SmiffyKMc’s gists

SmiffyKMc / Segmenting-And-Clustering.ipynb

Created August 30, 2019 20:28

Created on Cognitive Class Labs

Sorry, something went wrong. Reload?

Sorry, we cannot display this file.

Sorry, this file is invalid so it cannot be displayed.

SmiffyKMc / model_v1.py

Last active June 16, 2022 17:56

First version of the CNN

	import tensorflow as tf
	from tensorflow import keras
	from keras import layers

	inputs = keras.Input(shape=(256, 256, 3))
	x = layers.Rescaling(1./255)(inputs)
	x = layers.Conv2D(filters=32, kernel_size=3, activation=keras.activations.relu)(x)
	x = layers.MaxPooling2D(pool_size=2)(x)
	x = layers.Conv2D(filters=64, kernel_size=3, activation=keras.activations.relu)(x)
	x = layers.MaxPooling2D(pool_size=2)(x)

SmiffyKMc / data_preprocess.py

Created June 16, 2022 18:03

Steps done to clean the data and preprocess as datasets

	from tensorflow.keras.utils import image_dataset_from_directory

	test_dataset = image_dataset_from_directory(
	new_base_dir / "test",
	image_size=(256, 256),
	batch_size=32
	)

	train_dataset = image_dataset_from_directory(
	new_base_dir / "train",

SmiffyKMc / fit_model.py

Created June 16, 2022 18:15

Fitting the model and allowing for saving of the model based on the best validation loss

	callbacks = [
	keras.callbacks.ModelCheckpoint(
	filepath=f"{hotDogDir}hotdog_classifier_v1.keras",
	monitor="val_loss",
	save_best_only=True
	)
	]

	history = model.fit(
	train_dataset,

SmiffyKMc / model_evaluation.py

Created June 16, 2022 18:23

Evaluating the model

	test_model = keras.models.load_model(f'{hotDogDir}hotdog_classifier_v1.keras')
	test_loss, test_acc = test_model.evaluate(test_dataset)

	print(f"Test Acc: {test_acc:.3f}")

SmiffyKMc / data_augmentation.py

Created June 17, 2022 13:33

Example of the data augmentation and views from the plot

	data_augmentation = keras.Sequential(
	[
	layers.RandomRotation(0.3),
	layers.RandomZoom(0.2),
	layers.RandomFlip("horizontal"),
	]
	)

	plt.figure(figsize=(10, 10))
	for images, _ in test_dataset.take(1):

SmiffyKMc / model_v2.py

Created June 17, 2022 14:24

V2 of the CNN

	inputs = keras.Input(shape=(256, 256, 3))
	x = data_augmentation(inputs)
	x = layers.Rescaling(1./255)(x)
	x = layers.Conv2D(filters=32, kernel_size=3, activation=keras.activations.relu)(x)
	x = layers.MaxPooling2D(pool_size=2)(x)
	x = layers.Conv2D(filters=64, kernel_size=3, activation=keras.activations.relu)(x)
	x = layers.MaxPooling2D(pool_size=2)(x)
	x = layers.Conv2D(filters=128, kernel_size=3, activation=keras.activations.relu)(x)
	x = layers.MaxPooling2D(pool_size=2)(x)
	x = layers.Conv2D(filters=256, kernel_size=3, activation=keras.activations.relu)(x)

SmiffyKMc / model_v3.py

Created June 17, 2022 14:31

Using a convolutional base for the model

	conv_base = keras.applications.vgg16.VGG16(
	weights="imagenet",
	include_top=False,
	input_shape=(256, 256, 3)
	)

	conv_base.summary()

SmiffyKMc / get_features.py

Created June 17, 2022 15:01

Retrieving the features of the VGG16 model

	import numpy as np

	def get_features_and_labels(dataset):
	all_features = []
	all_labels = []
	for images, labels in dataset:
	preprocessed_images = keras.applications.vgg16.preprocess_input(images)
	features = conv_base.predict(preprocessed_images)
	all_features.append(features)
	all_labels.append(labels)

SmiffyKMc / model_v3.py

Created June 17, 2022 15:06

V3 of the model

	inputs = keras.Input(shape=(8, 8, 512))
	x = layers.Flatten()(inputs)
	x = layers.Dense(256)(x)
	x = layers.Dropout(0.5)(x)
	outputs = layers.Dense(1, activation= keras.activations.sigmoid)(x)
	model = keras.Model(inputs, outputs)

	model.compile(optimizer=keras.optimizers.RMSprop(),
	loss=keras.losses.BinaryCrossentropy(),
	metrics=["accuracy"])

Kieran McCarthy SmiffyKMc