Face Recognition.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 项目3：人脸识别\n",
    "\n",
    "欢迎来到机器学习工程师纳米学位的第三个项目！在此文件中，有些示例代码已经提供给你，但你还需要实现更多的功能让项目成功运行。除非有明确要求，你无须修改任何已给出的代码。每一部分都会有详细的指导，需要实现的部分也会在注释中以'TODO'标出。请仔细阅读所有的提示！\n",
    "\n",
    "除了实现代码外，你还必须回答一些与项目和你的实现有关的问题。每一个需要你回答的问题都会以'问题 X'为标题。请仔细阅读每个问题，并且在问题后的'回答问题'文字框中写出完整的答案。我们将根据你对问题的回答和撰写代码所实现的功能来对你提交的项目进行评分。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 任务介绍\n",
    "\n",
    "人脸识别是一个计算机视觉任务，任务是要通过一张带有人脸的图像，对图像中的人脸进行识别并判断是谁。关于人脸识别的任务，我们一定会用到2015年Google开发的FaceNet，这个模型由于其性能非常好而被广泛使用，并且该训练好的模型已经被开源。\n",
    "\n",
    "因此，本项目的任务将要学习人脸识别任务，在此项目中，我们将先按课程所学到的知识亲手搭建一个卷积神经网络，然后，我们将用高级的网络结构，比如ResNet50再次进行人脸识别任务，最后我们将用到预训练好的FaceNet模型。在这个过程中，我们还会用到数据增强和人脸抽取技术来提升人脸识别的精确度。\n",
    "\n",
    "在这个人脸识别项目中，我们将使用一个开源数据集[Five Celebrity Faces Dataset](kaggle.com/dansbecker/5-celebrity-faces-dataset)，这也是一个在Kaggle比赛中的一个数据集。我们也已经下载好了并放在`./5-celebrity-faces-dataset`中，数据集中包含五位名人的照片，Ben Affleck, Elton John, Jerry Seinfeld, Madonna, Mindy Kaling。文件下分`train`和`val`。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 数据准备\n",
    "我们首先要简单的观察数据，然后通过数据增强和人脸抽取技术对数据图像数据进行抽取。你需要在完成这些操作后，思考并回答相关的问题。\n",
    "\n",
    "### 显示一张图像\n",
    "\n",
    "所有 train下面的图像文件名都存入 images 列表中，并将该图像的人名按顺序存于 images_name 中"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "import cv2\n",
    "import matplotlib.pyplot as plt\n",
    "import os\n",
    "import random\n",
    "import pandas as pd\n",
    "%matplotlib inline\n",
    "\n",
    "data_root = \"./5-celebrity-faces-dataset/train/\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "import csv\n",
    "\n",
    "def read_file_log(pathName, cvsName):\n",
    "    with open(cvsName, 'w',encoding='utf-8') as f_cvs:\n",
    "        csv_writer = csv.writer(f_cvs)\n",
    "        #csv_writer.writerow(['fileNale', 'label'])\n",
    "        #csv_writer.writerow([\"file_path\", \"name\"])\n",
    "        all_dirs =  os.listdir(pathName)\n",
    "        for dir_name in all_dirs:\n",
    "            all_files =  os.listdir(pathName+dir_name)\n",
    "            for file_name in all_files:\n",
    "                child = os.path.join('%s/%s/%s' % (pathName, dir_name,file_name))\n",
    "                label = dir_name\n",
    "                ##print child.decode('gbk') # .decode('gbk')是解决中文显示乱码问题\n",
    "                #print(child,label)\n",
    "                csv_writer.writerow([child, label])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_log_file = './5-celebrity-faces-dataset/train_log.csv'\n",
    "read_file_log(data_root,train_log_file)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "def read_csv(file):\n",
    "    with open(file) as csvfile:\n",
    "        reader = csv.reader(csvfile)\n",
    "        images = []\n",
    "        images_name = []\n",
    "        for line in reader:\n",
    "            images.append(line[0])\n",
    "            images_name.append(line[1])\n",
    "    \n",
    "    return images,images_name\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# TODO: 把所有 train下面的图像文件名都存入 images 列表中，并将该图像的人名按顺序存于 images_name 中\n",
    "images,images_name = read_csv(train_log_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "./5-celebrity-faces-dataset/train//elton_john/httpssmediacacheakpinimgcomxfefdacfbfdeadajpg.jpg\n",
      "elton_john\n"
     ]
    }
   ],
   "source": [
    "print(images[0])\n",
    "print(images_name[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "从`images`中随机读取一张图像，使用 `cv2.imread`读取图像，然后使用`pyplot.imshow`显示图像。注意：你需要同时显示该图像对应的人名，以及打印该图像的`shape`。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "随机选取一张照片：\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mindy_kaling\n",
      "(345, 236, 3)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "('./5-celebrity-faces-dataset/train//mindy_kaling/httpssmediacacheakpinimgcomxbabafeacbaaacabffjpg.jpg',\n",
       " 'mindy_kaling')"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# TODO: 从images 中随机读取一张图像，并获得该图像中的人名\n",
    "from random import randrange\n",
    "def random_sample(images=images, images_name=images_name):\n",
    "    print(\"随机选取一张照片：\")\n",
    "    random_index = randrange(0,len(images))\n",
    "   \n",
    "    # TODO：从 images 和 images_name 随机读取一个图像文件路径以及该图像的人名\n",
    "    im_file, im_name = images[random_index],images_name[random_index]\n",
    "    # TODO：使用 cv2.imread 读取图像文件\n",
    "    img = cv2.imread(im_file)\n",
    "    img2=cv2.cvtColor(img, cv2.COLOR_BGR2RGB) # cv2默认为bgr顺序\n",
    "    # TODO：使用 plt.imshow 和 plt.show() 显示图像\n",
    "    #plt.imshow(img)\n",
    "    plt.imshow(img2)\n",
    "    plt.show()\n",
    "    \n",
    "    # 打印该图像的人名\n",
    "    print(im_name)\n",
    "    # 打印该图像的大小 shape\n",
    "    print(img.shape)\n",
    "    return im_file, im_name\n",
    "\n",
    "random_sample(images, images_name)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "你可以多次运行上面的代码来多观察一些人物图像，以此来对数据有一个大致的认知"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "用`cv2.imread`读取所有数据并存入`train_x`中，然后用 0,1,2,3,4 来标记 Ben Affleck, Elton John, Jerry Seinfeld, Madonna, Mindy Kaling，并将所有`images_name`数据存入`train_y`中。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_x = []\n",
    "train_y = []\n",
    "dict_name = {'ben_afflek':0,\n",
    "            'elton_john':1,\n",
    "            'jerry_seinfeld':2,\n",
    "            'madonna':3,\n",
    "            'mindy_kaling':4}\n",
    "for file,name in zip(images,images_name):\n",
    "    #print(file,name)\n",
    "    train_x.append(cv2.imread(file))\n",
    "    train_y.append(dict_name[name])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4]\n"
     ]
    }
   ],
   "source": [
    "print(train_y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 数据增强\n",
    "\n",
    "首先，我们需要打印训练集的数量。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "93\n"
     ]
    }
   ],
   "source": [
    "print(len(images))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "我们可以看到，训练集的图像数量比较少，这对模型建模并学习图像数据到人名标签的映射关系增加了难度，所以这里需要做数据增强的工作。这里有一份很不错的资料可以帮助你了解数据增强——[Data Augmentation](https://machinelearningmastery.com/how-to-configure-image-data-augmentation-when-training-deep-learning-neural-networks/)。\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "\n",
    "# TODO: 构造图像数据增强器\n",
    "\n",
    "data_gen = ImageDataGenerator(\n",
    "    rescale = .1, # TODO：随机缩放图像RGB值的倍数\n",
    "    rotation_range =0.15 , # TODO：随机旋转图像的范围\n",
    "    zoom_range = 0.1,  # TODO：随机缩放图像大小范围\n",
    "    width_shift_range = 0.2,  # TODO：随机水平方向平移图像(fraction of total width)\n",
    "    height_shift_range= 0.2,  # TODO：随机纵向平移图像(fraction of total height)\n",
    "    horizontal_flip=True,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "使用 `flow_from_directory` 遍历数据集 `./5-celebrity-faces-dataset/data`，来观察数据增强的表现。先得到一个图像迭代器，该迭代器每次都从路径里读取一个图像，并按照数据增强器的规则进行编辑图像"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 6 images belonging to 5 classes.\n",
      "['ben_afflek/httpcsvkmeuaeccjpg.jpg', 'elton_john/httpftqncomymusicLxZeltonjohnjpg.jpg', 'jerry_seinfeld/002.jpg', 'jerry_seinfeld/httpgraphicsnytimescomimagessectionmoviesfilmographyWireImagejpg.jpg', 'madonna/httpiamediaimdbcomimagesMMVBMTANDQNTAxNDVeQTJeQWpwZBbWUMDIMjQOTYVUXCRALjpg.jpg', 'mindy_kaling/httpgonetworthcomwpcontentuploadsthumbsjpg.jpg']\n"
     ]
    }
   ],
   "source": [
    "# 直接运行，得到一个图像迭代器，该迭代器每次都从路径里读取一个图像，并按照数据增强器的规则进行编辑图像\n",
    "dataflow_generator = data_gen.flow_from_directory(\n",
    "    \"./5-celebrity-faces-dataset/data\",\n",
    "    target_size=(160, 160),\n",
    "    batch_size=3,\n",
    "    color_mode='rgb',\n",
    "    class_mode='categorical')\n",
    "\n",
    "print(dataflow_generator.filenames)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['ben_afflek/httpcsvkmeuaeccjpg.jpg', 'elton_john/httpftqncomymusicLxZeltonjohnjpg.jpg', 'jerry_seinfeld/002.jpg', 'jerry_seinfeld/httpgraphicsnytimescomimagessectionmoviesfilmographyWireImagejpg.jpg', 'madonna/httpiamediaimdbcomimagesMMVBMTANDQNTAxNDVeQTJeQWpwZBbWUMDIMjQOTYVUXCRALjpg.jpg', 'mindy_kaling/httpgonetworthcomwpcontentuploadsthumbsjpg.jpg']\n",
      "6\n",
      "{'ben_afflek': 0, 'elton_john': 1, 'jerry_seinfeld': 2, 'madonna': 3, 'mindy_kaling': 4}\n"
     ]
    }
   ],
   "source": [
    "# TODO：从 迭代器中 读取10张图片，并显示图像\n",
    "from keras.preprocessing.image import array_to_img\n",
    "sample_count = 10\n",
    "i = 0\n",
    "filenames = dataflow_generator.filenames\n",
    "labels = dataflow_generator.class_indices\n",
    "print(filenames)\n",
    "print(len(filenames))\n",
    "print(labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3\n",
      "[0. 0. 0. 0. 1.]\n",
      "(160, 160, 3)\n",
      "[1. 0. 0. 0. 0.]\n",
      "(160, 160, 3)\n",
      "[0. 0. 1. 0. 0.]\n",
      "(160, 160, 3)\n",
      "3\n",
      "[0. 0. 0. 1. 0.]\n",
      "(160, 160, 3)\n",
      "[0. 1. 0. 0. 0.]\n",
      "(160, 160, 3)\n",
      "[0. 0. 1. 0. 0.]\n",
      "(160, 160, 3)\n",
      "3\n",
      "[0. 0. 1. 0. 0.]\n",
      "(160, 160, 3)\n",
      "[0. 0. 0. 0. 1.]\n",
      "(160, 160, 3)\n",
      "[0. 0. 1. 0. 0.]\n",
      "(160, 160, 3)\n",
      "3\n",
      "[0. 0. 0. 1. 0.]\n",
      "(160, 160, 3)\n",
      "[1. 0. 0. 0. 0.]\n",
      "(160, 160, 3)\n",
      "[0. 1. 0. 0. 0.]\n",
      "(160, 160, 3)\n",
      "3\n",
      "3\n",
      "3\n",
      "3\n",
      "3\n",
      "3\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 12 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#此处image_data是一个二维序列，\n",
    "#image_data[0][...]存放batch_size张图片\n",
    "#image_data[1][...]存放batch_size对应标签\n",
    "for image_data in dataflow_generator:\n",
    "    # TODO：使用 plt.imshow 和 plt.show() 显示图像\n",
    "    print(len(image_data[1]))\n",
    "    for j in range(0,len(image_data[1])):\n",
    "        if i >= 12:\n",
    "            break\n",
    "        plt.subplot(3,5,1+i)\n",
    "        image = image_data[0][j].astype('uint8')\n",
    "        #print(type(image_data))\n",
    "        plt.imshow(array_to_img(image))\n",
    "        #plt.imshow(image)\n",
    "        i += 1\n",
    "        print(image_data[1][j])          #label\n",
    "        print(image_data[0][0].shape) #image\n",
    "   \n",
    "    sample_count -= 1\n",
    "    if sample_count <= 0:\n",
    "        break"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**问题1**：观察以上人脸图像，简单说说产生的图像中存在哪些增强的部分，然后再详细阐述你对数据增强的思考，包括为什么数据增强能够帮助人脸识别？你需要参考一些论文，并列出你的引用。\n",
    "\n",
    "**问题回答：**\n",
    "\n",
    "左右平移，上下平移，旋转\n",
    "\n",
    "用CNN处理图片特征，具有平移不变形，旋转不变性。\n",
    "\n",
    "因为对同一个人拍照，可以采取不同角度和构图，所以做数据增强可以更好泛化\n",
    "\n",
    "数据增强能够补充数据数量，防止过拟合，增强模型的泛化能力。数据增强有助于产生更多的数据训练网络。增强模型的泛化的性能,使网络能够泛化到不在训练集中的图像。正如本项目的训练集的图像数量比较少，模型建模并学习就很困难。通过数据增强的方式，就可以在很少数据情况下产生足够多的数据，建立图片分类器。\n",
    "\n",
    "增强模型的泛化的性能，一般的手段有数据增强和正则化方法，而用于数据增强的一般方法有：随机裁剪、随机水平翻转、平移、旋转、增加噪音和生成网络方法等（前两个方法用的最多，也最有效）。\n",
    "\n",
    "https://nanonets.com/blog/data-augmentation-how-to-use-deep-learning-when-you-have-limited-data-part-2/\n",
    "\n",
    "https://machinelearningmastery.com/how-to-configure-image-data-augmentation-when-training-deep-learning-neural-networks/"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 人脸抽取\n",
    "\n",
    "在做人脸识别任务中，一项常用的图像数据处理的技术是人脸检测（Face Detection）。人脸检测是将输入的图片中的人脸部分自动检测出来，具体来说就是要通过预测一个矩形边界框（Bounding Box）从整个图像中定位人脸部分，这里的矩形边界框由矩形左下角坐标以及矩形高和宽来定义。人脸检测是一个比较成熟的任务，接下来在我们这个项目中，我们将使用 Multi-Task Cascaded Convolutional Neural Network，MTCNN，你也可以参考论文：[Joint Face Detection and Alignment Using Multitask Cascaded Convolutional Networks](https://arxiv.org/abs/1604.02878)，来学习人脸检测任务。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: mtcnn in /home/leon/anaconda3/lib/python3.7/site-packages (0.1.0)\r\n",
      "Requirement already satisfied: keras>=2.0.0 in /home/leon/anaconda3/lib/python3.7/site-packages (from mtcnn) (2.3.1)\r\n",
      "Requirement already satisfied: opencv-python>=4.1.0 in /home/leon/anaconda3/lib/python3.7/site-packages (from mtcnn) (4.2.0.32)\r\n",
      "Requirement already satisfied: pyyaml in /home/leon/anaconda3/lib/python3.7/site-packages (from keras>=2.0.0->mtcnn) (5.3.1)\r\n",
      "Requirement already satisfied: six>=1.9.0 in /home/leon/anaconda3/lib/python3.7/site-packages (from keras>=2.0.0->mtcnn) (1.14.0)\r\n",
      "Requirement already satisfied: scipy>=0.14 in /home/leon/anaconda3/lib/python3.7/site-packages (from keras>=2.0.0->mtcnn) (1.4.1)\r\n",
      "Requirement already satisfied: keras-applications>=1.0.6 in /home/leon/anaconda3/lib/python3.7/site-packages (from keras>=2.0.0->mtcnn) (1.0.8)\r\n",
      "Requirement already satisfied: keras-preprocessing>=1.0.5 in /home/leon/anaconda3/lib/python3.7/site-packages (from keras>=2.0.0->mtcnn) (1.1.0)\r\n",
      "Requirement already satisfied: h5py in /home/leon/anaconda3/lib/python3.7/site-packages (from keras>=2.0.0->mtcnn) (2.10.0)\r\n",
      "Requirement already satisfied: numpy>=1.9.1 in /home/leon/anaconda3/lib/python3.7/site-packages (from keras>=2.0.0->mtcnn) (1.18.1)\r\n"
     ]
    }
   ],
   "source": [
    "# 运行下面代码，安装 mtcnn\n",
    "!pip install mtcnn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 定义人脸抽取的函数\n",
    "from PIL import Image\n",
    "from mtcnn.mtcnn import MTCNN\n",
    "import numpy as np\n",
    "\n",
    "def extract_face(filename, image_size=(160, 160)):\n",
    "    # 加载图像\n",
    "    image = Image.open(filename)\n",
    "    # 转换RGB\n",
    "    image = image.convert('RGB')\n",
    "    # 转成 numpy.array 格式的数据\n",
    "    image_data = np.asarray(image)\n",
    "    # 创建一个人脸检测，\n",
    "    detector = MTCNN()\n",
    "    # 从图像中检测\n",
    "    results = detector.detect_faces(image_data)\n",
    "    # 返回的结果是图像中所有出现的人脸的矩形边界框，由于我们的图像中只有一张人脸，所所以只需要取结果中第一个\n",
    "    box_x, box_y, width, height = results[0]['box']\n",
    "    # 处理下标为负的情况\n",
    "    box_x, box_y = abs(box_x), abs(box_y)\n",
    "    box_x_up, box_y_up = box_x + width, box_y + height\n",
    "    # 获得人脸部分的数据\n",
    "    face = image_data[box_y:box_y_up, box_x:box_x_up]\n",
    "    print(\"face.shape\",face.shape)\n",
    "    \n",
    "    # TODO：把抽取出来的人脸图像 resize 至需要的图像大小，并返回numpy格式的数据\n",
    "    #face_array = cv2.resize(face,image_size,interpolation=cv2.INTER_CUBIC)\n",
    "    image = Image.fromarray(face)\n",
    "    image = image.resize(image_size)\n",
    "    face_array = np.asarray(image)\n",
    "\n",
    "    return face_array"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "随机选取一张照片：\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ben_afflek\n",
      "(297, 236, 3)\n",
      "face.shape (62, 45, 3)\n",
      "(160, 160, 3)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "ran_img_file, ran_img_name = random_sample()\n",
    "img = extract_face(ran_img_file)\n",
    "plt.imshow(img)\n",
    "#plt.show()\n",
