Python TensorFlow：在我自己的图像上训练

TensorFlow: training on my own image

提问by VICTOR

I am new to TensorFlow. I am looking for the help on the image recognition where I can train my own imagedataset.

我是 TensorFlow 的新手。我正在寻找有关图像识别的帮助，我可以在其中训练自己的图像数据集。

Is there any example for training the new dataset?

有没有训练新数据集的例子？

回答by Olivier Moindrot

If you are interested in how to input your own data in TensorFlow, you can look at this tutorial.
I've also written a guide with best practices for CS230 at Stanford here.

如果你对如何在 TensorFlow 中输入自己的数据感兴趣，可以看看这个教程。
我也写与CS230的最佳做法指南在斯坦福这里。

---

---

New answer (with tf.data) and with labels

新答案（带tf.data）和标签

With the introduction of tf.datain r1.4, we can create a batch of images without placeholders and without queues. The steps are the following:

通过引入tf.datain r1.4，我们可以创建一批没有占位符和队列的图像。步骤如下：

1. Create a list containing the filenames of the images and a corresponding list of labels
2. Create a tf.data.Datasetreading these filenames and labels
3. Preprocess the data
4. Create an iterator from the tf.data.Datasetwhich will yield the next batch

1. 创建一个包含图像文件名和相应标签列表的列表
2. 创建tf.data.Dataset读取这些文件名和标签
3. 预处理数据
4. 从中创建一个迭代器tf.data.Dataset，它将产生下一批

The code is:

代码是：

# step 1
filenames = tf.constant(['im_01.jpg', 'im_02.jpg', 'im_03.jpg', 'im_04.jpg'])
labels = tf.constant([0, 1, 0, 1])

# step 2: create a dataset returning slices of `filenames`
dataset = tf.data.Dataset.from_tensor_slices((filenames, labels))

# step 3: parse every image in the dataset using `map`
def _parse_function(filename, label):
    image_string = tf.read_file(filename)
    image_decoded = tf.image.decode_jpeg(image_string, channels=3)
    image = tf.cast(image_decoded, tf.float32)
    return image, label

dataset = dataset.map(_parse_function)
dataset = dataset.batch(2)

# step 4: create iterator and final input tensor
iterator = dataset.make_one_shot_iterator()
images, labels = iterator.get_next()

Now we can run directly sess.run([images, labels])without feeding any data through placeholders.

现在我们可以直接运行sess.run([images, labels])而无需通过占位符提供任何数据。

---

---

Old answer (with TensorFlow queues)

旧答案（使用 TensorFlow 队列）

To sum it up you have multiple steps:

总而言之，您有多个步骤：

1. Create a list of filenames (ex: the paths to your images)
2. Create a TensorFlow filename queue
3. Read and decode each image, resize them to a fixed size (necessary for batching)
4. Output a batch of these images

1. 创建文件名列表（例如：图像的路径）
2. 创建一个 TensorFlow文件名队列
3. 读取和解码每个图像，将它们调整为固定大小（批处理所需）
4. 输出一批这些图像

---

---

The simplest code would be:

最简单的代码是：

# step 1
filenames = ['im_01.jpg', 'im_02.jpg', 'im_03.jpg', 'im_04.jpg']

# step 2
filename_queue = tf.train.string_input_producer(filenames)

# step 3: read, decode and resize images
reader = tf.WholeFileReader()
filename, content = reader.read(filename_queue)
image = tf.image.decode_jpeg(content, channels=3)
image = tf.cast(image, tf.float32)
resized_image = tf.image.resize_images(image, [224, 224])

# step 4: Batching
image_batch = tf.train.batch([resized_image], batch_size=8)

回答by Madiyar

Based on @olivier-moindrot's answer, but for Tensorflow 2.0+:

基于@olivier-moindrot 的回答，但对于 Tensorflow 2.0+：

# step 1
filenames = tf.constant(['im_01.jpg', 'im_02.jpg', 'im_03.jpg', 'im_04.jpg'])
labels = tf.constant([0, 1, 0, 1])

# step 2: create a dataset returning slices of `filenames`
dataset = tf.data.Dataset.from_tensor_slices((filenames, labels))

def im_file_to_tensor(file, label):
    def _im_file_to_tensor(file, label):
        path = f"../foo/bar/{file.numpy().decode()}"
        im = tf.image.decode_jpeg(tf.io.read_file(path), channels=3)
        im = tf.cast(image_decoded, tf.float32) / 255.0
        return im, label
    return tf.py_function(_im_file_to_tensor, 
                          inp=(file, label), 
                          Tout=(tf.float32, tf.uint8))

dataset = dataset.map(im_file_to_tensor)

If you are hitting an issue similar to:

如果您遇到类似以下问题：

ValueError: Cannot take the length of Shape with unknown rank

ValueError：无法获取未知等级的形状的长度

when passing tf.data.Dataset tensors to model.fit, then take a look at https://github.com/tensorflow/tensorflow/issues/24520. A fix for the code snippet above would be:

将 tf.data.Dataset 张量传递给 model.fit 时，请查看https://github.com/tensorflow/tensorflow/issues/24520。上面代码片段的修复方法是：

def im_file_to_tensor(file, label):
    def _im_file_to_tensor(file, label):
        path = f"../foo/bar/{file.numpy().decode()}"
        im = tf.image.decode_jpeg(tf.io.read_file(path), channels=3)
        im = tf.cast(image_decoded, tf.float32) / 255.0
        return im, label

    file, label = tf.py_function(_im_file_to_tensor, 
                                 inp=(file, label), 
                                 Tout=(tf.float32, tf.uint8))
    file.set_shape([192, 192, 3])
    label.set_shape([])
    return (file, label)

回答by Tensorflow Support

2.0 Compatible Answer using Tensorflow Hub: Tensorflow Hubis a Provision/Product Offered by Tensorflow, which comprises the Models developed by Google, for Text and Image Datasets.

2.0 Compatible Answer using Tensorflow Hub：Tensorflow Hub是由 提供的一项规定/产品Tensorflow，其中包括由 Google 开发的用于文本和图像数据集的模型。

It saves Thousands of Hours of Training Time and Computational Effort, as it reuses the Existing Pre-Trained Model.

它saves Thousands of Hours of Training Time and Computational Effort，因为它重用了现有的预训练模型。

If we have an Image Dataset, we can take the Existing Pre-Trained Models from TF Hub and can adopt it to our Dataset.

如果我们有一个图像数据集，我们可以从 TF Hub 中获取现有的预训练模型，并将其应用于我们的数据集。

Code for Re-Training our Image Dataset using the Pre-Trained Model, MobileNet, is shown below:

使用预训练模型 MobileNet 重新训练我们的图像数据集的代码如下所示：

import itertools
import os

import matplotlib.pylab as plt
import numpy as np

import tensorflow as tf
import tensorflow_hub as hub

module_selection = ("mobilenet_v2_100_224", 224) #@param ["(\"mobilenet_v2_100_224\", 224)", "(\"inception_v3\", 299)"] {type:"raw", allow-input: true}
handle_base, pixels = module_selection
MODULE_HANDLE ="https://tfhub.dev/google/imagenet/{}/feature_vector/4".format(handle_base)
IMAGE_SIZE = (pixels, pixels)
print("Using {} with input size {}".format(MODULE_HANDLE, IMAGE_SIZE))

BATCH_SIZE = 32 #@param {type:"integer"}

#Here we need to Pass our Dataset

data_dir = tf.keras.utils.get_file(
    'flower_photos',
    'https://storage.googleapis.com/download.tensorflow.org/example_images/flower_photos.tgz',
    untar=True)

model = tf.keras.Sequential([
    hub.KerasLayer(MODULE_HANDLE, trainable=do_fine_tuning),
    tf.keras.layers.Dropout(rate=0.2),
    tf.keras.layers.Dense(train_generator.num_classes, activation='softmax',
                          kernel_regularizer=tf.keras.regularizers.l2(0.0001))
])
model.build((None,)+IMAGE_SIZE+(3,))
model.summary()

Complete Code for Image Retraining Tutorial can be found in this Github Link.

可以在此Github 链接中找到图像重新训练教程的完整代码。

More information about Tensorflow Hub can be found in this TF Blog.

更多关于 Tensorflow Hub 的信息可以在这个TF 博客中找到。

The Pre-Trained Modules related to Images can be found in this TF Hub Link.

与图像相关的预训练模块可以在这个TF Hub Link 中找到。

All the Pre-Trained Modules, related to Images, Text, Videos, etc.. can be found in this TF HUB Modules Link.

所有与图像、文本、视频等相关的预训练模块都可以在此TF HUB 模块链接中找到。

Finally, this is the Basic Page for Tensorflow Hub.

最后，这是Tensorflow Hub的基本页面。