遷移學(xué)習(xí)與retrain.py是一樣的

這是TensorFlow 實(shí)戰(zhàn)Google深度學(xué)習(xí)框架 ,鄭澤宇書(shū)上的代碼：其也是改編了retrain.py的代碼
這里和retrain.py是一樣的，因?yàn)閜b文件的參數(shù)無(wú)法繼續(xù)進(jìn)行訓(xùn)練了，所以取出pb文件模型圖里的頭和尾（頭是'DecodeJpeg/contents:0'，把圖片編碼成二進(jìn)制文件，在model里面會(huì)進(jìn)行剪裁，此尾是sofmax前面某一conv層，pool_3/_reshape:0），把圖片輸入inception_model后生成一個(gè)特征，作為新模型的輸入，這里BOTTLENECK_TENSOR_SIZE = 2048說(shuō)明pool_3/_reshape:0之后特征的shape為2048，新模型的輸出為n_classes，與你的具體任務(wù)有關(guān)，這flower有四類，所以為n_classes = 4，這里 n_classes = len(image_lists.keys())，看你在flower_potos文件夾建立幾個(gè)子個(gè)文件夾，flower_potos官方的文件有這四類

開(kāi)始上代碼

import glob
import os.path
import random
import numpy as np
import tensorflow as tf
from tensorflow.python.platform import gfile


# #### 1. 模型和樣本路徑的設(shè)置

# In[2]:

BOTTLENECK_TENSOR_SIZE = 2048
BOTTLENECK_TENSOR_NAME = 'pool_3/_reshape:0'
JPEG_DATA_TENSOR_NAME = 'DecodeJpeg/contents:0'


MODEL_DIR = 'inception_model'
MODEL_FILE= 'classify_image_graph_def.pb'

CACHE_DIR = 'bottleneck'
INPUT_DATA = 'flower_photos'

VALIDATION_PERCENTAGE = 10
TEST_PERCENTAGE = 10


# #### 2. 神經(jīng)網(wǎng)絡(luò)參數(shù)的設(shè)置

# In[3]:

LEARNING_RATE = 0.01
STEPS = 4000
BATCH = 100


# #### 3. 把樣本中所有的圖片列表并按訓(xùn)練、驗(yàn)證、測(cè)試數(shù)據(jù)分開(kāi)

# In[4]:

def create_image_lists(testing_percentage, validation_percentage):

    result = {}
    sub_dirs = [x[0] for x in os.walk(INPUT_DATA)]
    is_root_dir = True
    for sub_dir in sub_dirs:
        if is_root_dir:
            is_root_dir = False
            continue

        extensions = ['jpg', 'jpeg', 'JPG', 'JPEG']

        file_list = []
        dir_name = os.path.basename(sub_dir)
        for extension in extensions:
            file_glob = os.path.join(INPUT_DATA, dir_name, '*.' + extension)
            file_list.extend(glob.glob(file_glob))
        if not file_list: continue

        label_name = dir_name.lower()
        
        # 初始化
        training_images = []
        testing_images = []
        validation_images = []
        for file_name in file_list:
            base_name = os.path.basename(file_name)
            
            # 隨機(jī)劃分?jǐn)?shù)據(jù)
            chance = np.random.randint(100)
            if chance < validation_percentage:
                validation_images.append(base_name)
            elif chance < (testing_percentage + validation_percentage):
                testing_images.append(base_name)
            else:
                training_images.append(base_name)

        result[label_name] = {
            'dir': dir_name,
            'training': training_images,
            'testing': testing_images,
            'validation': validation_images,
            }
    return result


# #### 4. 定義函數(shù)通過(guò)類別名稱、所屬數(shù)據(jù)集和圖片編號(hào)獲取一張圖片的地址。

# In[5]:

def get_image_path(image_lists, image_dir, label_name, index, category):
    label_lists = image_lists[label_name]
    category_list = label_lists[category]
    mod_index = index % len(category_list)
    base_name = category_list[mod_index]
    sub_dir = label_lists['dir']
    full_path = os.path.join(image_dir, sub_dir, base_name)
    return full_path


