jamesonthecrow

# Convert to mobile formats
import coremltools
import tensorflow as tf
import tempfile

def convert_to_coreml(model):
    return coremltools.converters.keras.convert(
        model,
        input_names=['input'],
        output_names=['digit']

jamesonthecrow

import keras
from keras.datasets import mnist
keras.backend.clear_session()

(x_train, y_train), (x_test, y_test) = mnist.load_data()


def build_model():
    input = keras.layers.Input((28, 28, 1))
    out = keras.layers.Conv2D(16, 3, strides=2, activation='relu')(input)

jamesonthecrow

import UIKit
import AVFoundation
import Fritz

class ViewController: UIViewController, UIImagePickerControllerDelegate,
UINavigationControllerDelegate {
    
  @IBOutlet var imageView: UIImageView!
  var maskView: UIImageView!
  var backgroundView: UIImageView!

jamesonthecrow

// Run the image through the model to identify pixels belonging to a pet.
FritzVisionSegmentResult segmentResult = predictor.predict(visionImage);

jamesonthecrow

// Determine how to rotate the image from the camera used.
int imgRotation = FritzVisionOrientation.getImageRotationFromCamera(this, cameraId);

// Create a FritzVisionImage object from android.media.Image
FritzVisionImage visionImage = FritzVisionImage.fromMediaImage(image, imgRotation);

jamesonthecrow

extension ViewController: AVCaptureVideoDataOutputSampleBufferDelegate {

  func captureOutput(_ output: AVCaptureOutput, didOutput sampleBuffer: CMSampleBuffer, from connection: AVCaptureConnection) {
    let image = FritzVisionImage(buffer: sampleBuffer)
    image.metadata = FritzVisionImageMetadata()
    image.metadata?.orientation = FritzImageOrientation(from: connection)

    guard let result = try? visionModel.predict(image) else { return }
    
    let mask = result.buildSingleClassMask(

jamesonthecrow

// ...
import Fritz

class ViewController: UIViewController {

    var cameraView: UIImageView!
    
    override func viewDidLoad() {
        super.viewDidLoad()

jamesonthecrow

from coremltools.models.neural_network import flexible_shape_utils

def make_mlmodel_flexible(spec, size_range=(100, 1920):
    """Make input and output sizes of a Core ML model flexible.

    Args:
        spec (NeuralNetwork_pb2): a Core ML neural network spec
        size_range ([Int]): a tuple containing the min and max input sizes.
    """
    size_range_spec = flexible_shape_utils.NeuralNetworkImageSizeRange()

jamesonthecrow

# Create a model with a normal convolution.
inpt = keras.layers.Input(shape=(500, 500, 3))
out = keras.layers.Conv2D(10, 10)(inpt)
model = keras.models.Model(inpt, out)
mlmodel = coremltools.converters.keras.convert(model)
mlmodel.save('convolution.mlmodel')

# Create a model with a dialted (atrous) convolution.
inpt = keras.layers.Input(shape=(500, 500, 3))
out = keras.layers.Conv2D(10, 10, dilation_rate=4)(inpt)

jamesonthecrow

// PUT A BREAKPOINT HERE

// Compile the model.
let compiledModelURL = try! MLModel.compileModel(at: assetPath!)

// Initialize the model for use on a specific set of hardware
let config = MLModelConfiguration()
config.computeUnits = .all // can be .all, .cpuAndGPU, or .cpuOnly
let mlmodel = try! MLModel(contentsOf: compiledModelURL, configuration: config)