JulieProst

from pathlib import Path

import cv2.cv2 as cv2
import numpy as np
from cachier import cachier

ROOT_FOLDER = Path(__file__).resolve().parents[0]
CACHIER_FOLDER = ROOT_FOLDER / "cache"

NUMBER_OF_ITERATIONS = 5

JulieProst

# Binarize input image
gray_image = cv2.cvtColor(original_image, cv2.COLOR_RGB2GRAY)
show_image(gray_image, "Gray image")

binarized_image = cv2.adaptiveThreshold(
    gray_image,
    maxValue=1,
    adaptiveMethod=cv2.ADAPTIVE_THRESH_GAUSSIAN_C,
    thresholdType=cv2.THRESH_BINARY,
    blockSize=9,

JulieProst

import cv2.cv2 as cv2
import numpy as np

# Read the original image
IMAGES_FOLDER = "Path to your image folder"
original_image = cv2.imread(str(IMAGES_FOLDER / "spoon.jpg"))

# Define boundary rectangle containing the foreground object
height, width, _ = original_image.shape
left_margin_proportion = 0.3

JulieProst

N_EPOCH_SEARCH = 40

tuner.search(x_train, y_train, epochs=N_EPOCH_SEARCH, validation_split=0.1)

JulieProst

tuner.search_space_summary()

JulieProst

HYPERBAND_MAX_EPOCHS = 40
MAX_TRIALS = 20
EXECUTION_PER_TRIAL = 2

JulieProst

# Show a summary of the search
tuner.results_summary()

# Retrieve the best model.
best_model = tuner.get_best_models(num_models=1)[0]

# Evaluate the best model.
loss, accuracy = best_model.evaluate(x_test, y_test)

JulieProst

hp.Float(
    'learning_rate',
    min_value=1e-5,
    max_value=1e-2,
    sampling='LOG',
    default=1e-3
)

JulieProst

model.compile(
    optimizer=keras.optimizers.Adam(1e-3),
    loss='sparse_categorical_crossentropy',
    metrics=['accuracy']
)

JulieProst

Dense(
    units=hp.Int(
        'units',
        min_value=32,
        max_value=512,
        step=32,
        default=128
    ),
    activation=hp.Choice(
        'dense_activation',