# #### 5. 定義函數(shù)獲取Inception-v3模型處理之后的特征向量的文件地址。

# In[6]:

def get_bottleneck_path(image_lists, label_name, index, category):
    return get_image_path(image_lists, CACHE_DIR, label_name, index, category) + '.txt'


# #### 6. 定義函數(shù)使用加載的訓(xùn)練好的Inception-v3模型處理一張圖片，得到這個(gè)圖片的特征向量。

# In[7]:

def run_bottleneck_on_image(sess, image_data, image_data_tensor, bottleneck_tensor):

    bottleneck_values = sess.run(bottleneck_tensor, {image_data_tensor: image_data})

    bottleneck_values = np.squeeze(bottleneck_values)
    return bottleneck_values


# #### 7. 定義函數(shù)會(huì)先試圖尋找已經(jīng)計(jì)算且保存下來(lái)的特征向量，如果找不到則先計(jì)算這個(gè)特征向量，然后保存到文件。

# In[8]:

def get_or_create_bottleneck(sess, image_lists, label_name, index, category, jpeg_data_tensor, bottleneck_tensor):
    label_lists = image_lists[label_name]
    sub_dir = label_lists['dir']
    sub_dir_path = os.path.join(CACHE_DIR, sub_dir)
    if not os.path.exists(sub_dir_path): os.makedirs(sub_dir_path)
    bottleneck_path = get_bottleneck_path(image_lists, label_name, index, category)
    if not os.path.exists(bottleneck_path):

        image_path = get_image_path(image_lists, INPUT_DATA, label_name, index, category)

        image_data = gfile.FastGFile(image_path, 'rb').read()

        bottleneck_values = run_bottleneck_on_image(sess, image_data, jpeg_data_tensor, bottleneck_tensor)

        bottleneck_string = ','.join(str(x) for x in bottleneck_values)
        with open(bottleneck_path, 'w') as bottleneck_file:
            bottleneck_file.write(bottleneck_string)
    else:

        with open(bottleneck_path, 'r') as bottleneck_file:
            bottleneck_string = bottleneck_file.read()
        bottleneck_values = [float(x) for x in bottleneck_string.split(',')]

    return bottleneck_values


# #### 8. 這個(gè)函數(shù)隨機(jī)獲取一個(gè)batch的圖片作為訓(xùn)練數(shù)據(jù)。

# In[9]:

def get_random_cached_bottlenecks(sess, n_classes, image_lists, how_many, category, jpeg_data_tensor, bottleneck_tensor):
    bottlenecks = []
    ground_truths = []
    for _ in range(how_many):
        label_index = random.randrange(n_classes)
        label_name = list(image_lists.keys())[label_index]
        image_index = random.randrange(65536)
        bottleneck = get_or_create_bottleneck(
            sess, image_lists, label_name, image_index, category, jpeg_data_tensor, bottleneck_tensor)
        ground_truth = np.zeros(n_classes, dtype=np.float32)
        ground_truth[label_index] = 1.0
        bottlenecks.append(bottleneck)
        ground_truths.append(ground_truth)

    return bottlenecks, ground_truths


# #### 9. 這個(gè)函數(shù)獲取全部的測(cè)試數(shù)據(jù)，并計(jì)算正確率。

# In[10]:

def get_test_bottlenecks(sess, image_lists, n_classes, jpeg_data_tensor, bottleneck_tensor):
    bottlenecks = []
    ground_truths = []
    label_name_list = list(image_lists.keys())
    for label_index, label_name in enumerate(label_name_list):
        category = 'testing'
        for index, unused_base_name in enumerate(image_lists[label_name][category]):
            bottleneck = get_or_create_bottleneck(sess, image_lists, label_name, index, category,jpeg_data_tensor, bottleneck_tensor)
            ground_truth = np.zeros(n_classes, dtype=np.float32)
            ground_truth[label_index] = 1.0
            bottlenecks.append(bottleneck)
            ground_truths.append(ground_truth)
    return bottlenecks, ground_truths