    "print(img.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**问题2：**通过多次运行以上代码并观察人脸抽取后的图像，你认为人脸检测对人脸识别有帮助吗？为什么？你需要参考一些论文，并列出你的reference。\n",
    "\n",
    "**回答问题：**\n",
    "\n",
    "当然有帮助，只有准确检测出人脸区域，才能进一步做人脸识别\n",
    "\n",
    "从数据整理的角度，通过检测是否有图像包含人脸，可以更好的清理训练图像数据。\n",
    "\n",
    "人脸识别的所需的信息既有特征，又有结构。人类检测的矩形边界框定义人脸，其实是最基本的结构。\n",
    "\n",
    "人脸识别过程中常常要做人脸对齐（Face Alignment），而人脸检测是进行这一步骤的基础。对齐的人类图像，会降低模型识别的难度。\n",
    "\n",
    "现实场景的图像可能包括大量背景信息，而不是我们关注的人类。人类对齐可以有助于把人脸和背景分离，更好的让模型专注于人类的特征。\n",
    "\n",
    "Bruce Cheen. 人脸检测和人脸识别.https://blog.csdn.net/czp_374/article/details/81162923\n",
    "\n",
    "如何应用MTCNN和FaceNet模型实现人脸检测及识别.http://www.uml.org.cn/ai/201806124.asp"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 构造数据\n",
    "\n",
    "现在我们可以应用以上的数据增强和人脸检测技术来构造完整的数据。\n",
    "\n",
    "这里可以直接使用前面定义好的图像数据增强器 `data_gen`，然后使用 [ImageDataGenerator](https://keras.io/preprocessing/image/)中的 `random_transform` 对单个图像做随机增强操作。\n",
    "\n",
    "另外，在构造数据之前，你需要先构造一个人名到类别的映射，使得在构造数据的label的时候将string格式的人名转换为int格式的类别。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "编程练习：\n",
    "1. 你需要 构造人名字典，将 ben_afflek、elton_john、jerry_seinfeld、madonna、mindy_kaling 分别映射到 0-1-2-3-4\n",
    "2. 定义 `load_dataset`函数，遍历train或者val文件夹，读取文件夹下5个人名文件夹，以该文件夹名映射至0到4 的标签；然后分别从人名文件夹中遍历所有图像文件，读取图像，如果是train文件夹下的图像，则需要用`data_gen.random_transform`来增强图像数据，增强次数为augment_times；如果是val文件夹下的图像，则不需要进行图像增强"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "def extract_face2(detector,image, image_size=(160, 160)):\n",
    "    # 转成 numpy.array 格式的数据\n",
    "    image_data = image\n",
    "    # 创建一个人脸检测，\n",
    "    # 从图像中检测\n",
    "    results = detector.detect_faces(image_data)\n",
    "    flag = True\n",
    "    if len(results) > 0:\n",
    "        # 返回的结果是图像中所有出现的人脸的矩形边界框，由于我们的图像中只有一张人脸，所所以只需要取结果中第一个\n",
    "        box_x, box_y, width, height = results[0]['box']\n",
    "        # 处理下标为负的情况\n",
    "        box_x, box_y = abs(box_x), abs(box_y)\n",
    "        box_x_up, box_y_up = box_x + width, box_y + height\n",
    "        # 获得人脸部分的数据\n",
    "        face = image_data[box_y:box_y_up, box_x:box_x_up]\n",
    "        #print(\"face.shape\",face.shape)\n",
    "\n",
    "        # TODO：把抽取出来的人脸图像 resize 至需要的图像大小，并返回numpy格式的数据\n",
    "        flag = True\n",
    "        face_array = cv2.resize(face,image_size,interpolation=cv2.INTER_CUBIC)\n",
    "    else:\n",
    "        flag = False\n",
    "        face_array = None\n",
    "\n",
    "    return flag,face_array\n",
    "# TODO：构造人名字典，将 ben_afflek、elton_john、jerry_seinfeld、madonna、mindy_kaling 分别映射到 0-1-2-3-4\n",
    "name_dict = {'ben_afflek':0,\n",
    "            'elton_john':1,\n",
    "            'jerry_seinfeld':2,\n",
    "            'madonna':3,\n",
    "            'mindy_kaling':4}\n",
    "\n",
    "\n",
    "# TODO：定义数据加载函数，data_dir为文件路径，augment_times为数据增强次数，is_train为判断是训练集还是测试集（测试集不需要数据增强）\n",
    "def load_dataset2(data_dir = \"./5-celebrity-faces-dataset/train/\", augment_times=2, is_train=True):\n",
    "    data_x = []\n",
    "    data_y = []\n",
    "    images = []\n",
    "    labels = []\n",
    "    if is_train:\n",
    "        data_gen2 = ImageDataGenerator(rescale = .3, rotation_range =0.2 , zoom_range = 0.2,  width_shift_range = 0.2,  height_shift_range= 0.2)\n",
    "        dataflow_generator2 = data_gen2.flow_from_directory(data_dir,target_size=(160, 160),batch_size=1,color_mode='rgb',class_mode='categorical')\n",
    "        labels_dict = dataflow_generator2.class_indices\n",
    "        print(\"labels_dict\",labels_dict)\n",
    "        sample_count = len(dataflow_generator2.filenames)*augment_times\n",
    "        print('sample_count:',sample_count)\n",
    "        #filenames = dataflow_generator2.filenames\n",
    "        #labels = dataflow_generator2.class_indices\n",
    "        #print(filenames)\n",
    "        #print(labels)\n",
    "        for image_data in dataflow_generator:\n",
    "            # TODO：使用 plt.imshow 和 plt.show() 显示图像\n",
    "            #print(len(image_data[1]))\n",
    "            for j in range(0,len(image_data[1])):\n",
    "                image = image_data[0][j].astype('uint8')\n",
    "                images.append(image)\n",
    "                labels.append(image_data[1][j])\n",
    "\n",
    "            #print(image_data[1])          #label\n",
    "            #print(image_data[0][0].shape) #image\n",
    "\n",
    "            sample_count -= 1\n",
    "            if sample_count <= 0:\n",
    "                images = np.array(images)\n",
    "                labels = np.array(labels)\n",
    "                break\n",
    "    else:\n",
    "        images = []\n",
    "        labels = []\n",
    "        all_dirs =  os.listdir(data_dir)\n",
    "        print(\"all_dirs:\",all_dirs)\n",
    "        for dir_name in all_dirs:\n",
    "            all_files =  os.listdir(data_dir+dir_name)\n",
    "            for file_name in all_files:\n",
    "                im_file = os.path.join('%s/%s/%s' % (data_dir, dir_name,file_name))\n",
    "                img = Image.open(filename)\n",
    "                image = image.convert('RGB')\n",
    "                label = dir_name\n",
    "                images.append(img)\n",
    "                labels.append(name_dict[label])\n",
    "                #print(\"im_file:{},label:{}\".format(im_file,label))\n",
    "    \n",
    "        \n",
    "        images = np.array(images)\n",
    "        labels = np.eye(5)[np.array(labels)]\n",
    "\n",
    "    detector = MTCNN()\n",
    "    \n",
    "    for i in range(0,len(labels)):\n",
    "        flag,face = extract_face2(detector,images[i])\n",
    "        if flag:\n",
    "            data_x.append(face)\n",
    "            data_y.append(labels[i])\n",
    "            \n",
    "        \n",
    "    data_x =np.array(data_x)\n",
    "    data_y =np.array(data_y)\n",
    "    \n",
    "    return data_x, data_y\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "def extract_face3(detector, filename, image_size=(160, 160)):\n",
    "    # 加载图像\n",
    "    image = Image.open(filename)\n",
    "    # 转换RGB\n",
    "    image = image.convert('RGB')\n",
    "    # 转成 numpy.array 格式的数据\n",
    "    image_data = np.asarray(image)\n",
    "    \n",
    "    # 从图像中检测\n",
    "    results = detector.detect_faces(image_data)\n",
    "    # 返回的结果是图像中所有出现的人脸的矩形边界框，由于我们的图像中只有一张人脸，所所以只需要取结果中第一个\n",
    "    box_x, box_y, width, height = results[0]['box']\n",
    "    # 处理下标为负的情况\n",
    "    box_x, box_y = abs(box_x), abs(box_y)\n",
    "    box_x_up, box_y_up = box_x + width, box_y + height\n",
    "    # 获得人脸部分的数据\n",
    "    face = image_data[box_y:box_y_up, box_x:box_x_up]\n",
    "\n",
    "    # TODO：把抽取出来的人脸图像 resize 至需要的图像大小，并返回numpy格式的数据\n",
    "    image = Image.fromarray(face)\n",
    "    image = image.resize(image_size)\n",
    "    face_array = np.asarray(image)\n",
    "    \n",
    "    return face_array"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "# TODO：构造人名字典，将 ben_afflek、elton_john、jerry_seinfeld、madonna、mindy_kaling 分别映射到 0-1-2-3-4\n",
    "name_dict = {'ben_afflek':0,\n",
    "            'elton_john':1,\n",
    "            'jerry_seinfeld':2,\n",
    "            'madonna':3,\n",
    "            'mindy_kaling':4}\n",
    "\n",
    "# TODO：定义数据加载函数，data_dir为文件路径，augment_times为数据增强次数，is_train为判断是训练集还是测试集（测试集不需要数据增强）\n",
    "def load_dataset(data_dir = \"./5-celebrity-faces-dataset/train/\", augment_times=2, is_train=True):\n",
    "    data_x = []\n",
    "    data_y = []\n",
    "    detector = MTCNN()\n",
    "    # TODO:\n",
    "    for subdir in os.listdir(data_dir):\n",
    "        path = os.path.join(data_dir, subdir)\n",
    "        for filename in os.listdir(path):\n",
    "            face = extract_face3(detector,os.path.join(path, filename))\n",
    "            #print(face.shape)\n",
    "            data_x.append(face)\n",
    "            data_y.append(name_dict[subdir])\n",
    "            # 如果是测试数据，则不需要进行数据增强\n",
    "            if is_train:\n",
    "                for _ in range(augment_times):\n",
    "                    face_aug = data_gen.random_transform(face)\n",
    "                    data_x.append(face_aug)\n",
    "                    data_y.append(name_dict[subdir])\n",
    "                \n",
    "    return data_x, data_y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_x, train_y = load_dataset(\"./5-celebrity-faces-dataset/train/\", augment_times=2, is_train=True)\n",
    "test_x, test_y = load_dataset(\"./5-celebrity-faces-dataset/val/\", is_train=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 最终构造好训练和测试数据\n",
    "train_X = np.asarray(train_x)\n",
    "train_Y = np.eye(5)[np.array(train_y)] #独热编码\n",
    "test_X = np.asarray(test_x)\n",
    "test_Y = np.eye(5)[np.array(test_y)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(279, 160, 160, 3)\n",
      "(279, 5) [[0. 1. 0. 0. 0.]\n",
      " [0. 1. 0. 0. 0.]\n",
      " [0. 1. 0. 0. 0.]\n",
      " ...\n",
      " [0. 0. 0. 0. 1.]\n",
      " [0. 0. 0. 0. 1.]\n",
      " [0. 0. 0. 0. 1.]]\n"
     ]
    }
   ],
   "source": [
    "#print(train_y)\n",
    "print(train_X.shape)\n",
    "print(train_Y.shape,train_Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "index = [i for i in range(len(train_y))]    \n",
    "random.shuffle(index)   \n",
    "\n",
    "train_X = train_X[index]  \n",
    "train_Y = train_Y[index]  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "279\n",
      "(160, 160, 3)\n",
      "279\n",
      "[[1. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 1.]\n",
      " [1. 0. 0. 0. 0.]\n",
      " ...\n",
      " [0. 1. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 1.]\n",
      " [0. 0. 0. 1. 0.]]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f8f403eabd0>"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "print(len(train_X))\n",
    "print(train_X[0].shape)\n",
    "print(len(train_Y))\n",
    "print(train_Y)\n",
    "plt.imshow(train_X[10])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 构建一个卷积神经网络\n",
    "\n",
    "创建一个卷积神经网络来对人脸进行分类。在你代码块的最后，执行 `model.summary()` 来输出你模型的总结信息。\n",
    "。\n",
    "\n",
    "![Sample CNN](images/sample_cnn.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**问题3：**在下方的代码块中尝试使用 Keras 搭建卷积网络的架构，并回答相关的问题。\n",
    "\n",
    "1. 你可以尝试自己搭建一个卷积网络的模型，那么你需要回答你搭建卷积网络的具体步骤（用了哪些层）以及为什么这样搭建。\n",
    "2. 你也可以根据上图提示的步骤搭建卷积网络，那么请说明如上的架构能够在该问题上取得的表现。\n",
    "\n",
    "\n",
    "**回答问题：**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_2\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_88 (Conv2D)           (None, 158, 158, 32)      896       \n",
      "_________________________________________________________________\n",
      "max_pooling2d_45 (MaxPooling (None, 79, 79, 32)        0         \n",
      "_________________________________________________________________\n",
      "dropout_4 (Dropout)          (None, 79, 79, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_89 (Conv2D)           (None, 77, 77, 64)        18496     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_46 (MaxPooling (None, 38, 38, 64)        0         \n",
      "_________________________________________________________________\n",
      "dropout_5 (Dropout)          (None, 38, 38, 64)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_90 (Conv2D)           (None, 36, 36, 128)       73856     \n",
      "_________________________________________________________________\n",
      "global_average_pooling2d_2 ( (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dropout_6 (Dropout)          (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_51 (Dense)             (None, 5)                 645       \n",
      "=================================================================\n",
      "Total params: 93,893\n",
      "Trainable params: 93,893\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from keras.layers import Conv2D, MaxPooling2D, GlobalAveragePooling2D\n",
    "from keras.layers import Dropout, Flatten, Dense\n",
    "from keras.models import Sequential\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "### TODO: 定义你的网络架构\n",
    "\n",
    "model.add(Conv2D(filters=32, kernel_size=3, padding='valid', activation='relu', input_shape=(160, 160, 3)))\n",
    "#model.add(Conv2D(32, (3,3), input_shape=(160, 160, 3), activation=\"relu\"))\n",
    "model.add(MaxPooling2D(pool_size=2))\n",
    "#model.add(MaxPooling2D(pool_size=(2,2)))\n",
    "model.add(Dropout(0.5))\n",
    "\n",
    "model.add(Conv2D(filters=64, kernel_size=3, padding='valid', activation='relu'))\n",
    "model.add(MaxPooling2D(pool_size=2))\n",
    "\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Conv2D(filters=128, kernel_size=3, padding='valid', activation='relu'))\n",
    "\n",
    "model.add(GlobalAveragePooling2D())\n",
    "\n",
    "model.add(Dropout(0.5))\n",
    "\n",
    "model.add(Dense(5, activation='softmax'))\n",
    "\n",
    "model.summary()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "keras.callbacks.ModelCheckpoint(filepath, monitor='val_loss', verbose=0, save_best_only=False, save_weights_only=False, mode='auto', period=1)\n",
    "\n",
    "filepath 可以包括命名格式选项，可以由 epoch 的值和 logs 的键（由 on_epoch_end 参数传递）来填充。\n",
    "\n",
    "例如：如果 filepath 是 weights.{epoch:02d}-{val_loss:.2f}.hdf5， 那么模型被保存的的文件名就会有训练轮数和验证损失。\n",
    "\n",
    "参数\n",
    "\n",
    "filepath: 字符串，保存模型的路径。\n",
    "\n",
    "monitor: 被监测的数据。\n",
    "\n",
    "verbose: 详细信息模式，0 或者 1 。\n",
    "\n",
    "save_best_only: 如果 save_best_only=True， 被监测数据的最佳模型就不会被覆盖。\n",
    "\n",
    "mode: {auto, min, max} 的其中之一。 如果 save_best_only=True，那么是否覆盖保存文件的决定就取决于被监测数据的最大或者最小值。 对于 val_acc，模式就会是 max，而对于 val_loss，模式就需要是 min，等等。 在 auto 模式中，方向会自动从被监测的数据的名字中判断出来。\n",
    "\n",
    "save_weights_only: 如果 True，那么只有模型的权重会被保存 (model.save_weights(filepath))， 否则的话，整个模型会被保存 (model.save(filepath))。\n",
    "\n",
    "period: 每个检查点之间的间隔（训练轮数）。\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.callbacks import ModelCheckpoint\n",
    " \n",
    "filepath='weights.best.hdf5'\n",
    "    # 有一次提升, 则覆盖一次.\n",
    "checkpointer = ModelCheckpoint(filepath='face.weights.best.hdf5', verbose=1, save_best_only=True)\n",
    "callbacks_list = [checkpointer]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "279/279 [==============================] - 4s 14ms/step - loss: 26.0778 - accuracy: 0.2401\n",
      "Epoch 2/50\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/leon/anaconda3/lib/python3.7/site-packages/keras/callbacks/callbacks.py:707: RuntimeWarning: Can save best model only with val_loss available, skipping.\n",
      "  'skipping.' % (self.monitor), RuntimeWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "279/279 [==============================] - 3s 13ms/step - loss: 2.4940 - accuracy: 0.2616\n",
