Nearest Neighbor Classifier

This topic demonstrates the process of creating an updatable empty k-nearest neighbor model using coremltools.

Create the Classifier

  1. Create the classifier and apply its properties:
number_of_dimensions = 128

from coremltools.models.nearest_neighbors import KNearestNeighborsClassifierBuilder
builder = KNearestNeighborsClassifierBuilder(input_name='input',
                                             output_name='output',
                                             number_of_dimensions=number_of_dimensions,
                                             default_class_label='defaultLabel',
                                             number_of_neighbors=3,
                                             weighting_scheme='inverse_distance',
                                             index_type='linear')

builder.author = 'Core ML Tools Example'
builder.license = 'MIT'
builder.description = 'Classifies {} dimension vector based on 3 nearest neighbors'.format(number_of_dimensions)

builder.spec.description.input[0].shortDescription = 'Input vector to classify'
builder.spec.description.output[0].shortDescription = 'Predicted label. Defaults to \'defaultLabel\''
builder.spec.description.output[1].shortDescription = 'Probabilities / score for each possible label.'

builder.spec.description.trainingInput[0].shortDescription = 'Example input vector'
builder.spec.description.trainingInput[1].shortDescription = 'Associated true label of each example vector'

Note: An empty knn model is updatable by default:

# By default an empty knn model is updatable
builder.is_updatable
True
  1. Confirm the number of dimensions are set correctly:
# Let's confirm the number of dimension is set correctly
builder.number_of_dimensions
128

Set the Number of Neighbors Value

  1. Verify the current number of neighbors value:
# Let's check what the value of 'numberOfNeighbors' is
builder.number_of_neighbors
3

Note: The number of neighbors is bounded by the default range:

# The number of neighbors is bounded by the default range...
builder.number_of_neighbors_allowed_range()
(1, 1000)

If you set the number of neighbors to a value outside of this default range, an ValueError will occur as shown in the Out tab:

# If we try to set the number of neighbors to a value outside of this range
builder.number_of_neighbors = 1001
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-17-e8bea591e72c> in <module>
      1 # If we try to set the number of neighbors to a value outside of this range
----> 2 builder.number_of_neighbors = 1001

~/eng/sources/coreml/coremltools/coremltools/models/nearest_neighbors/builder.py in number_of_neighbors(self, number_of_neighbors)
    312                 self.spec.kNearestNeighborsClassifier.numberOfNeighbors.defaultValue = number_of_neighbors
    313             else:
--> 314                 raise ValueError('number_of_neighbors is not within range bounds')
    315         else:
    316             spec_values = self.spec.kNearestNeighborsClassifier.numberOfNeighbors.set.values

ValueError: number_of_neighbors is not within range bounds
  1. Change the bounds for the number of neighbors. Individual values can be set for the numberOfNeighbors parameter:
# Instead of a range, you can a set individual values that are valid for the numberOfNeighbors parameter.
builder.set_number_of_neighbors_with_bounds(3, allowed_set={ 1, 3, 5 })

3, Verify change using the number_of_neighbors_allowed_set() method.

# Check out the results of the previous operation
builder.number_of_neighbors_allowed_set()
{1, 3, 5}
  1. The number of neighbors value can now be set without an error:
# And now if you attempt to set it to an invalid value...
builder.number_of_neighbors = 4
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-20-98c77c72c722> in <module>
      1 # And now if you attempt to set it to an invalid value...
----> 2 builder.number_of_neighbors = 4

~/eng/sources/coreml/coremltools/coremltools/models/nearest_neighbors/builder.py in number_of_neighbors(self, number_of_neighbors)
    320                     self.spec.kNearestNeighborsClassifier.numberOfNeighbors.defaultValue = number_of_neighbors
    321                     return
--> 322             raise ValueError('number_of_neighbors is not an allowed value')
    323 
    324     def set_number_of_neighbors_with_bounds(self, number_of_neighbors, allowed_range=None, allowed_set=None):

ValueError: number_of_neighbors is not valid

If desired, you can revert back to a valid range:

# And of course you can go back to a valid range
builder.set_number_of_neighbors_with_bounds(3, allowed_range=(1, 30))

Set the Index Type

  1. Verify the current index type:
# Let's see what the index type is
builder.index_type
'linear'
  1. Set the index and leaf size:
# Let's set the index to kd_tree with leaf size of 30
builder.set_index_type('kd_tree', 30)
builder.index_type
'kd_tree'
  1. Save the model:
mlmodel_updatable_path = './UpdatableKNN.mlmodel'

# Save the updated spec
from coremltools.models import MLModel
mlmodel_updatable = MLModel(builder.spec)
mlmodel_updatable.save(mlmodel_updatable_path)