复现LLP-GAN代码（使用CIFAR-10数据集）

代码地址

LLP-GAN

CIFAR-10数据集

CIFAR-10数据集

环境准备

python3.6
tensorflow1.14
keras2.2.5
其他的不用指定版本

数据集改为本地文件

环境准备好了以后，可以直接运行good-llp-gan.py，这里代码会去下载CIFAR-10数据集，会很慢，我们可以先下来来，然后本地直接加载。

下载好cifar-10-python.tar.gz，解压，放在dataset文件夹下。

load_local_cifar10.py

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

from tensorflow.keras import backend as K
import numpy as np
import os

import sys
from six.moves import cPickle


def load_batch(fpath, label_key='labels'):
    """Internal utility for parsing CIFAR data.

    # Arguments
        fpath: path the file to parse.
        label_key: key for label data in the retrieve
            dictionary.

    # Returns
        A tuple `(data, labels)`.
    """
    with open(fpath, 'rb') as f:
        if sys.version_info < (3,):
            d = cPickle.load(f)
        else:
            d = cPickle.load(f, encoding='bytes')
            # decode utf8
            d_decoded = {}
            for k, v in d.items():
                d_decoded[k.decode('utf8')] = v
            d = d_decoded
    data = d['data']
    labels = d[label_key]

    data = data.reshape(data.shape[0], 3, 32, 32)
    return data, labels


def load_data(ROOT):
    """Loads CIFAR10 dataset.
    # Returns
        Tuple of Numpy arrays: `(x_train, y_train), (x_test, y_test)`.
    """
    # dirname = 'cifar-10-batches-py'
    # origin = 'https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz'
    # path = get_file(dirname, origin=origin, untar=True)
    path = ROOT

    num_train_samples = 50000

    x_train = np.empty((num_train_samples, 3, 32, 32), dtype='uint8')
    y_train = np.empty((num_train_samples,), dtype='uint8')

    for i in range(1, 6):
        fpath = os.path.join(path, 'data_batch_' + str(i))
        (x_train[(i - 1) * 10000: i * 10000, :, :, :],
         y_train[(i - 1) * 10000: i * 10000]) = load_batch(fpath)

    fpath = os.path.join(path, 'test_batch')
    x_test, y_test = load_batch(fpath)

    y_train = np.reshape(y_train, (len(y_train), 1))
    y_test = np.reshape(y_test, (len(y_test), 1))

    if K.image_data_format() == 'channels_last':
        x_train = x_train.transpose(0, 2, 3, 1)
        x_test = x_test.transpose(0, 2, 3, 1)

    return (x_train, y_train), (x_test, y_test)

然后修改utils.py中的load_data()函数

def load_data():
    if hp.target_dataset == "CIFAR-10":
        if os.path.exists(hp.DATASET_DIR + hp.target_dataset):
            print("load data from pickle")
            with open(hp.DATASET_DIR + hp.target_dataset + "/train_X.pkl", 'rb') as f:
                train_X = pickle.load(f)
            with open(hp.DATASET_DIR + hp.target_dataset + "/train_y.pkl", 'rb') as f:
                train_y = pickle.load(f)
            with open(hp.DATASET_DIR + hp.target_dataset + "/valid_X.pkl", 'rb') as f:
                valid_X = pickle.load(f)
            with open(hp.DATASET_DIR + hp.target_dataset + "/valid_y.pkl", 'rb') as f:
                valid_y = pickle.load(f)
            with open(hp.DATASET_DIR + hp.target_dataset + "/test_X.pkl", 'rb') as f:
                test_X = pickle.load(f)
            with open(hp.DATASET_DIR + hp.target_dataset + "/test_y.pkl", 'rb') as f:
                test_y = pickle.load(f)
        else:
            # (cifar_X_1, cifar_y_1), (cifar_X_2, cifar_y_2) = cifar10.load_data() 
            (cifar_X_1, cifar_y_1), (cifar_X_2, cifar_y_2) = load_local_cifar10.load_data('dataset/cifar-10-batches-py')
            cifar_X = np.r_[cifar_X_1, cifar_X_2]
            cifar_y = np.r_[cifar_y_1, cifar_y_2]

            cifar_X = cifar_X.astype('float32') / 255.0
            cifar_y = np.eye(10)[cifar_y.astype('int32').flatten()]

            train_X, test_X, train_y, test_y = train_test_split(cifar_X, cifar_y, test_size=5000,
                                                                random_state=hp.RANDOM_STATE)
            train_X, valid_X, train_y, valid_y = train_test_split(train_X, train_y, test_size=5000,
                                                                  random_state=hp.RANDOM_STATE)

            os.mkdir(hp.DATASET_DIR + hp.target_dataset)
            with open(hp.DATASET_DIR + hp.target_dataset + "/train_X.pkl", 'wb') as f1:
                pickle.dump(train_X, f1)
            with open(hp.DATASET_DIR + hp.target_dataset + "/train_y.pkl", 'wb') as f1:
                pickle.dump(train_y, f1)
            with open(hp.DATASET_DIR + hp.target_dataset + "/valid_X.pkl", 'wb') as f1:
                pickle.dump(valid_X, f1)
            with open(hp.DATASET_DIR + hp.target_dataset + "/valid_y.pkl", 'wb') as f1:
                pickle.dump(valid_y, f1)
            with open(hp.DATASET_DIR + hp.target_dataset + "/test_X.pkl", 'wb') as f1:
                pickle.dump(test_X, f1)
            with open(hp.DATASET_DIR + hp.target_dataset + "/test_y.pkl", 'wb') as f1:
                pickle.dump(test_y, f1)

    return train_X, train_y, valid_X, valid_y, test_X, test_y

修改good-llp-gan.py

# 注掉
# trainx, trainy, validx, validy, testx, testy  = utils.load_data()


cifar10_dir = 'dataset/cifar-10-batches-py'
trainx, trainy, validx, validy, testx, testy = utils.load_data()

然后运行，也可以使用gpu加速，只要加入这两行(tensorflow1.0)

import keras.backend.tensorflow_backend as KTF

KTF.set_session(tf.Session(config=tf.ConfigProto(device_count={'gpu': 0})))