      "Epoch 3/50\n",
      "279/279 [==============================] - 4s 13ms/step - loss: 1.7938 - accuracy: 0.2581\n",
      "Epoch 4/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.5913 - accuracy: 0.2724\n",
      "Epoch 5/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.5567 - accuracy: 0.2939\n",
      "Epoch 6/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.5438 - accuracy: 0.2903\n",
      "Epoch 7/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.5417 - accuracy: 0.2939\n",
      "Epoch 8/50\n",
      "279/279 [==============================] - 4s 13ms/step - loss: 1.5788 - accuracy: 0.3154\n",
      "Epoch 9/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.4584 - accuracy: 0.3333\n",
      "Epoch 10/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.5625 - accuracy: 0.3656\n",
      "Epoch 11/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.4231 - accuracy: 0.3441 1s - loss: 1.3652 - \n",
      "Epoch 12/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.4331 - accuracy: 0.3835\n",
      "Epoch 13/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.4003 - accuracy: 0.3692\n",
      "Epoch 14/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.4405 - accuracy: 0.3978\n",
      "Epoch 15/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.3647 - accuracy: 0.4337\n",
      "Epoch 16/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.3740 - accuracy: 0.4301\n",
      "Epoch 17/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.3513 - accuracy: 0.4337\n",
      "Epoch 18/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.5335 - accuracy: 0.4480\n",
      "Epoch 19/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.2962 - accuracy: 0.4946\n",
      "Epoch 20/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.2116 - accuracy: 0.5018\n",
      "Epoch 21/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.2669 - accuracy: 0.4767\n",
      "Epoch 22/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.4022 - accuracy: 0.4839\n",
      "Epoch 23/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.2814 - accuracy: 0.4839\n",
      "Epoch 24/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.2838 - accuracy: 0.4588\n",
      "Epoch 25/50\n",
      "279/279 [==============================] - 3s 13ms/step - loss: 1.1781 - accuracy: 0.5556\n",
      "Epoch 26/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.2070 - accuracy: 0.5233\n",
      "Epoch 27/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.1924 - accuracy: 0.5305\n",
      "Epoch 28/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.1622 - accuracy: 0.5556\n",
      "Epoch 29/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.0972 - accuracy: 0.5663\n",
      "Epoch 30/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.0933 - accuracy: 0.5520\n",
      "Epoch 31/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.0730 - accuracy: 0.5806\n",
      "Epoch 32/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.1130 - accuracy: 0.5699\n",
      "Epoch 33/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 0.9869 - accuracy: 0.6129\n",
      "Epoch 34/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.0865 - accuracy: 0.5986\n",
      "Epoch 35/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.1155 - accuracy: 0.5663\n",
      "Epoch 36/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 0.9479 - accuracy: 0.6237\n",
      "Epoch 37/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.0270 - accuracy: 0.5735\n",
      "Epoch 38/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.0286 - accuracy: 0.5986\n",
      "Epoch 39/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 0.9911 - accuracy: 0.6022\n",
      "Epoch 40/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.0030 - accuracy: 0.6165\n",
      "Epoch 41/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 0.9127 - accuracy: 0.6452\n",
      "Epoch 42/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 0.9546 - accuracy: 0.6201\n",
      "Epoch 43/50\n",
      "279/279 [==============================] - 3s 13ms/step - loss: 0.9753 - accuracy: 0.6380\n",
      "Epoch 44/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 0.8964 - accuracy: 0.6595\n",
      "Epoch 45/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.0048 - accuracy: 0.6308\n",
      "Epoch 46/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 0.9374 - accuracy: 0.6344\n",
      "Epoch 47/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 1.0781 - accuracy: 0.5771\n",
      "Epoch 48/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 0.8160 - accuracy: 0.6810\n",
      "Epoch 49/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 0.8886 - accuracy: 0.6416\n",
      "Epoch 50/50\n",
      "279/279 [==============================] - 3s 12ms/step - loss: 0.8515 - accuracy: 0.6595\n"
     ]
    }
   ],
   "source": [
    "# 直接运行编译模型和训练模型\n",
    "\n",
    "# 编译模型\n",
    "model.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "# 模型训练\n",
    "history1 = model.fit(train_X, train_Y, batch_size=16, epochs=50,callbacks=callbacks_list)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 模型测试\n",
    "\n",
    "你需要编写一个自动测试模型准确率的函数。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Simple CNN Accuracy: 0.560\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import fbeta_score, accuracy_score\n",
    "from sklearn.metrics import classification_report\n",
    "\n",
    "def metric_accuracy(model, test_X, test_Y, model_name):\n",
    "    preds_Y = model.predict(test_X)\n",
    "    correct = 0.\n",
    "    for pr, y in zip(preds_Y, test_Y):\n",
    "        pr_cls = np.argmax(pr)\n",
    "        if y[pr_cls] == 1:\n",
    "            correct += 1\n",
    "    accuracy = correct / len(preds_Y)\n",
    "    \n",
    "    print()\n",
    "    \n",
    "    print(\"%s Accuracy: %.3f\" % (model_name, accuracy))\n",
    "\n",
    "\n",
    "def metric_accuracy2(model, test_X, test_Y, model_name):\n",
    "    preds_Y = model.predict(test_X)\n",
    "    #TODO：通过预测值preds_Y以及真实值test_Y，来计算准确率\n",
    "    #print(preds_Y)\n",
    "    max_index = np.argmax(preds_Y,axis=1) #横轴比较\n",
    "    #print(len(max_index),max_index)\n",
    "\n",
    "    preds_Y = np.zeros( preds_Y.shape)\n",
    "    for i in range(0,len(max_index)):\n",
    "        preds_Y[i][max_index[i]] = 1\n",
    "\n",
    "    #print(len(preds_Y),preds_Y)\n",
    "    print(preds_Y.shape)\n",
    "    print(preds_Y[0])\n",
    "    print(\"=\"*88)\n",
    "    #print(len(test_Y),test_Y)\n",
    "    print(test_Y.shape)\n",
    "    print(test_Y[0])\n",
    "    \n",
    "    accuracy = accuracy_score(preds_Y,test_Y)\n",
    "    #classification_report(test_Y,preds_Y)\n",
    "    print()\n",
    "    \n",
    "    print(\"%s Accuracy: %.3f\" % (model_name, accuracy))\n",
    "    \n",
    "metric_accuracy(model, test_X, test_Y, \"Simple CNN\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(279, 5)\n",
      "[0. 0. 0. 0. 1.]\n",
      "========================================================================================\n",
      "(279, 5)\n",
      "[1. 0. 0. 0. 0.]\n",
      "\n",
      "Simple CNN Accuracy: 0.591\n"
     ]
    }
   ],
   "source": [
    "metric_accuracy2(model, train_X, train_Y, \"Simple CNN\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 进阶 CNN 模型架构\n",
    "在计算机视觉任务中，有一些复杂的高级CNN模型架构，比如ResNet、VGG、Inception 等等，他们能够对图像有一个非常好的表达。并且，已经有人把这些模型在非常大的图像数据上训练好了参数，这使得预训练的大模型能够对图像有一个很好的特征表达。这种在大规模图像数据上学到的图像特征，能够迁移到人脸图像的特征表示。\n",
    "\n",
    "在这一小节，我们利用预训练好的 ResNet50,抽取图像特征，然后再去做人脸识别。虽然 ResNet50 在各种图像上面进行预训练的，但是该模型对图像结构特征信息的学习也能够帮助人脸识别任务中的预测。\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/leon/anaconda3/lib/python3.7/site-packages/keras_applications/resnet50.py:265: UserWarning: The output shape of `ResNet50(include_top=False)` has been changed since Keras 2.2.0.\n",
      "  warnings.warn('The output shape of `ResNet50(include_top=False)` '\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_22\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_22 (InputLayer)           (None, 160, 160, 3)  0                                            \n",
      "__________________________________________________________________________________________________\n",
      "conv1_pad (ZeroPadding2D)       (None, 166, 166, 3)  0           input_22[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "conv1 (Conv2D)                  (None, 80, 80, 64)   9472        conv1_pad[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "bn_conv1 (BatchNormalization)   (None, 80, 80, 64)   256         conv1[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "activation_1 (Activation)       (None, 80, 80, 64)   0           bn_conv1[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "pool1_pad (ZeroPadding2D)       (None, 82, 82, 64)   0           activation_1[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_47 (MaxPooling2D) (None, 40, 40, 64)   0           pool1_pad[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "res2a_branch2a (Conv2D)         (None, 40, 40, 64)   4160        max_pooling2d_47[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "bn2a_branch2a (BatchNormalizati (None, 40, 40, 64)   256         res2a_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_2 (Activation)       (None, 40, 40, 64)   0           bn2a_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res2a_branch2b (Conv2D)         (None, 40, 40, 64)   36928       activation_2[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "bn2a_branch2b (BatchNormalizati (None, 40, 40, 64)   256         res2a_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_3 (Activation)       (None, 40, 40, 64)   0           bn2a_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res2a_branch2c (Conv2D)         (None, 40, 40, 256)  16640       activation_3[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "res2a_branch1 (Conv2D)          (None, 40, 40, 256)  16640       max_pooling2d_47[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "bn2a_branch2c (BatchNormalizati (None, 40, 40, 256)  1024        res2a_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "bn2a_branch1 (BatchNormalizatio (None, 40, 40, 256)  1024        res2a_branch1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "add_1 (Add)                     (None, 40, 40, 256)  0           bn2a_branch2c[0][0]              \n",
      "                                                                 bn2a_branch1[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "activation_4 (Activation)       (None, 40, 40, 256)  0           add_1[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "res2b_branch2a (Conv2D)         (None, 40, 40, 64)   16448       activation_4[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "bn2b_branch2a (BatchNormalizati (None, 40, 40, 64)   256         res2b_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_5 (Activation)       (None, 40, 40, 64)   0           bn2b_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res2b_branch2b (Conv2D)         (None, 40, 40, 64)   36928       activation_5[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "bn2b_branch2b (BatchNormalizati (None, 40, 40, 64)   256         res2b_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_6 (Activation)       (None, 40, 40, 64)   0           bn2b_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res2b_branch2c (Conv2D)         (None, 40, 40, 256)  16640       activation_6[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "bn2b_branch2c (BatchNormalizati (None, 40, 40, 256)  1024        res2b_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_2 (Add)                     (None, 40, 40, 256)  0           bn2b_branch2c[0][0]              \n",
      "                                                                 activation_4[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "activation_7 (Activation)       (None, 40, 40, 256)  0           add_2[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "res2c_branch2a (Conv2D)         (None, 40, 40, 64)   16448       activation_7[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "bn2c_branch2a (BatchNormalizati (None, 40, 40, 64)   256         res2c_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_8 (Activation)       (None, 40, 40, 64)   0           bn2c_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res2c_branch2b (Conv2D)         (None, 40, 40, 64)   36928       activation_8[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "bn2c_branch2b (BatchNormalizati (None, 40, 40, 64)   256         res2c_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_9 (Activation)       (None, 40, 40, 64)   0           bn2c_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res2c_branch2c (Conv2D)         (None, 40, 40, 256)  16640       activation_9[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "bn2c_branch2c (BatchNormalizati (None, 40, 40, 256)  1024        res2c_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_3 (Add)                     (None, 40, 40, 256)  0           bn2c_branch2c[0][0]              \n",
      "                                                                 activation_7[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "activation_10 (Activation)      (None, 40, 40, 256)  0           add_3[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "res3a_branch2a (Conv2D)         (None, 20, 20, 128)  32896       activation_10[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3a_branch2a (BatchNormalizati (None, 20, 20, 128)  512         res3a_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_11 (Activation)      (None, 20, 20, 128)  0           bn3a_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res3a_branch2b (Conv2D)         (None, 20, 20, 128)  147584      activation_11[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3a_branch2b (BatchNormalizati (None, 20, 20, 128)  512         res3a_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_12 (Activation)      (None, 20, 20, 128)  0           bn3a_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res3a_branch2c (Conv2D)         (None, 20, 20, 512)  66048       activation_12[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res3a_branch1 (Conv2D)          (None, 20, 20, 512)  131584      activation_10[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3a_branch2c (BatchNormalizati (None, 20, 20, 512)  2048        res3a_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "bn3a_branch1 (BatchNormalizatio (None, 20, 20, 512)  2048        res3a_branch1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "add_4 (Add)                     (None, 20, 20, 512)  0           bn3a_branch2c[0][0]              \n",
