Question

我有500个RGB彩色图像标签（200x300像素）未标记，用于无监督学习（CNN，GAN，自动编码器）。我想将图像集导入到tensorflow，而不是MNIST示例。我需要将它们转换为CSV文件吗？

import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt

from tensorflow.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets("./", one_hot=True)

Answer 1

这是我训练GAN的代码的一部分。像这样读取文件。这是做到这一点的一种方法。

filenames = tf.train.string_input_producer(
    tf.train.match_filenames_once("D:/TensorFlow/resizedimages/*.png"))

参考代码是这个。

def train():
    filenames = tf.train.string_input_producer(
        tf.train.match_filenames_once("D:/TensorFlow/resizedimages/*.png"))
    reader = tf.WholeFileReader()
    _, input = reader.read(filenames)
    input = tf.image.decode_png(input, channels=3)
    input.set_shape([299, 299, 3])

    batch = tf.train.batch([input],
                           batch_size=30)

    init = (tf.global_variables_initializer(), tf.local_variables_initializer())
    with tf.Session() as sess:
        sess.run(init)
        coord = tf.train.Coordinator()
        threads = tf.train.start_queue_runners(coord=coord)
        train_writer = tf.summary.FileWriter('D:/TensorFlow/logs/1/train', sess.graph)
        tf.summary.image("Image", GeneratedImage)
        merge = tf.summary.merge_all()

        for it in range(100):
            _, X_batch =  sess.run([input,batch])
            summary,_ = sess.run([merge,D_optimizer], feed_dict={Z : samplefromuniformdistribution(20,100), X: X_batch, keep_prob: keep_prob_value})
            summary,_ = sess.run([merge,G_optimizer],feed_dict={ Z : samplefromuniformdistribution(20,100), X: X_batch, keep_prob: keep_prob_value})

            train_writer.add_summary(summary, it)
            train_writer.flush()

        train_writer.close()
        coord.request_stop()
        coord.join(threads)

对于无监督学习，如何生成图像集

1 个答案: