コード例 #1
0
def main():

    tensorboard_directory = './tmp/tensorboard/001'
    tensorboard_paths = [
        r'C:\Users\parth\Documents\GitHub\Kaggle-Santander-Value-Prediction-Challenge\tmp\tensorboard\001'
    ]
    tensorboard_names = ['rmse']

    # Model Parameters
    # --------------------------------------------------------------------------

    use_dropout = False
    use_batch_norm = False

    # Dropout inputs
    #     use : to use dropout in this layer
    #     rate : dropout rate
    dropout_parameters = [{
        'use': True,
        'rate': 0.5
    }, {
        'use': True,
        'rate': 0.5
    }, {
        'use': True,
        'rate': 0.5
    }, {
        'use': True,
        'rate': 0.5
    }]

    # Fully Connected Layers unit size
    fc_parameters = [{
        'units': 5000
    }, {
        'units': 5000
    }, {
        'units': 5000
    }, {
        'units': 5000
    }]
    num_dense = len(fc_parameters)

    data_shape = [None, 4990]
    batch_size = 500
    val_size = 5000
    epochs = 100000
    learning_rate = 0.001
    session = tf.Session()

    Tensorboard.make(paths=tensorboard_paths,
                     names=tensorboard_names,
                     host='127.0.0.1',
                     port='6006',
                     output=True,
                     start=False)

    dropout_parameters = []

    model = Model(sess=session,
                  data_shape=data_shape,
                  num_classes=1,
                  num_dense=2,
                  learning_rate=learning_rate,
                  use_batch_norm=use_batch_norm,
                  use_dropout=use_dropout,
                  dropout_parameters=dropout_parameters,
                  fc_parameters=fc_parameters,
                  tensorboard_directory=tensorboard_directory)

    train_data, train_labels = get_data()
    train_data, val_data, train_labels, val_labels = train_test_split(
        train_data, train_labels, test_size=0.30)

    print('> Training Data: {} {}'.format(train_data.shape,
                                          train_labels.shape))
    print('> Val Data: {} {}'.format(val_data.shape, val_labels.shape))
    # print('> Test Data: {} {}'.format(test_data.shape, test_labels.shape))

    model.train_data(data=train_data, labels=train_labels)

    model.val_data(data=val_data, labels=val_labels)

    model.train(batch_size=batch_size, epochs=epochs)
コード例 #2
0
def train(restore=True):
    tensorboard_directory = r'./tmp/tensorboard/015'
    tensorboard_paths = [
        r'C:\Users\parth\Documents\GitHub\Facial-Recognition\tmp\tensorboard\005',
        r'C:\Users\parth\Documents\GitHub\Facial-Recognition\tmp\tensorboard\007',
        r'C:\Users\parth\Documents\GitHub\Facial-Recognition\tmp\tensorboard\008',
        r'C:\Users\parth\Documents\GitHub\Facial-Recognition\tmp\tensorboard\009',
        r'C:\Users\parth\Documents\GitHub\Facial-Recognition\tmp\tensorboard\010',
        r'C:\Users\parth\Documents\GitHub\Facial-Recognition\tmp\tensorboard\011',
        r'C:\Users\parth\Documents\GitHub\Facial-Recognition\tmp\tensorboard\012',
        r'C:\Users\parth\Documents\GitHub\Facial-Recognition\tmp\tensorboard\013',
        r'C:\Users\parth\Documents\GitHub\Facial-Recognition\tmp\tensorboard\014',
        r'C:\Users\parth\Documents\GitHub\Facial-Recognition\tmp\tensorboard\015'
    ]
    tensorboard_names = [
        'mine-64x3-64x2', 'mine-64x4-150x1', 'mine-64x3-543-62x1',
        'mine-64x3-543-62x2-dr20', 'mine-64x3-543-62x2',
        'full_image-64x3-543-62x2', 'full_image-64x3-345-62x2',
        'full_image-64x3-357-62x2', 'aug_full_image-64x3-357-62x2', 'cropped'
    ]
    Tensorboard.make(paths=tensorboard_paths,
                     names=tensorboard_names,
                     host='127.0.0.1',
                     _print=True)