      "                                                                 bn3a_branch1[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "activation_13 (Activation)      (None, 20, 20, 512)  0           add_4[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "res3b_branch2a (Conv2D)         (None, 20, 20, 128)  65664       activation_13[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3b_branch2a (BatchNormalizati (None, 20, 20, 128)  512         res3b_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_14 (Activation)      (None, 20, 20, 128)  0           bn3b_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res3b_branch2b (Conv2D)         (None, 20, 20, 128)  147584      activation_14[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3b_branch2b (BatchNormalizati (None, 20, 20, 128)  512         res3b_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_15 (Activation)      (None, 20, 20, 128)  0           bn3b_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res3b_branch2c (Conv2D)         (None, 20, 20, 512)  66048       activation_15[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3b_branch2c (BatchNormalizati (None, 20, 20, 512)  2048        res3b_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_5 (Add)                     (None, 20, 20, 512)  0           bn3b_branch2c[0][0]              \n",
      "                                                                 activation_13[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_16 (Activation)      (None, 20, 20, 512)  0           add_5[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "res3c_branch2a (Conv2D)         (None, 20, 20, 128)  65664       activation_16[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3c_branch2a (BatchNormalizati (None, 20, 20, 128)  512         res3c_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_17 (Activation)      (None, 20, 20, 128)  0           bn3c_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res3c_branch2b (Conv2D)         (None, 20, 20, 128)  147584      activation_17[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3c_branch2b (BatchNormalizati (None, 20, 20, 128)  512         res3c_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_18 (Activation)      (None, 20, 20, 128)  0           bn3c_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res3c_branch2c (Conv2D)         (None, 20, 20, 512)  66048       activation_18[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3c_branch2c (BatchNormalizati (None, 20, 20, 512)  2048        res3c_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_6 (Add)                     (None, 20, 20, 512)  0           bn3c_branch2c[0][0]              \n",
      "                                                                 activation_16[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_19 (Activation)      (None, 20, 20, 512)  0           add_6[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "res3d_branch2a (Conv2D)         (None, 20, 20, 128)  65664       activation_19[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3d_branch2a (BatchNormalizati (None, 20, 20, 128)  512         res3d_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_20 (Activation)      (None, 20, 20, 128)  0           bn3d_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res3d_branch2b (Conv2D)         (None, 20, 20, 128)  147584      activation_20[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3d_branch2b (BatchNormalizati (None, 20, 20, 128)  512         res3d_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_21 (Activation)      (None, 20, 20, 128)  0           bn3d_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res3d_branch2c (Conv2D)         (None, 20, 20, 512)  66048       activation_21[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3d_branch2c (BatchNormalizati (None, 20, 20, 512)  2048        res3d_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_7 (Add)                     (None, 20, 20, 512)  0           bn3d_branch2c[0][0]              \n",
      "                                                                 activation_19[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_22 (Activation)      (None, 20, 20, 512)  0           add_7[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "res4a_branch2a (Conv2D)         (None, 10, 10, 256)  131328      activation_22[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4a_branch2a (BatchNormalizati (None, 10, 10, 256)  1024        res4a_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_23 (Activation)      (None, 10, 10, 256)  0           bn4a_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4a_branch2b (Conv2D)         (None, 10, 10, 256)  590080      activation_23[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4a_branch2b (BatchNormalizati (None, 10, 10, 256)  1024        res4a_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_24 (Activation)      (None, 10, 10, 256)  0           bn4a_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4a_branch2c (Conv2D)         (None, 10, 10, 1024) 263168      activation_24[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4a_branch1 (Conv2D)          (None, 10, 10, 1024) 525312      activation_22[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4a_branch2c (BatchNormalizati (None, 10, 10, 1024) 4096        res4a_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "bn4a_branch1 (BatchNormalizatio (None, 10, 10, 1024) 4096        res4a_branch1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "add_8 (Add)                     (None, 10, 10, 1024) 0           bn4a_branch2c[0][0]              \n",
      "                                                                 bn4a_branch1[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "activation_25 (Activation)      (None, 10, 10, 1024) 0           add_8[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "res4b_branch2a (Conv2D)         (None, 10, 10, 256)  262400      activation_25[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4b_branch2a (BatchNormalizati (None, 10, 10, 256)  1024        res4b_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_26 (Activation)      (None, 10, 10, 256)  0           bn4b_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4b_branch2b (Conv2D)         (None, 10, 10, 256)  590080      activation_26[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4b_branch2b (BatchNormalizati (None, 10, 10, 256)  1024        res4b_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_27 (Activation)      (None, 10, 10, 256)  0           bn4b_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4b_branch2c (Conv2D)         (None, 10, 10, 1024) 263168      activation_27[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4b_branch2c (BatchNormalizati (None, 10, 10, 1024) 4096        res4b_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_9 (Add)                     (None, 10, 10, 1024) 0           bn4b_branch2c[0][0]              \n",
      "                                                                 activation_25[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_28 (Activation)      (None, 10, 10, 1024) 0           add_9[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "res4c_branch2a (Conv2D)         (None, 10, 10, 256)  262400      activation_28[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4c_branch2a (BatchNormalizati (None, 10, 10, 256)  1024        res4c_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_29 (Activation)      (None, 10, 10, 256)  0           bn4c_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4c_branch2b (Conv2D)         (None, 10, 10, 256)  590080      activation_29[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4c_branch2b (BatchNormalizati (None, 10, 10, 256)  1024        res4c_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_30 (Activation)      (None, 10, 10, 256)  0           bn4c_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4c_branch2c (Conv2D)         (None, 10, 10, 1024) 263168      activation_30[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4c_branch2c (BatchNormalizati (None, 10, 10, 1024) 4096        res4c_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_10 (Add)                    (None, 10, 10, 1024) 0           bn4c_branch2c[0][0]              \n",
      "                                                                 activation_28[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_31 (Activation)      (None, 10, 10, 1024) 0           add_10[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res4d_branch2a (Conv2D)         (None, 10, 10, 256)  262400      activation_31[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4d_branch2a (BatchNormalizati (None, 10, 10, 256)  1024        res4d_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_32 (Activation)      (None, 10, 10, 256)  0           bn4d_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4d_branch2b (Conv2D)         (None, 10, 10, 256)  590080      activation_32[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4d_branch2b (BatchNormalizati (None, 10, 10, 256)  1024        res4d_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_33 (Activation)      (None, 10, 10, 256)  0           bn4d_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4d_branch2c (Conv2D)         (None, 10, 10, 1024) 263168      activation_33[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4d_branch2c (BatchNormalizati (None, 10, 10, 1024) 4096        res4d_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_11 (Add)                    (None, 10, 10, 1024) 0           bn4d_branch2c[0][0]              \n",
      "                                                                 activation_31[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_34 (Activation)      (None, 10, 10, 1024) 0           add_11[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res4e_branch2a (Conv2D)         (None, 10, 10, 256)  262400      activation_34[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4e_branch2a (BatchNormalizati (None, 10, 10, 256)  1024        res4e_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_35 (Activation)      (None, 10, 10, 256)  0           bn4e_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4e_branch2b (Conv2D)         (None, 10, 10, 256)  590080      activation_35[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4e_branch2b (BatchNormalizati (None, 10, 10, 256)  1024        res4e_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_36 (Activation)      (None, 10, 10, 256)  0           bn4e_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4e_branch2c (Conv2D)         (None, 10, 10, 1024) 263168      activation_36[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4e_branch2c (BatchNormalizati (None, 10, 10, 1024) 4096        res4e_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_12 (Add)                    (None, 10, 10, 1024) 0           bn4e_branch2c[0][0]              \n",
      "                                                                 activation_34[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_37 (Activation)      (None, 10, 10, 1024) 0           add_12[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res4f_branch2a (Conv2D)         (None, 10, 10, 256)  262400      activation_37[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4f_branch2a (BatchNormalizati (None, 10, 10, 256)  1024        res4f_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_38 (Activation)      (None, 10, 10, 256)  0           bn4f_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4f_branch2b (Conv2D)         (None, 10, 10, 256)  590080      activation_38[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4f_branch2b (BatchNormalizati (None, 10, 10, 256)  1024        res4f_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_39 (Activation)      (None, 10, 10, 256)  0           bn4f_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4f_branch2c (Conv2D)         (None, 10, 10, 1024) 263168      activation_39[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4f_branch2c (BatchNormalizati (None, 10, 10, 1024) 4096        res4f_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_13 (Add)                    (None, 10, 10, 1024) 0           bn4f_branch2c[0][0]              \n",
      "                                                                 activation_37[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_40 (Activation)      (None, 10, 10, 1024) 0           add_13[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res5a_branch2a (Conv2D)         (None, 5, 5, 512)    524800      activation_40[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn5a_branch2a (BatchNormalizati (None, 5, 5, 512)    2048        res5a_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_41 (Activation)      (None, 5, 5, 512)    0           bn5a_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res5a_branch2b (Conv2D)         (None, 5, 5, 512)    2359808     activation_41[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn5a_branch2b (BatchNormalizati (None, 5, 5, 512)    2048        res5a_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_42 (Activation)      (None, 5, 5, 512)    0           bn5a_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res5a_branch2c (Conv2D)         (None, 5, 5, 2048)   1050624     activation_42[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res5a_branch1 (Conv2D)          (None, 5, 5, 2048)   2099200     activation_40[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn5a_branch2c (BatchNormalizati (None, 5, 5, 2048)   8192        res5a_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "bn5a_branch1 (BatchNormalizatio (None, 5, 5, 2048)   8192        res5a_branch1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "add_14 (Add)                    (None, 5, 5, 2048)   0           bn5a_branch2c[0][0]              \n",
      "                                                                 bn5a_branch1[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "activation_43 (Activation)      (None, 5, 5, 2048)   0           add_14[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res5b_branch2a (Conv2D)         (None, 5, 5, 512)    1049088     activation_43[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn5b_branch2a (BatchNormalizati (None, 5, 5, 512)    2048        res5b_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_44 (Activation)      (None, 5, 5, 512)    0           bn5b_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res5b_branch2b (Conv2D)         (None, 5, 5, 512)    2359808     activation_44[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn5b_branch2b (BatchNormalizati (None, 5, 5, 512)    2048        res5b_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_45 (Activation)      (None, 5, 5, 512)    0           bn5b_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res5b_branch2c (Conv2D)         (None, 5, 5, 2048)   1050624     activation_45[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn5b_branch2c (BatchNormalizati (None, 5, 5, 2048)   8192        res5b_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_15 (Add)                    (None, 5, 5, 2048)   0           bn5b_branch2c[0][0]              \n",