# #### 10. 定義主函數(shù)。

# In[11]:

def main():
    image_lists = create_image_lists(TEST_PERCENTAGE, VALIDATION_PERCENTAGE)
    n_classes = len(image_lists.keys())
    
    # 讀取已經(jīng)訓(xùn)練好的Inception-v3模型。
    with gfile.FastGFile(os.path.join(MODEL_DIR, MODEL_FILE), 'rb') as f:
        graph_def = tf.GraphDef()
        graph_def.ParseFromString(f.read())
    bottleneck_tensor, jpeg_data_tensor = tf.import_graph_def(
        graph_def, return_elements=[BOTTLENECK_TENSOR_NAME, JPEG_DATA_TENSOR_NAME])

    # 定義新的神經(jīng)網(wǎng)絡(luò)輸入
    bottleneck_input = tf.placeholder(tf.float32, [None, BOTTLENECK_TENSOR_SIZE], name='BottleneckInputPlaceholder')
    ground_truth_input = tf.placeholder(tf.float32, [None, n_classes], name='GroundTruthInput')
    
    # 定義一層全鏈接層
    with tf.name_scope('final_training_ops'):
        weights = tf.Variable(tf.truncated_normal([BOTTLENECK_TENSOR_SIZE, n_classes], stddev=0.001))
        biases = tf.Variable(tf.zeros([n_classes]))
        logits = tf.matmul(bottleneck_input, weights) + biases
        final_tensor = tf.nn.softmax(logits)
        
    # 定義交叉熵?fù)p失函數(shù)。
    cross_entropy = tf.nn.softmax_cross_entropy_with_logits(logits=logits, labels=ground_truth_input)
    cross_entropy_mean = tf.reduce_mean(cross_entropy)
    train_step = tf.train.GradientDescentOptimizer(LEARNING_RATE).minimize(cross_entropy_mean)
    
    # 計(jì)算正確率。
    with tf.name_scope('evaluation'):
        correct_prediction = tf.equal(tf.argmax(final_tensor, 1), tf.argmax(ground_truth_input, 1))
        evaluation_step = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))

    with tf.Session() as sess:
        init = tf.global_variables_initializer()
        sess.run(init)
        # 訓(xùn)練過(guò)程。
        for i in range(STEPS):
 
            train_bottlenecks, train_ground_truth = get_random_cached_bottlenecks(
                sess, n_classes, image_lists, BATCH, 'training', jpeg_data_tensor, bottleneck_tensor)
            sess.run(train_step, feed_dict={bottleneck_input: train_bottlenecks, ground_truth_input: train_ground_truth})

            if i % 100 == 0 or i + 1 == STEPS:
                validation_bottlenecks, validation_ground_truth = get_random_cached_bottlenecks(
                    sess, n_classes, image_lists, BATCH, 'validation', jpeg_data_tensor, bottleneck_tensor)
                validation_accuracy = sess.run(evaluation_step, feed_dict={
                    bottleneck_input: validation_bottlenecks, ground_truth_input: validation_ground_truth})
                print('Step %d: Validation accuracy on random sampled %d examples = %.1f%%' %
                    (i, BATCH, validation_accuracy * 100))
            
        # 在最后的測(cè)試數(shù)據(jù)上測(cè)試正確率。
        test_bottlenecks, test_ground_truth = get_test_bottlenecks(
            sess, image_lists, n_classes, jpeg_data_tensor, bottleneck_tensor)
        test_accuracy = sess.run(evaluation_step, feed_dict={
            bottleneck_input: test_bottlenecks, ground_truth_input: test_ground_truth})
        print('Final test accuracy = %.1f%%' % (test_accuracy * 100))

if __name__ == '__main__':
    main()

來(lái)個(gè)圖吧