Edoh Kodjo kedevked

	app = Flask(__name__)
	graph = tf.get_default_graph()

	@app.route('/predict', methods=['POST', 'GET'])
	@cross_origin()
	def predict_():
	'''
	returns a flower prediction
	'''
	global graph

kedevked / dog_breed_flask.py

Created April 28, 2019 17:43

run flask without thread

	if __name__ == "__main__":
	app.run("0.0.0.0", port=5000, debug=False, threaded=False)

kedevked / dog_breed_global_average_pooling.py

Created April 28, 2019 17:37

transfer learning with only GlobalAveragePooling layer

	model = Sequential()
	model.add(GlobalAveragePooling2D(train_inception_v3.shape[1:]))
	model.add(Dense(133, activation='softmax'))

kedevked / dog_breed_transfer.learning.py

Created April 28, 2019 17:26

transfer learning with inception v3

	bottleneck_features = np.load('inception_v3_file.npz')
	train_inception_v3 = bottleneck_features['train']
	valid_inception_v3 = bottleneck_features['valid']
	test_inception_v3 = bottleneck_features['test']


	model = Sequential()
	model.add(Dense(input_shape=train_inception_v3.shape[1:]), units=16))
	model.add(Conv2D(filters = 16, kernel_size=2, strides=1, padding ='valid', activation='relu'))
	model.add(MaxPooling2D(pool_size = (2), strides=(2), padding ='valid'))

kedevked / dog_breed.conv.py

Created April 28, 2019 17:09

convolutional layers for dog breed classifier

	model.add(Dense(input_shape=[224, 224, 3], units=16))
	model.add(Conv2D(filters = 16, kernel_size=2, strides=1, padding ='valid', activation='relu'))
	model.add(MaxPooling2D(pool_size = (2), strides=(2), padding ='valid'))
	model.add(Conv2D(filters = 32, kernel_size=2, strides=1, padding ='valid', activation='relu'))
	model.add(MaxPooling2D(pool_size = (2), strides=(2), padding ='valid'))
	model.add(Conv2D(filters = 64, kernel_size=2, strides=1, padding ='valid', activation='relu'))
	model.add(MaxPooling2D(pool_size = (2), strides=(2), padding ='same'))
	model.add(GlobalAveragePooling2D())
	model.add(Dense(units=133, activation='softmax'))

kedevked / tfjs-webworker.html

Created March 14, 2019 02:32

use tensorflow.js in a web worker

	<head>
	<script src="https://cdn.jsdelivr.net/npm/@tensorflow/[email protected]/dist/tf.min.js"></script>
	<script>
	const worker_function = () => {

	onmessage = () => {
	console.log('from web worker')
	this.window = this
	importScripts('https://cdn.jsdelivr.net/npm/[email protected]/setImmediate.min.js')
	importScripts('https://cdn.jsdelivr.net/npm/@tensorflow/[email protected]')

kedevked / mnist-sequential.js

Created March 8, 2019 21:38

build a sequential model for transfert learning

	async function buildModel(featureExtractor, units) {
	return tf.sequential({
	layers: [
	// Flattens the input to a vector so we can use it in a dense layer. While
	// technically a layer, this only performs a reshape (and has no training
	// parameters).
	// slice so as not to take the batch size
	tf.layers.flatten(
	{ inputShape: featureExtractor.outputs[0].shape.slice(1) }),
	// add all the layers of the model to train