      "                                                                 activation_43[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_46 (Activation)      (None, 5, 5, 2048)   0           add_15[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res5c_branch2a (Conv2D)         (None, 5, 5, 512)    1049088     activation_46[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn5c_branch2a (BatchNormalizati (None, 5, 5, 512)    2048        res5c_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_47 (Activation)      (None, 5, 5, 512)    0           bn5c_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res5c_branch2b (Conv2D)         (None, 5, 5, 512)    2359808     activation_47[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn5c_branch2b (BatchNormalizati (None, 5, 5, 512)    2048        res5c_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_48 (Activation)      (None, 5, 5, 512)    0           bn5c_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res5c_branch2c (Conv2D)         (None, 5, 5, 2048)   1050624     activation_48[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn5c_branch2c (BatchNormalizati (None, 5, 5, 2048)   8192        res5c_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_16 (Add)                    (None, 5, 5, 2048)   0           bn5c_branch2c[0][0]              \n",
      "                                                                 activation_46[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_49 (Activation)      (None, 5, 5, 2048)   0           add_16[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "sequential_3 (Sequential)       (None, 5)            52434949    activation_49[0][0]              \n",
      "==================================================================================================\n",
      "Total params: 76,022,661\n",
      "Trainable params: 75,969,541\n",
      "Non-trainable params: 53,120\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "import keras \n",
    "from keras.models import Model, Sequential\n",
    "from keras.layers import Conv2D, MaxPooling2D, Flatten, Dense, Dropout\n",
    "\n",
    "# 模型底层使用 ResNet50 对原始图像进行建模，特征抽取\n",
    "resnet50_weights = \"./models/resnet50_weights.h5\"\n",
    "resnet = keras.applications.resnet50.ResNet50(weights=None, include_top=False, input_shape=(160, 160, 3))\n",
    "resnet.load_weights(resnet50_weights)\n",
    "\n",
    "# TODO：自己定义模型顶层，使用抽取后的特征进行人脸识别\n",
    "resnet_face = Sequential()\n",
    "resnet_face.add(Flatten(input_shape=resnet.output_shape[1:]))\n",
    "#resnet_face.add(Dense(2048, activation=\"relu\"))\n",
    "#resnet_face.add(Dropout(0.5))\n",
    "resnet_face.add(Dense(1024, activation=\"relu\"))\n",
    "resnet_face.add(Dropout(0.5))\n",
    "resnet_face.add(Dense(5, activation='softmax'))\n",
    "\n",
    "resnet_face_model = Model(inputs=resnet.input, outputs=resnet_face(resnet.output))\n",
    "resnet_face_model.summary()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "176\n",
      "<keras.engine.input_layer.InputLayer object at 0x7f8f10e0ed50>\n"
     ]
    }
   ],
   "source": [
    "print(len(resnet_face_model.layers))\n",
    "print(resnet_face_model.layers[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "for layer in resnet_face_model.layers[:10]:\n",
    "    layer.trainable = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "279/279 [==============================] - 40s 144ms/step - loss: 22.2836 - accuracy: 0.3011\n",
      "Epoch 2/50\n",
      "279/279 [==============================] - 36s 128ms/step - loss: 2.1383 - accuracy: 0.4409\n",
      "Epoch 3/50\n",
      "279/279 [==============================] - 36s 128ms/step - loss: 2.2940 - accuracy: 0.6237\n",
      "Epoch 4/50\n",
      "279/279 [==============================] - 36s 128ms/step - loss: 1.8373 - accuracy: 0.6344\n",
      "Epoch 5/50\n",
      "279/279 [==============================] - 36s 129ms/step - loss: 2.1793 - accuracy: 0.6487\n",
      "Epoch 6/50\n",
      "279/279 [==============================] - 36s 129ms/step - loss: 1.2294 - accuracy: 0.7491\n",
      "Epoch 7/50\n",
      "279/279 [==============================] - 36s 127ms/step - loss: 1.2793 - accuracy: 0.7706\n",
      "Epoch 8/50\n",
      "279/279 [==============================] - 36s 128ms/step - loss: 2.4417 - accuracy: 0.6846\n",
      "Epoch 9/50\n",
      "279/279 [==============================] - 37s 131ms/step - loss: 1.5691 - accuracy: 0.7097\n",
      "Epoch 10/50\n",
      "279/279 [==============================] - 36s 128ms/step - loss: 1.9518 - accuracy: 0.6953\n",
      "Epoch 11/50\n",
      "279/279 [==============================] - 36s 130ms/step - loss: 1.2105 - accuracy: 0.7204\n",
      "Epoch 12/50\n",
      "279/279 [==============================] - 36s 130ms/step - loss: 1.7098 - accuracy: 0.7455\n",
      "Epoch 13/50\n",
      "279/279 [==============================] - 36s 128ms/step - loss: 1.3649 - accuracy: 0.8387\n",
      "Epoch 14/50\n",
      "279/279 [==============================] - 36s 129ms/step - loss: 0.9004 - accuracy: 0.8100\n",
      "Epoch 15/50\n",
      "279/279 [==============================] - 36s 128ms/step - loss: 1.0018 - accuracy: 0.8423\n",
      "Epoch 16/50\n",
      "279/279 [==============================] - 36s 128ms/step - loss: 1.2191 - accuracy: 0.8029\n",
      "Epoch 17/50\n",
      "279/279 [==============================] - 36s 129ms/step - loss: 1.2040 - accuracy: 0.8387\n",
      "Epoch 18/50\n",
      "279/279 [==============================] - 36s 128ms/step - loss: 1.0593 - accuracy: 0.8674\n",
      "Epoch 19/50\n",
      "279/279 [==============================] - 36s 129ms/step - loss: 1.1277 - accuracy: 0.8387\n",
      "Epoch 20/50\n",
      "279/279 [==============================] - 36s 129ms/step - loss: 1.0861 - accuracy: 0.8566\n",
      "Epoch 21/50\n",
      "279/279 [==============================] - 36s 129ms/step - loss: 0.6751 - accuracy: 0.8781\n",
      "Epoch 22/50\n",
      "279/279 [==============================] - 36s 128ms/step - loss: 0.8087 - accuracy: 0.8710\n",
      "Epoch 23/50\n",
      "279/279 [==============================] - 36s 130ms/step - loss: 0.9367 - accuracy: 0.8961\n",
      "Epoch 24/50\n",
      "279/279 [==============================] - 36s 129ms/step - loss: 0.4484 - accuracy: 0.9211\n",
      "Epoch 25/50\n",
      "279/279 [==============================] - 36s 127ms/step - loss: 0.6853 - accuracy: 0.9104\n",
      "Epoch 26/50\n",
      "279/279 [==============================] - 36s 128ms/step - loss: 0.1116 - accuracy: 0.9570\n",
      "Epoch 27/50\n",
      "279/279 [==============================] - 36s 130ms/step - loss: 0.4736 - accuracy: 0.9319\n",
      "Epoch 28/50\n",
      "279/279 [==============================] - 36s 130ms/step - loss: 0.1739 - accuracy: 0.9713\n",
      "Epoch 29/50\n",
      "279/279 [==============================] - 36s 129ms/step - loss: 0.6488 - accuracy: 0.9247\n",
      "Epoch 30/50\n",
      "279/279 [==============================] - 36s 128ms/step - loss: 1.3177 - accuracy: 0.8817\n",
      "Epoch 31/50\n",
      "279/279 [==============================] - 36s 130ms/step - loss: 0.6947 - accuracy: 0.8961\n",
      "Epoch 32/50\n",
      "279/279 [==============================] - 36s 129ms/step - loss: 0.3140 - accuracy: 0.9283\n",
      "Epoch 33/50\n",
      "279/279 [==============================] - 36s 129ms/step - loss: 0.3956 - accuracy: 0.9534\n",
      "Epoch 34/50\n",
      "279/279 [==============================] - 37s 131ms/step - loss: 0.5650 - accuracy: 0.9355\n",
      "Epoch 35/50\n",
      "279/279 [==============================] - 37s 131ms/step - loss: 1.7614 - accuracy: 0.9032\n",
      "Epoch 36/50\n",
      "279/279 [==============================] - 36s 129ms/step - loss: 0.1630 - accuracy: 0.9785\n",
      "Epoch 37/50\n",
      "279/279 [==============================] - 36s 128ms/step - loss: 0.2337 - accuracy: 0.9677\n",
      "Epoch 38/50\n",
      "279/279 [==============================] - 36s 127ms/step - loss: 0.8454 - accuracy: 0.9140\n",
      "Epoch 39/50\n",
      "279/279 [==============================] - 36s 128ms/step - loss: 0.3609 - accuracy: 0.9677\n",
      "Epoch 40/50\n",
      "279/279 [==============================] - 36s 130ms/step - loss: 0.7178 - accuracy: 0.9211\n",
      "Epoch 41/50\n",
      "279/279 [==============================] - 36s 129ms/step - loss: 0.5907 - accuracy: 0.9319\n",
      "Epoch 42/50\n",
      "279/279 [==============================] - 36s 129ms/step - loss: 0.0660 - accuracy: 0.9642\n",
      "Epoch 43/50\n",
      "279/279 [==============================] - 36s 129ms/step - loss: 0.1703 - accuracy: 0.9534\n",
      "Epoch 44/50\n",
      "279/279 [==============================] - 36s 128ms/step - loss: 0.1541 - accuracy: 0.9570\n",
      "Epoch 45/50\n",
      "279/279 [==============================] - 36s 129ms/step - loss: 0.9227 - accuracy: 0.9427\n",
      "Epoch 46/50\n",
      "279/279 [==============================] - 36s 129ms/step - loss: 0.5299 - accuracy: 0.9713\n",
      "Epoch 47/50\n",
      "279/279 [==============================] - 36s 130ms/step - loss: 0.0997 - accuracy: 0.9713\n",
      "Epoch 48/50\n",
      "279/279 [==============================] - 37s 131ms/step - loss: 0.4657 - accuracy: 0.9677\n",
      "Epoch 49/50\n",
      "279/279 [==============================] - 36s 129ms/step - loss: 0.1146 - accuracy: 0.9821\n",
      "Epoch 50/50\n",
      "279/279 [==============================] - 36s 128ms/step - loss: 0.0016 - accuracy: 1.0000\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.callbacks.History at 0x7f8f12308310>"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 设置同样的训练参数，直接运行\n",
    "\n",
    "## 编译模型\n",
    "resnet_face_model.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "# 模型训练\n",
    "resnet_face_model.fit(train_X, train_Y, batch_size=16, epochs=50)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "模型测试"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "ResNet50 Accuracy: 0.440\n"
     ]
    }
   ],
   "source": [
    "# 直接运行，测试 resnet_face_model 的准确率\n",
    "metric_accuracy(resnet_face_model, test_X, test_Y, \"ResNet50\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "ResNet50 Accuracy: 0.545\n"
     ]
    }
   ],
   "source": [
    "metric_accuracy(resnet_face_model, train_X, train_Y, \"ResNet50\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**问题5：**对比 ResNet50 模型和 CNN 模型的结果，请你分析为什么 ResNet50 模型能够取得更好的结果？\n",
    "\n",
    "**回答问题：**\n",
    "\n",
    "1.ResNet50模型相比CNN更深，卷积核更多，可以采样到更多的特征。\n",
    "\n",
    "2.ResNet50模型有结果大量数据训练过的参数；"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**问题6：**上面我们使用了预训练好的 ResNet50，即`resnet.load_weights(resnet50_weights)`，那么加载预训练好的参数对该任务有帮助吗？你需要通过做对比实验，即不加载预训练好的参数，然后在下面的代码框中重新跑一遍 ResNet50 的模型，来作为对比说明加载预训练是否有帮助\n",
    "\n",
    "**回答问题：**\n",
    "\n",
    "1.预加载参数，模型收敛的更快；\n",
    "\n",
    "2.从实验结果看预加载的参数，泛化能力不如重新训练所有参数？（这点与我认知刚好相反，请老师指点）"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_23\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_23 (InputLayer)           (None, 160, 160, 3)  0                                            \n",
      "__________________________________________________________________________________________________\n",
      "conv1_pad (ZeroPadding2D)       (None, 166, 166, 3)  0           input_23[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "conv1 (Conv2D)                  (None, 80, 80, 64)   9472        conv1_pad[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "bn_conv1 (BatchNormalization)   (None, 80, 80, 64)   256         conv1[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "activation_50 (Activation)      (None, 80, 80, 64)   0           bn_conv1[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "pool1_pad (ZeroPadding2D)       (None, 82, 82, 64)   0           activation_50[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_48 (MaxPooling2D) (None, 40, 40, 64)   0           pool1_pad[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "res2a_branch2a (Conv2D)         (None, 40, 40, 64)   4160        max_pooling2d_48[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "bn2a_branch2a (BatchNormalizati (None, 40, 40, 64)   256         res2a_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_51 (Activation)      (None, 40, 40, 64)   0           bn2a_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res2a_branch2b (Conv2D)         (None, 40, 40, 64)   36928       activation_51[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn2a_branch2b (BatchNormalizati (None, 40, 40, 64)   256         res2a_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_52 (Activation)      (None, 40, 40, 64)   0           bn2a_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res2a_branch2c (Conv2D)         (None, 40, 40, 256)  16640       activation_52[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res2a_branch1 (Conv2D)          (None, 40, 40, 256)  16640       max_pooling2d_48[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "bn2a_branch2c (BatchNormalizati (None, 40, 40, 256)  1024        res2a_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "bn2a_branch1 (BatchNormalizatio (None, 40, 40, 256)  1024        res2a_branch1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "add_17 (Add)                    (None, 40, 40, 256)  0           bn2a_branch2c[0][0]              \n",
      "                                                                 bn2a_branch1[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "activation_53 (Activation)      (None, 40, 40, 256)  0           add_17[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res2b_branch2a (Conv2D)         (None, 40, 40, 64)   16448       activation_53[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn2b_branch2a (BatchNormalizati (None, 40, 40, 64)   256         res2b_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_54 (Activation)      (None, 40, 40, 64)   0           bn2b_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res2b_branch2b (Conv2D)         (None, 40, 40, 64)   36928       activation_54[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn2b_branch2b (BatchNormalizati (None, 40, 40, 64)   256         res2b_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_55 (Activation)      (None, 40, 40, 64)   0           bn2b_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res2b_branch2c (Conv2D)         (None, 40, 40, 256)  16640       activation_55[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn2b_branch2c (BatchNormalizati (None, 40, 40, 256)  1024        res2b_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_18 (Add)                    (None, 40, 40, 256)  0           bn2b_branch2c[0][0]              \n",
