Python normalize_training_dataset Beispiele

Beispiel #1

Datei: main_RNN_LSTM_tensorflow.py Projekt: antkillerfarm/antkillerfarm_crazy

ALLOWED_LABELS_MAP = {}
for i in range(0, len(ALLOWED_LABELS)):
    ALLOWED_LABELS_MAP[str(i)] = ALLOWED_LABELS[i]

print(time.asctime(time.localtime(time.time())))
dataset_train_features, dataset_train_labels, labels_one_hot_map, dataset_train_filenames =\
	get_audio_dataset_features_labels(DATASET_PATH, ALLOWED_LABELS, type='train')
audio_filenames = get_audio_test_dataset_filenames(DATASET_PATH)

print('dataset_train_features.shape:', dataset_train_features.shape,
	  'dataset_train_labels.shape:', dataset_train_labels.shape)

# normalize training and testing features dataset
print('Normalizing datasets')
dataset_train_features, min_value, max_value \
	= normalize_training_dataset(dataset_train_features)

# randomize shuffle
#print('Shuffling training dataset')
#dataset_train_features, dataset_train_labels \
#	= shuffle_randomize(dataset_train_features, dataset_train_labels)

# divide training set into training and validation
train_len = 22000
dataset_validation_features, dataset_validation_labels \
	= dataset_train_features[train_len:dataset_train_features.shape[0], :],\
	  dataset_train_labels[train_len:dataset_train_labels.shape[0], :]
dataset_validation_filenames = []
for i in range(train_len, len(dataset_train_filenames)):
    dataset_validation_filenames.append(dataset_train_filenames[i])
dataset_train_features, dataset_train_labels \

Beispiel #2

Datei: main_RNN_LSTM_tensorflow.py Projekt: SeongJunChoi/TensorFlow_Speech_Recognition_Challenge

    'silence', 'unknown'
]
ALLOWED_LABELS_MAP = {}
for i in range(0, len(ALLOWED_LABELS)):
    ALLOWED_LABELS_MAP[str(i)] = ALLOWED_LABELS[i]

dataset_train_features, dataset_train_labels, labels_one_hot_map = get_audio_dataset_features_labels(
    DATASET_PATH, ALLOWED_LABELS, type='train')
audio_filenames = get_audio_test_dataset_filenames(DATASET_PATH)

print('dataset_train_features.shape:', dataset_train_features.shape,
      'dataset_train_labels.shape:', dataset_train_labels.shape)

# normalize training and testing features dataset
print('Normalizing datasets')
dataset_train_features, min_value, max_value = normalize_training_dataset(
    dataset_train_features)

# randomize shuffle
print('Shuffling training dataset')
dataset_train_features, dataset_train_labels = shuffle_randomize(
    dataset_train_features, dataset_train_labels)

# divide training set into training and validation
dataset_validation_features, dataset_validation_labels = dataset_train_features[
    57000:dataset_train_features.shape[0], :], dataset_train_labels[
        57000:dataset_train_labels.shape[0], :]
dataset_train_features, dataset_train_labels = dataset_train_features[
    0:57000, :], dataset_train_labels[0:57000, :]
print('dataset_validation_features.shape:', dataset_validation_features.shape,
      'dataset_validation_labels.shape:', dataset_validation_labels.shape)

Beispiel #3

]
ALLOWED_LABELS_MAP = {}
for i in range(0, len(ALLOWED_LABELS)):
    ALLOWED_LABELS_MAP[str(i)] = ALLOWED_LABELS[i]

dataset_train_features, dataset_train_labels, labels_one_hot_map, dataset_train_filenames =\
 get_audio_dataset_features_labels(DATASET_PATH, ALLOWED_LABELS, type='train')
audio_filenames = get_audio_test_dataset_filenames(DATASET_PATH)

print('dataset_train_features.shape:', dataset_train_features.shape,
      'dataset_train_labels.shape:', dataset_train_labels.shape)

# normalize training and testing features dataset
print('Normalizing datasets')
dataset_train_features, min_value, max_value \
 = normalize_training_dataset(dataset_train_features)

# randomize shuffle
#print('Shuffling training dataset')
#dataset_train_features, dataset_train_labels \
#	= shuffle_randomize(dataset_train_features, dataset_train_labels)

# divide training set into training and validation
train_len = 3500
dataset_validation_features, dataset_validation_labels \
 = dataset_train_features[train_len:dataset_train_features.shape[0], :],\
   dataset_train_labels[train_len:dataset_train_labels.shape[0], :]
dataset_validation_filenames = []
for i in range(train_len, len(dataset_train_filenames)):
    dataset_validation_filenames.append(dataset_train_filenames[i])
dataset_train_features, dataset_train_labels \