def generate_prediction(description, testing_file_content, testing_image_feature_dict, prediction_file_prefix, feature_extension):
"""Generate prediction.
:param description: the folder name of the working directory
:type description: string
:param testing_file_content: the content in the testing file
:type testing_file_content: numpy array
:param testing_image_feature_dict: the features of the testing images which is saved in a dict
:type testing_image_feature_dict: dict
:param prediction_file_prefix: the prefix of the prediction file
:type prediction_file_prefix: string
:param feature_extension: the extension of the feature files
:type feature_extension: string
:return: the prediction file will be saved to disk
:rtype: None
"""
print("\nGenerating prediction ...")
working_directory = common.get_working_directory(description)
model_path_rule = os.path.join(working_directory, "*" + common.KERAS_MODEL_EXTENSION)
metric_list = METRIC_LIST_DICT[feature_extension]
for model_path in sorted(glob.glob(model_path_rule)):
model_name = os.path.basename(os.path.splitext(model_path)[0])
print("\nWorking on {} ...".format(model_name))
# Init a keras model with specific weights
final_feature = solution_basic.get_final_feature(testing_image_feature_dict.values()[0], testing_image_feature_dict.values()[0], metric_list)
dimension = final_feature.size
model = keras_related.init_model(dimension)
model.load_weights(model_path)
# Add progress bar
progress_bar = pyprind.ProgBar(testing_file_content.shape[0], monitor=True)
# Generate prediction
prediction_list = []
for _, file_1_name, file_2_name in testing_file_content:
file_1_feature = testing_image_feature_dict[file_1_name]
file_2_feature = testing_image_feature_dict[file_2_name]
final_feature = solution_basic.get_final_feature(file_1_feature, file_2_feature, metric_list)
final_feature = final_feature.reshape(1, -1)
probability_estimates = model.predict_proba(final_feature, verbose=0)
prediction = probability_estimates[0, 1]
prediction_list.append(prediction)
# Update progress bar
progress_bar.update()
# Report tracking information
print(progress_bar)
# Write prediction
prediction_file_name = prediction_file_prefix + model_name + "_" + str(int(time.time())) + ".csv"
solution_basic.write_prediction(testing_file_content, np.array(prediction_list), prediction_file_name)
def perform_training(image_feature_list, image_index_list, description, feature_extension, nb_epoch):
"""Perform training phase.
:param image_feature_list: the features of the images
:type image_feature_list: list
:param image_index_list: the indexes of the images
:type image_index_list: list
:param description: the folder name of the working directory
:type description: string
:param feature_extension: the extension of the feature files
:type feature_extension: string
:param nb_epoch: the maximum number of epochs
:type nb_epoch: int
:return: the model files will be saved to disk
:rtype: None
"""
print("Performing training phase ...")
# Reset the working directory
common.reset_working_directory(description)
working_directory = common.get_working_directory(description)
# Cross Validation
fold_num = 5
best_score_array = np.zeros(fold_num)
best_score_index_array = np.zeros(fold_num)
label_kfold = LabelKFold(image_index_list, n_folds=fold_num)
# Add progress bar
progress_bar = pyprind.ProgBar(fold_num, monitor=True)
metric_list = METRIC_LIST_DICT[feature_extension]
for fold_index, fold_item in enumerate(label_kfold):
print("\nWorking on the {:d}/{:d} fold ...".format(fold_index + 1, fold_num))
# Generate final data set
X_train, Y_train = solution_basic.convert_to_final_data_set(image_feature_list, image_index_list, fold_item[0], 1, metric_list)
X_test, Y_test = solution_basic.convert_to_final_data_set(image_feature_list, image_index_list, fold_item[1], None, metric_list)
# Perform training
model_name = "Model_{:d}".format(fold_index + 1) + common.KERAS_MODEL_EXTENSION
model_path = os.path.join(working_directory, model_name)
best_score_index, best_score = keras_related.train_model(X_train, Y_train, X_test, Y_test, model_path, nb_epoch)
best_score_array[fold_index] = best_score
best_score_index_array[fold_index] = best_score_index
print("For the {:d} fold, the Keras model achieved the score {:.4f} at the {:d} epoch.".format(fold_index + 1, best_score, best_score_index))
# Update progress bar
progress_bar.update()
# Report tracking information
print(progress_bar)
print("\nThe best score is {:.4f} and the highest epoch is {:d}.".format(np.max(best_score_array), np.max(best_score_index_array).astype(np.int)))
def generate_prediction(description, testing_file_content,
testing_image_feature_dict, prediction_file_prefix,
feature_extension):
"""Generate prediction.
