns Cat Or Dog as class label'
+
+
+coreml_model.save('catdogmodel.mlmodel')
+
+coreml_model_path = "./catdogmodel.mlmodel"
+
+spec = coremltools.utils.load_spec(coreml_model_path)
+builder = coremltools.models.neural_network.NeuralNetworkBuilder(spec=spec)
+builder.inspect_layers(last=3)
+builder.inspect_input_features()
+
+neuralnetwork_spec = builder.spec
+
+neuralnetwork_spec.description.input[0].type.imageType.width = 150
+neuralnetwork_spec.description.input[0].type.imageType.height = 150
+
+# Set input and output description
+neuralnetwork_spec.description.input[0].shortDescription = 'Takes as input an image'
+neuralnetwork_spec.description.output[0].shortDescription = 'Prediction as cat or dog'
+
+# Provide metadata
+neuralnetwork_spec.description.metadata.author = 'Anupam Chugh'
+neuralnetwork_spec.description.metadata.license = 'MIT'
+neuralnetwork_spec.description.metadata.shortDescription = (
+ 'Cat Dog Classifier converted from a Keras model')
+
+
+model_spec = builder.spec
+
+# make_updatable method is used to make a layer updatable. It requires a list of layer names.
+# dense_5 and dense_6 are two innerProduct layer in this example and we make them updatable.
+builder.make_updatable(['dense_5', 'dense_6'])
+
+# Categorical Cross Entropy or Mean Squared Error can be chosen for the loss layer.
+builder.set_categorical_cross_entropy_loss(name='lossLayer', input='output')
+
+
+# in addition of the loss layer, an optimizer must also be defined. SGD and Adam optimizers are supported.
+# SGD has been used for this example. To use SGD, one must set lr(learningRate) and batch(miniBatchSize) (momentum is an optional parameter).
+from coremltools.models.neural_network import SgdParams
+builder.set_sgd_optimizer(SgdParams(lr=0.01, batch=5))
+
+# The number of epochs must be set as follows.
+builder.set_epochs(1)
+
+model_spec.isUpdatable = True
+model_spec.specificationVersion = coremltools._MINIMUM_UPDATABLE_SPEC_VERSION
+
+
+# Set training inputs descriptions
+model_spec.description.trainingInput[0].shortDescription = 'Image for training and updating the model'
+model_spec.description.trainingInput[1].shortDescription = 'Set the value as Cat or Dog and update the model'
+
+
+# save the updated spec
+coremltools.utils.save_spec(model_spec, "CatDogUpdatable.mlmodel")