      "                                                                 activation_53[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_56 (Activation)      (None, 40, 40, 256)  0           add_18[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res2c_branch2a (Conv2D)         (None, 40, 40, 64)   16448       activation_56[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn2c_branch2a (BatchNormalizati (None, 40, 40, 64)   256         res2c_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_57 (Activation)      (None, 40, 40, 64)   0           bn2c_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res2c_branch2b (Conv2D)         (None, 40, 40, 64)   36928       activation_57[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn2c_branch2b (BatchNormalizati (None, 40, 40, 64)   256         res2c_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_58 (Activation)      (None, 40, 40, 64)   0           bn2c_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res2c_branch2c (Conv2D)         (None, 40, 40, 256)  16640       activation_58[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn2c_branch2c (BatchNormalizati (None, 40, 40, 256)  1024        res2c_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_19 (Add)                    (None, 40, 40, 256)  0           bn2c_branch2c[0][0]              \n",
      "                                                                 activation_56[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_59 (Activation)      (None, 40, 40, 256)  0           add_19[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res3a_branch2a (Conv2D)         (None, 20, 20, 128)  32896       activation_59[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3a_branch2a (BatchNormalizati (None, 20, 20, 128)  512         res3a_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_60 (Activation)      (None, 20, 20, 128)  0           bn3a_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res3a_branch2b (Conv2D)         (None, 20, 20, 128)  147584      activation_60[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3a_branch2b (BatchNormalizati (None, 20, 20, 128)  512         res3a_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_61 (Activation)      (None, 20, 20, 128)  0           bn3a_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res3a_branch2c (Conv2D)         (None, 20, 20, 512)  66048       activation_61[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res3a_branch1 (Conv2D)          (None, 20, 20, 512)  131584      activation_59[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3a_branch2c (BatchNormalizati (None, 20, 20, 512)  2048        res3a_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "bn3a_branch1 (BatchNormalizatio (None, 20, 20, 512)  2048        res3a_branch1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "add_20 (Add)                    (None, 20, 20, 512)  0           bn3a_branch2c[0][0]              \n",
      "                                                                 bn3a_branch1[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "activation_62 (Activation)      (None, 20, 20, 512)  0           add_20[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res3b_branch2a (Conv2D)         (None, 20, 20, 128)  65664       activation_62[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3b_branch2a (BatchNormalizati (None, 20, 20, 128)  512         res3b_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_63 (Activation)      (None, 20, 20, 128)  0           bn3b_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res3b_branch2b (Conv2D)         (None, 20, 20, 128)  147584      activation_63[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3b_branch2b (BatchNormalizati (None, 20, 20, 128)  512         res3b_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_64 (Activation)      (None, 20, 20, 128)  0           bn3b_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res3b_branch2c (Conv2D)         (None, 20, 20, 512)  66048       activation_64[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3b_branch2c (BatchNormalizati (None, 20, 20, 512)  2048        res3b_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_21 (Add)                    (None, 20, 20, 512)  0           bn3b_branch2c[0][0]              \n",
      "                                                                 activation_62[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_65 (Activation)      (None, 20, 20, 512)  0           add_21[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res3c_branch2a (Conv2D)         (None, 20, 20, 128)  65664       activation_65[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3c_branch2a (BatchNormalizati (None, 20, 20, 128)  512         res3c_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_66 (Activation)      (None, 20, 20, 128)  0           bn3c_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res3c_branch2b (Conv2D)         (None, 20, 20, 128)  147584      activation_66[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3c_branch2b (BatchNormalizati (None, 20, 20, 128)  512         res3c_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_67 (Activation)      (None, 20, 20, 128)  0           bn3c_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res3c_branch2c (Conv2D)         (None, 20, 20, 512)  66048       activation_67[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3c_branch2c (BatchNormalizati (None, 20, 20, 512)  2048        res3c_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_22 (Add)                    (None, 20, 20, 512)  0           bn3c_branch2c[0][0]              \n",
      "                                                                 activation_65[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_68 (Activation)      (None, 20, 20, 512)  0           add_22[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res3d_branch2a (Conv2D)         (None, 20, 20, 128)  65664       activation_68[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3d_branch2a (BatchNormalizati (None, 20, 20, 128)  512         res3d_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_69 (Activation)      (None, 20, 20, 128)  0           bn3d_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res3d_branch2b (Conv2D)         (None, 20, 20, 128)  147584      activation_69[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3d_branch2b (BatchNormalizati (None, 20, 20, 128)  512         res3d_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_70 (Activation)      (None, 20, 20, 128)  0           bn3d_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res3d_branch2c (Conv2D)         (None, 20, 20, 512)  66048       activation_70[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn3d_branch2c (BatchNormalizati (None, 20, 20, 512)  2048        res3d_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_23 (Add)                    (None, 20, 20, 512)  0           bn3d_branch2c[0][0]              \n",
      "                                                                 activation_68[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_71 (Activation)      (None, 20, 20, 512)  0           add_23[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res4a_branch2a (Conv2D)         (None, 10, 10, 256)  131328      activation_71[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4a_branch2a (BatchNormalizati (None, 10, 10, 256)  1024        res4a_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_72 (Activation)      (None, 10, 10, 256)  0           bn4a_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4a_branch2b (Conv2D)         (None, 10, 10, 256)  590080      activation_72[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4a_branch2b (BatchNormalizati (None, 10, 10, 256)  1024        res4a_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_73 (Activation)      (None, 10, 10, 256)  0           bn4a_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4a_branch2c (Conv2D)         (None, 10, 10, 1024) 263168      activation_73[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4a_branch1 (Conv2D)          (None, 10, 10, 1024) 525312      activation_71[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4a_branch2c (BatchNormalizati (None, 10, 10, 1024) 4096        res4a_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "bn4a_branch1 (BatchNormalizatio (None, 10, 10, 1024) 4096        res4a_branch1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "add_24 (Add)                    (None, 10, 10, 1024) 0           bn4a_branch2c[0][0]              \n",
      "                                                                 bn4a_branch1[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "activation_74 (Activation)      (None, 10, 10, 1024) 0           add_24[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res4b_branch2a (Conv2D)         (None, 10, 10, 256)  262400      activation_74[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4b_branch2a (BatchNormalizati (None, 10, 10, 256)  1024        res4b_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_75 (Activation)      (None, 10, 10, 256)  0           bn4b_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4b_branch2b (Conv2D)         (None, 10, 10, 256)  590080      activation_75[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4b_branch2b (BatchNormalizati (None, 10, 10, 256)  1024        res4b_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_76 (Activation)      (None, 10, 10, 256)  0           bn4b_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4b_branch2c (Conv2D)         (None, 10, 10, 1024) 263168      activation_76[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4b_branch2c (BatchNormalizati (None, 10, 10, 1024) 4096        res4b_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_25 (Add)                    (None, 10, 10, 1024) 0           bn4b_branch2c[0][0]              \n",
      "                                                                 activation_74[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_77 (Activation)      (None, 10, 10, 1024) 0           add_25[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res4c_branch2a (Conv2D)         (None, 10, 10, 256)  262400      activation_77[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4c_branch2a (BatchNormalizati (None, 10, 10, 256)  1024        res4c_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_78 (Activation)      (None, 10, 10, 256)  0           bn4c_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4c_branch2b (Conv2D)         (None, 10, 10, 256)  590080      activation_78[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4c_branch2b (BatchNormalizati (None, 10, 10, 256)  1024        res4c_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_79 (Activation)      (None, 10, 10, 256)  0           bn4c_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4c_branch2c (Conv2D)         (None, 10, 10, 1024) 263168      activation_79[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4c_branch2c (BatchNormalizati (None, 10, 10, 1024) 4096        res4c_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_26 (Add)                    (None, 10, 10, 1024) 0           bn4c_branch2c[0][0]              \n",
      "                                                                 activation_77[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_80 (Activation)      (None, 10, 10, 1024) 0           add_26[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res4d_branch2a (Conv2D)         (None, 10, 10, 256)  262400      activation_80[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4d_branch2a (BatchNormalizati (None, 10, 10, 256)  1024        res4d_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_81 (Activation)      (None, 10, 10, 256)  0           bn4d_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4d_branch2b (Conv2D)         (None, 10, 10, 256)  590080      activation_81[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4d_branch2b (BatchNormalizati (None, 10, 10, 256)  1024        res4d_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_82 (Activation)      (None, 10, 10, 256)  0           bn4d_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4d_branch2c (Conv2D)         (None, 10, 10, 1024) 263168      activation_82[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4d_branch2c (BatchNormalizati (None, 10, 10, 1024) 4096        res4d_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_27 (Add)                    (None, 10, 10, 1024) 0           bn4d_branch2c[0][0]              \n",
      "                                                                 activation_80[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_83 (Activation)      (None, 10, 10, 1024) 0           add_27[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res4e_branch2a (Conv2D)         (None, 10, 10, 256)  262400      activation_83[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4e_branch2a (BatchNormalizati (None, 10, 10, 256)  1024        res4e_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_84 (Activation)      (None, 10, 10, 256)  0           bn4e_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4e_branch2b (Conv2D)         (None, 10, 10, 256)  590080      activation_84[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4e_branch2b (BatchNormalizati (None, 10, 10, 256)  1024        res4e_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_85 (Activation)      (None, 10, 10, 256)  0           bn4e_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4e_branch2c (Conv2D)         (None, 10, 10, 1024) 263168      activation_85[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4e_branch2c (BatchNormalizati (None, 10, 10, 1024) 4096        res4e_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_28 (Add)                    (None, 10, 10, 1024) 0           bn4e_branch2c[0][0]              \n",