    epochs = 50000
    use_batch_norm = True
    use_dropout = True

    learning_rate = 0.001

    # Conv2d inputs
    #     filters : Integer, dimensionality of the output space (ie. the number of filters in the convolution)
    #     kernel_size : An integer or tuple/list of 2 integers, specifying the height and width of the 2D convolution window
    #                   Can be a single integer to specify the same value for all spatial dimensions
    #     strides : An integer or tuple/list of 2 integers, specifying the strides of the convolution along the height and width
    #               Can be a single integer to specify the same value for all spatial dimensions
    conv2d_specifications = [[{
        'filters': 64,
        'kernel_size': [3, 3],
        'strides': (1, 1)
    }], [{
        'filters': 64,
        'kernel_size': [5, 5],
        'strides': (1, 1)
    }], [{
        'filters': 64,
        'kernel_size': [7, 7],
        'strides': (1, 1)
    }]]

    # Max Pool inputs
    #     pool_size : An integer or tuple/list of 2 integers: (pool_height, pool_width) specifying the size of the pooling window
    #                 Can be a single integer to specify the same value for all spatial dimensions
    #     strides : n integer or tuple/list of 2 integers, specifying the strides of the pooling operation
    #               Can be a single integer to specify the same value for all spatial dimensions
    max_pool_specifications = [[{
        'use': True,
        'pool_size': [3, 3],
        'strides': [1, 1]
    }], [{
        'use': True,
        'pool_size': [3, 3],
        'strides': [1, 1]
    }], [{
        'use': True,
        'pool_size': [3, 3],
        'strides': [1, 1]
    }]]

    # Dropout inputs
    #     use : to use dropout in this layer
    #     rate : dropout rate
    dropout_parameters = [{
        'use': True,
        'rate': 0.5
    }, {
        'use': True,
        'rate': 0.5
    }]

    # Fully Connected Layers unit size
    num_dense = 2
    fc_parameters = [{'units': 62}, {'units': 62}]

    # Data
    #------------------------------------------------------------------------------

    dataset = fetch_lfw_people(data_home=None,
                               resize=1.0,
                               color=True,
                               download_if_missing=True,
                               min_faces_per_person=20)

    images = dataset.images
    labels = dataset.target

    data_shape = [63, 63, 3]
    aug = Augment()
    model = Model(sess=tf.Session(),
                  data_shape=data_shape,
                  num_classes=len(set(labels)),
                  num_dense=num_dense,
                  learning_rate=learning_rate,
                  use_batch_norm=use_dropout,
                  use_dropout=use_dropout,
                  conv_parameters=conv2d_specifications,
                  max_pool_parameters=max_pool_specifications,
                  dropout_parameters=dropout_parameters,
                  fc_parameters=fc_parameters,
                  tensorboard_directory=tensorboard_directory)

    labels_encoded = np.zeros((len(labels), len(set(labels))))
    labels_encoded[np.arange(len(labels)), labels] = 1

    croped = []
    for _ in range(2):
        croped.append(aug.randomCropAll(images, 63, 63))

    # a1 = aug.augment(images=images,
    #                  operations=['flip_h', 'brightness', 'hue'],
    #                  width=63,
    #                  height=63)
    #
    # a2 = aug.augment(images=images,
    #                  operations=['flip_h', 'hue'],
    #                  width=63,
    #                  height=63)
    #
    # a3 = aug.augment(images=images,
    #                  operations=['flip_h', 'contrast', 'hue'],
    #                  width=63,
    #                  height=63)

    images_selected = np.concatenate((croped[0], croped[1]), axis=0)
    labels_selected = np.concatenate([labels_encoded, labels_encoded], axis=0)
    if type(images_selected).__module__ is not np.__name__:
        images_selected = np.array(images_selected)
    if type(labels_selected).__module__ is not np.__name__:
        labels_selected = np.array(labels_selected)

    print('> Data Shape: {}'.format(images_selected.shape))

    X_train, X_test, y_train, y_test = train_test_split(images_selected,
                                                        labels_selected,
                                                        test_size=0.30)
    X_train, X_val, y_train, y_val = train_test_split(images_selected,
                                                      labels_selected,
                                                      test_size=0.30)

    print('> Number of Classes: {}'.format(len(set(labels))))

    batch_size = 30

    model.train_data(data=X_train, labels=y_train)
    model.val_data(data=X_val, labels=y_val)
    model.test_data(data=X_test, labels=y_test)

    model.train(batch_size=batch_size,
                batch_size_val=batch_size,
                epochs=epochs,
                is_restore=restore)

    return tensorboard_directory