:param description: the folder name of the working directory
:type description: string
:param testing_file_content: the content in the testing file
:type testing_file_content: numpy array
:param testing_image_feature_dict: the features of the testing images which is saved in a dict
:type testing_image_feature_dict: dict
:param prediction_file_prefix: the prefix of the prediction file
:type prediction_file_prefix: string
:param feature_extension: the extension of the feature files
:type feature_extension: string
:return: the prediction file will be saved to disk
:rtype: None
"""
print("\nGenerating prediction ...")
working_directory = common.get_working_directory(description)
model_path_rule = os.path.join(working_directory,
"*" + common.SCIKIT_LEARN_EXTENSION)
metric_list = METRIC_LIST_DICT[feature_extension]
for model_path in sorted(glob.glob(model_path_rule)):
model_name = os.path.basename(os.path.splitext(model_path)[0])
print("\nWorking on {} ...".format(model_name))
# Load the sklearn model
classifier = joblib.load(model_path)
# Add progress bar
progress_bar = pyprind.ProgBar(testing_file_content.shape[0],
monitor=True)
# Generate prediction
prediction_list = []
for _, file_1_name, file_2_name in testing_file_content:
file_1_feature = testing_image_feature_dict[file_1_name]
file_2_feature = testing_image_feature_dict[file_2_name]
final_feature = solution_basic.get_final_feature(
file_1_feature, file_2_feature, metric_list)
final_feature = final_feature.reshape(1, -1)
probability_estimates = classifier.predict_proba(final_feature)
prediction = probability_estimates[0, 1]
prediction_list.append(prediction)
# Update progress bar
progress_bar.update()
# Report tracking information
print(progress_bar)
# Write prediction
prediction_file_name = prediction_file_prefix + model_name + "_" + str(
int(time.time())) + ".csv"
solution_basic.write_prediction(testing_file_content,
np.array(prediction_list),
prediction_file_name)
def perform_training(image_feature_list, image_index_list, description,
feature_extension):
"""Perform training phase.
:param image_feature_list: the features of the images
:type image_feature_list: list
:param image_index_list: the indexes of the images
:type image_index_list: list
:param description: the folder name of the working directory
:type description: string
:param feature_extension: the extension of the feature files
:type feature_extension: string
:return: the model files will be saved to disk
:rtype: None
"""
print("Performing training phase ...")
# Reset the working directory
common.reset_working_directory(description)
working_directory = common.get_working_directory(description)
# Cross Validation
fold_num = 5
best_score_array = np.zeros(fold_num)
label_kfold = LabelKFold(image_index_list, n_folds=fold_num)
# Add progress bar
progress_bar = pyprind.ProgBar(fold_num, monitor=True)
metric_list = METRIC_LIST_DICT[feature_extension]
for fold_index, fold_item in enumerate(label_kfold):
print("\nWorking on the {:d}/{:d} fold ...".format(
fold_index + 1, fold_num))
# Generate final data set
X_train, Y_train = solution_basic.convert_to_final_data_set(
image_feature_list, image_index_list, fold_item[0], 1, metric_list)
X_test, Y_test = solution_basic.convert_to_final_data_set(
image_feature_list, image_index_list, fold_item[1], None,
metric_list)
# Perform training
model_name = "Model_{:d}".format(fold_index +
1) + common.SCIKIT_LEARN_EXTENSION
model_path = os.path.join(working_directory, model_name)
best_score = sklearn_related.train_model(X_train, Y_train, X_test,
Y_test, model_path)
best_score_array[fold_index] = best_score
print(
"For the {:d} fold, the sklearn model achieved the score {:.4f}.".
format(fold_index + 1, best_score))
# Update progress bar
progress_bar.update()
# Report tracking information
print(progress_bar)
print("\nThe best score is {:.4f}.".format(np.max(best_score_array)))