      "                                                                 activation_83[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_86 (Activation)      (None, 10, 10, 1024) 0           add_28[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res4f_branch2a (Conv2D)         (None, 10, 10, 256)  262400      activation_86[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4f_branch2a (BatchNormalizati (None, 10, 10, 256)  1024        res4f_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_87 (Activation)      (None, 10, 10, 256)  0           bn4f_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4f_branch2b (Conv2D)         (None, 10, 10, 256)  590080      activation_87[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4f_branch2b (BatchNormalizati (None, 10, 10, 256)  1024        res4f_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_88 (Activation)      (None, 10, 10, 256)  0           bn4f_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res4f_branch2c (Conv2D)         (None, 10, 10, 1024) 263168      activation_88[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn4f_branch2c (BatchNormalizati (None, 10, 10, 1024) 4096        res4f_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_29 (Add)                    (None, 10, 10, 1024) 0           bn4f_branch2c[0][0]              \n",
      "                                                                 activation_86[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_89 (Activation)      (None, 10, 10, 1024) 0           add_29[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res5a_branch2a (Conv2D)         (None, 5, 5, 512)    524800      activation_89[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn5a_branch2a (BatchNormalizati (None, 5, 5, 512)    2048        res5a_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_90 (Activation)      (None, 5, 5, 512)    0           bn5a_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res5a_branch2b (Conv2D)         (None, 5, 5, 512)    2359808     activation_90[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn5a_branch2b (BatchNormalizati (None, 5, 5, 512)    2048        res5a_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_91 (Activation)      (None, 5, 5, 512)    0           bn5a_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res5a_branch2c (Conv2D)         (None, 5, 5, 2048)   1050624     activation_91[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res5a_branch1 (Conv2D)          (None, 5, 5, 2048)   2099200     activation_89[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn5a_branch2c (BatchNormalizati (None, 5, 5, 2048)   8192        res5a_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "bn5a_branch1 (BatchNormalizatio (None, 5, 5, 2048)   8192        res5a_branch1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "add_30 (Add)                    (None, 5, 5, 2048)   0           bn5a_branch2c[0][0]              \n",
      "                                                                 bn5a_branch1[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "activation_92 (Activation)      (None, 5, 5, 2048)   0           add_30[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res5b_branch2a (Conv2D)         (None, 5, 5, 512)    1049088     activation_92[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn5b_branch2a (BatchNormalizati (None, 5, 5, 512)    2048        res5b_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_93 (Activation)      (None, 5, 5, 512)    0           bn5b_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res5b_branch2b (Conv2D)         (None, 5, 5, 512)    2359808     activation_93[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn5b_branch2b (BatchNormalizati (None, 5, 5, 512)    2048        res5b_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_94 (Activation)      (None, 5, 5, 512)    0           bn5b_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res5b_branch2c (Conv2D)         (None, 5, 5, 2048)   1050624     activation_94[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn5b_branch2c (BatchNormalizati (None, 5, 5, 2048)   8192        res5b_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_31 (Add)                    (None, 5, 5, 2048)   0           bn5b_branch2c[0][0]              \n",
      "                                                                 activation_92[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_95 (Activation)      (None, 5, 5, 2048)   0           add_31[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "res5c_branch2a (Conv2D)         (None, 5, 5, 512)    1049088     activation_95[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn5c_branch2a (BatchNormalizati (None, 5, 5, 512)    2048        res5c_branch2a[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_96 (Activation)      (None, 5, 5, 512)    0           bn5c_branch2a[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res5c_branch2b (Conv2D)         (None, 5, 5, 512)    2359808     activation_96[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn5c_branch2b (BatchNormalizati (None, 5, 5, 512)    2048        res5c_branch2b[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "activation_97 (Activation)      (None, 5, 5, 512)    0           bn5c_branch2b[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "res5c_branch2c (Conv2D)         (None, 5, 5, 2048)   1050624     activation_97[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "bn5c_branch2c (BatchNormalizati (None, 5, 5, 2048)   8192        res5c_branch2c[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_32 (Add)                    (None, 5, 5, 2048)   0           bn5c_branch2c[0][0]              \n",
      "                                                                 activation_95[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "activation_98 (Activation)      (None, 5, 5, 2048)   0           add_32[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "sequential_4 (Sequential)       (None, 5)            52434949    activation_98[0][0]              \n",
      "==================================================================================================\n",
      "Total params: 76,022,661\n",
      "Trainable params: 75,969,541\n",
      "Non-trainable params: 53,120\n",
      "__________________________________________________________________________________________________\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/leon/anaconda3/lib/python3.7/site-packages/keras_applications/resnet50.py:265: UserWarning: The output shape of `ResNet50(include_top=False)` has been changed since Keras 2.2.0.\n",
      "  warnings.warn('The output shape of `ResNet50(include_top=False)` '\n"
     ]
    }
   ],
   "source": [
    "# 重新跑一遍不加载预训练参数的 ResNet50 的模型，请在此处写完整的code\n",
    "\n",
    "import keras \n",
    "from keras.models import Model, Sequential\n",
    "from keras.layers import Conv2D, MaxPooling2D, Flatten, Dense, Dropout\n",
    "\n",
    "# 模型底层使用 ResNet50 对原始图像进行建模，特征抽取\n",
    "resnet50_weights = \"./models/resnet50_weights.h5\"\n",
    "resnet2 = keras.applications.resnet50.ResNet50(weights=None, include_top=False, input_shape=(160, 160, 3))\n",
    "#resnet.load_weights(resnet50_weights)\n",
    "\n",
    "# TODO：自己定义模型顶层，使用抽取后的特征进行人脸识别\n",
    "resnet_face2 = Sequential()\n",
    "resnet_face2.add(Flatten(input_shape=resnet2.output_shape[1:]))\n",
    "resnet_face2.add(Dense(1024, activation=\"relu\"))\n",
    "resnet_face2.add(Dropout(0.5))\n",
    "resnet_face2.add(Dense(5, activation=\"softmax\"))\n",
    "\n",
    "resnet_face_model2 = Model(inputs=resnet2.input, outputs=resnet_face2(resnet2.output))\n",
    "resnet_face_model2.summary()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "279/279 [==============================] - 43s 152ms/step - loss: 132.4508 - accuracy: 0.1864\n",
      "Epoch 2/50\n",
      "279/279 [==============================] - 37s 134ms/step - loss: 14.0330 - accuracy: 0.2222\n",
      "Epoch 3/50\n",
      "279/279 [==============================] - 38s 135ms/step - loss: 11.7406 - accuracy: 0.1864\n",
      "Epoch 4/50\n",
      "279/279 [==============================] - 38s 136ms/step - loss: 15.9189 - accuracy: 0.2366\n",
      "Epoch 5/50\n",
      "279/279 [==============================] - 38s 136ms/step - loss: 7.9958 - accuracy: 0.3190\n",
      "Epoch 6/50\n",
      "279/279 [==============================] - 38s 136ms/step - loss: 8.3574 - accuracy: 0.2294\n",
      "Epoch 7/50\n",
      "279/279 [==============================] - 38s 134ms/step - loss: 6.7789 - accuracy: 0.3047\n",
      "Epoch 8/50\n",
      "279/279 [==============================] - 38s 137ms/step - loss: 5.9448 - accuracy: 0.3011\n",
      "Epoch 9/50\n",
      "279/279 [==============================] - 38s 136ms/step - loss: 3.0968 - accuracy: 0.3047\n",
      "Epoch 10/50\n",
      "279/279 [==============================] - 38s 137ms/step - loss: 2.3825 - accuracy: 0.3584\n",
      "Epoch 11/50\n",
      "279/279 [==============================] - 38s 136ms/step - loss: 2.0711 - accuracy: 0.4122\n",
      "Epoch 12/50\n",
      "279/279 [==============================] - 37s 134ms/step - loss: 1.8503 - accuracy: 0.4373\n",
      "Epoch 13/50\n",
      "279/279 [==============================] - 38s 135ms/step - loss: 1.3798 - accuracy: 0.4875\n",
      "Epoch 14/50\n",
      "279/279 [==============================] - 38s 138ms/step - loss: 1.3486 - accuracy: 0.5090\n",
      "Epoch 15/50\n",
      "279/279 [==============================] - 38s 136ms/step - loss: 1.4802 - accuracy: 0.5090\n",
      "Epoch 16/50\n",
      "279/279 [==============================] - 37s 134ms/step - loss: 1.3546 - accuracy: 0.5735\n",
      "Epoch 17/50\n",
      "279/279 [==============================] - 38s 137ms/step - loss: 1.1176 - accuracy: 0.6201\n",
      "Epoch 18/50\n",
      "279/279 [==============================] - 38s 138ms/step - loss: 1.1948 - accuracy: 0.6129\n",
      "Epoch 19/50\n",
      "279/279 [==============================] - 38s 135ms/step - loss: 1.4476 - accuracy: 0.6523\n",
      "Epoch 20/50\n",
      "279/279 [==============================] - 38s 137ms/step - loss: 0.8687 - accuracy: 0.6774\n",
      "Epoch 21/50\n",
      "279/279 [==============================] - 38s 137ms/step - loss: 1.4439 - accuracy: 0.6201\n",
      "Epoch 22/50\n",
      "279/279 [==============================] - 38s 136ms/step - loss: 0.7777 - accuracy: 0.7204\n",
      "Epoch 23/50\n",
      "279/279 [==============================] - 39s 138ms/step - loss: 1.1037 - accuracy: 0.6703\n",
      "Epoch 24/50\n",
      "279/279 [==============================] - 38s 135ms/step - loss: 0.7374 - accuracy: 0.7527\n",
      "Epoch 25/50\n",
      "279/279 [==============================] - 38s 135ms/step - loss: 0.9613 - accuracy: 0.7168\n",
      "Epoch 26/50\n",
      "279/279 [==============================] - 37s 134ms/step - loss: 0.9790 - accuracy: 0.7061\n",
      "Epoch 27/50\n",
      "279/279 [==============================] - 38s 135ms/step - loss: 0.7050 - accuracy: 0.7814\n",
      "Epoch 28/50\n",
      "279/279 [==============================] - 37s 134ms/step - loss: 0.6342 - accuracy: 0.7957\n",
      "Epoch 29/50\n",
      "279/279 [==============================] - 38s 135ms/step - loss: 0.4063 - accuracy: 0.8423\n",
      "Epoch 30/50\n",
      "279/279 [==============================] - 38s 135ms/step - loss: 0.4354 - accuracy: 0.8530\n",
      "Epoch 31/50\n",
      "279/279 [==============================] - 38s 134ms/step - loss: 0.7407 - accuracy: 0.8208\n",
      "Epoch 32/50\n",
      "279/279 [==============================] - 37s 134ms/step - loss: 0.5902 - accuracy: 0.8351\n",
      "Epoch 33/50\n",
      "279/279 [==============================] - 38s 136ms/step - loss: 0.5117 - accuracy: 0.8172\n",
      "Epoch 34/50\n",
      "279/279 [==============================] - 38s 136ms/step - loss: 0.3024 - accuracy: 0.8961\n",
      "Epoch 35/50\n",
      "279/279 [==============================] - 38s 136ms/step - loss: 0.2872 - accuracy: 0.8996\n",
      "Epoch 36/50\n",
      "279/279 [==============================] - 38s 135ms/step - loss: 0.4893 - accuracy: 0.8602\n",
      "Epoch 37/50\n",
      "279/279 [==============================] - 37s 134ms/step - loss: 0.3688 - accuracy: 0.8495\n",
      "Epoch 38/50\n",
      "279/279 [==============================] - 37s 134ms/step - loss: 0.4784 - accuracy: 0.8889\n",
      "Epoch 39/50\n",
      "279/279 [==============================] - 38s 137ms/step - loss: 0.3949 - accuracy: 0.8781\n",
      "Epoch 40/50\n",
      "279/279 [==============================] - 38s 135ms/step - loss: 0.1795 - accuracy: 0.9462\n",
      "Epoch 41/50\n",
      "279/279 [==============================] - 38s 136ms/step - loss: 0.5709 - accuracy: 0.8387\n",
      "Epoch 42/50\n",
      "279/279 [==============================] - 37s 134ms/step - loss: 0.4801 - accuracy: 0.8638\n",
      "Epoch 43/50\n",
      "279/279 [==============================] - 37s 134ms/step - loss: 0.3383 - accuracy: 0.9032\n",
      "Epoch 44/50\n",
      "279/279 [==============================] - 37s 134ms/step - loss: 0.2252 - accuracy: 0.9247\n",
      "Epoch 45/50\n",
      "279/279 [==============================] - 38s 136ms/step - loss: 0.2521 - accuracy: 0.9247\n",
      "Epoch 46/50\n",
      "279/279 [==============================] - 38s 135ms/step - loss: 0.2604 - accuracy: 0.9283\n",
      "Epoch 47/50\n",
      "279/279 [==============================] - 37s 134ms/step - loss: 0.1640 - accuracy: 0.9462\n",
      "Epoch 48/50\n",
      "279/279 [==============================] - 37s 134ms/step - loss: 0.2064 - accuracy: 0.9498\n",
      "Epoch 49/50\n",
      "279/279 [==============================] - 38s 135ms/step - loss: 0.1343 - accuracy: 0.9498\n",
      "Epoch 50/50\n",
      "279/279 [==============================] - 38s 135ms/step - loss: 0.4854 - accuracy: 0.9355\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.callbacks.History at 0x7f8ee86000d0>"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## 编译模型\n",
    "resnet_face_model2.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "# 模型训练\n",
    "resnet_face_model2.fit(train_X, train_Y, batch_size=16, epochs=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "ResNet50 Accuracy: 0.280\n"
     ]
    }
   ],
   "source": [
    "metric_accuracy(resnet_face_model2, test_X, test_Y, \"ResNet50\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "ResNet50 Accuracy: 0.315\n"
     ]
    }
   ],
   "source": [
    "metric_accuracy(resnet_face_model2, train_X, train_Y, \"ResNet50\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### FaceNet \n",
    "\n",
    "上一小节中，我们利用了预训练好的 ResNet50 来抽取图像特征，而这一小节我们将利用预训练好的 FaceNet 来抽取人脸特征。我们已经知道 ResNet50 是在大规模数据上建模学习图像特征的，这里面的数据是多种多样的，不限制于人脸图像，而 [FaceNet](https://arxiv.org/pdf/1503.03832.pdf) 是专门对于人脸进行特征抽取的工具。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[<tf.Tensor 'input_1_1:0' shape=(?, 160, 160, 3) dtype=float32>]\n",
      "[<tf.Tensor 'Bottleneck_BatchNorm/cond/Merge:0' shape=(?, 128) dtype=float32>]\n",
      "Model: \"inception_resnet_v1\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_1 (InputLayer)            (None, 160, 160, 3)  0                                            \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_1a_3x3 (Conv2D)          (None, 79, 79, 32)   864         input_1[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_1a_3x3_BatchNorm (BatchN (None, 79, 79, 32)   96          Conv2d_1a_3x3[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_1a_3x3_Activation (Activ (None, 79, 79, 32)   0           Conv2d_1a_3x3_BatchNorm[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_2a_3x3 (Conv2D)          (None, 77, 77, 32)   9216        Conv2d_1a_3x3_Activation[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_2a_3x3_BatchNorm (BatchN (None, 77, 77, 32)   96          Conv2d_2a_3x3[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_2a_3x3_Activation (Activ (None, 77, 77, 32)   0           Conv2d_2a_3x3_BatchNorm[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_2b_3x3 (Conv2D)          (None, 77, 77, 64)   18432       Conv2d_2a_3x3_Activation[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_2b_3x3_BatchNorm (BatchN (None, 77, 77, 64)   192         Conv2d_2b_3x3[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_2b_3x3_Activation (Activ (None, 77, 77, 64)   0           Conv2d_2b_3x3_BatchNorm[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "MaxPool_3a_3x3 (MaxPooling2D)   (None, 38, 38, 64)   0           Conv2d_2b_3x3_Activation[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_3b_1x1 (Conv2D)          (None, 38, 38, 80)   5120        MaxPool_3a_3x3[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_3b_1x1_BatchNorm (BatchN (None, 38, 38, 80)   240         Conv2d_3b_1x1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_3b_1x1_Activation (Activ (None, 38, 38, 80)   0           Conv2d_3b_1x1_BatchNorm[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_4a_3x3 (Conv2D)          (None, 36, 36, 192)  138240      Conv2d_3b_1x1_Activation[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_4a_3x3_BatchNorm (BatchN (None, 36, 36, 192)  576         Conv2d_4a_3x3[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_4a_3x3_Activation (Activ (None, 36, 36, 192)  0           Conv2d_4a_3x3_BatchNorm[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_4b_3x3 (Conv2D)          (None, 17, 17, 256)  442368      Conv2d_4a_3x3_Activation[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_4b_3x3_BatchNorm (BatchN (None, 17, 17, 256)  768         Conv2d_4b_3x3[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_4b_3x3_Activation (Activ (None, 17, 17, 256)  0           Conv2d_4b_3x3_BatchNorm[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   8192        Conv2d_4b_3x3_Activation[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_1_Branch_2_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_1_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   8192        Conv2d_4b_3x3_Activation[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_1_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_1_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_1_Branch_2_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_1_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_1_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_0_Conv2d_1x1 ( (None, 17, 17, 32)   8192        Conv2d_4b_3x3_Activation[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_1_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   9216        Block35_1_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_0_Conv2d_1x1_B (None, 17, 17, 32)   96          Block35_1_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_1_Branch_1_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   96          Block35_1_Branch_2_Conv2d_0c_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_0_Conv2d_1x1_A (None, 17, 17, 32)   0           Block35_1_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_1_Branch_1_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   0           Block35_1_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Concatenate (Concaten (None, 17, 17, 96)   0           Block35_1_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block35_1_Branch_1_Conv2d_0b_3x3_\n",
      "                                                                 Block35_1_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Conv2d_1x1 (Conv2D)   (None, 17, 17, 256)  24832       Block35_1_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_ScaleSum (Lambda)     (None, 17, 17, 256)  0           Conv2d_4b_3x3_Activation[0][0]   \n",
      "                                                                 Block35_1_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Activation (Activatio (None, 17, 17, 256)  0           Block35_1_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   8192        Block35_1_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_2_Branch_2_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_2_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   8192        Block35_1_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_2_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_2_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_2_Branch_2_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_2_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_2_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_0_Conv2d_1x1 ( (None, 17, 17, 32)   8192        Block35_1_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_2_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   9216        Block35_2_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_0_Conv2d_1x1_B (None, 17, 17, 32)   96          Block35_2_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_2_Branch_1_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   96          Block35_2_Branch_2_Conv2d_0c_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_0_Conv2d_1x1_A (None, 17, 17, 32)   0           Block35_2_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_2_Branch_1_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   0           Block35_2_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Concatenate (Concaten (None, 17, 17, 96)   0           Block35_2_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block35_2_Branch_1_Conv2d_0b_3x3_\n",
      "                                                                 Block35_2_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Conv2d_1x1 (Conv2D)   (None, 17, 17, 256)  24832       Block35_2_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_ScaleSum (Lambda)     (None, 17, 17, 256)  0           Block35_1_Activation[0][0]       \n",
      "                                                                 Block35_2_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Activation (Activatio (None, 17, 17, 256)  0           Block35_2_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   8192        Block35_2_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_3_Branch_2_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_3_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   8192        Block35_2_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_3_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_3_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_3_Branch_2_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_3_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_3_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_0_Conv2d_1x1 ( (None, 17, 17, 32)   8192        Block35_2_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_3_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   9216        Block35_3_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_0_Conv2d_1x1_B (None, 17, 17, 32)   96          Block35_3_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_3_Branch_1_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   96          Block35_3_Branch_2_Conv2d_0c_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_0_Conv2d_1x1_A (None, 17, 17, 32)   0           Block35_3_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_3_Branch_1_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   0           Block35_3_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Concatenate (Concaten (None, 17, 17, 96)   0           Block35_3_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block35_3_Branch_1_Conv2d_0b_3x3_\n",
      "                                                                 Block35_3_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Conv2d_1x1 (Conv2D)   (None, 17, 17, 256)  24832       Block35_3_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_ScaleSum (Lambda)     (None, 17, 17, 256)  0           Block35_2_Activation[0][0]       \n",
      "                                                                 Block35_3_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Activation (Activatio (None, 17, 17, 256)  0           Block35_3_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   8192        Block35_3_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_4_Branch_2_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_4_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   8192        Block35_3_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_4_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_4_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_4_Branch_2_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_4_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_4_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_0_Conv2d_1x1 ( (None, 17, 17, 32)   8192        Block35_3_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_4_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   9216        Block35_4_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_0_Conv2d_1x1_B (None, 17, 17, 32)   96          Block35_4_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_4_Branch_1_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   96          Block35_4_Branch_2_Conv2d_0c_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_0_Conv2d_1x1_A (None, 17, 17, 32)   0           Block35_4_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_4_Branch_1_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   0           Block35_4_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Concatenate (Concaten (None, 17, 17, 96)   0           Block35_4_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block35_4_Branch_1_Conv2d_0b_3x3_\n",
      "                                                                 Block35_4_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Conv2d_1x1 (Conv2D)   (None, 17, 17, 256)  24832       Block35_4_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_ScaleSum (Lambda)     (None, 17, 17, 256)  0           Block35_3_Activation[0][0]       \n",
      "                                                                 Block35_4_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Activation (Activatio (None, 17, 17, 256)  0           Block35_4_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   8192        Block35_4_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_5_Branch_2_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_5_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   8192        Block35_4_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_5_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_5_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_5_Branch_2_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_5_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_5_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_0_Conv2d_1x1 ( (None, 17, 17, 32)   8192        Block35_4_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_5_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   9216        Block35_5_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_0_Conv2d_1x1_B (None, 17, 17, 32)   96          Block35_5_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_5_Branch_1_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   96          Block35_5_Branch_2_Conv2d_0c_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_0_Conv2d_1x1_A (None, 17, 17, 32)   0           Block35_5_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_5_Branch_1_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   0           Block35_5_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Concatenate (Concaten (None, 17, 17, 96)   0           Block35_5_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block35_5_Branch_1_Conv2d_0b_3x3_\n",
      "                                                                 Block35_5_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Conv2d_1x1 (Conv2D)   (None, 17, 17, 256)  24832       Block35_5_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_ScaleSum (Lambda)     (None, 17, 17, 256)  0           Block35_4_Activation[0][0]       \n",
      "                                                                 Block35_5_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Activation (Activatio (None, 17, 17, 256)  0           Block35_5_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_1_Conv2d_0a_1x1 (None, 17, 17, 192)  49152       Block35_5_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_1_Conv2d_0a_1x1 (None, 17, 17, 192)  576         Mixed_6a_Branch_1_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_1_Conv2d_0a_1x1 (None, 17, 17, 192)  0           Mixed_6a_Branch_1_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_1_Conv2d_0b_3x3 (None, 17, 17, 192)  331776      Mixed_6a_Branch_1_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_1_Conv2d_0b_3x3 (None, 17, 17, 192)  576         Mixed_6a_Branch_1_Conv2d_0b_3x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_1_Conv2d_0b_3x3 (None, 17, 17, 192)  0           Mixed_6a_Branch_1_Conv2d_0b_3x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_0_Conv2d_1a_3x3 (None, 8, 8, 384)    884736      Block35_5_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_1_Conv2d_1a_3x3 (None, 8, 8, 256)    442368      Mixed_6a_Branch_1_Conv2d_0b_3x3_A\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_0_Conv2d_1a_3x3 (None, 8, 8, 384)    1152        Mixed_6a_Branch_0_Conv2d_1a_3x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_1_Conv2d_1a_3x3 (None, 8, 8, 256)    768         Mixed_6a_Branch_1_Conv2d_1a_3x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_0_Conv2d_1a_3x3 (None, 8, 8, 384)    0           Mixed_6a_Branch_0_Conv2d_1a_3x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_1_Conv2d_1a_3x3 (None, 8, 8, 256)    0           Mixed_6a_Branch_1_Conv2d_1a_3x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_2_MaxPool_1a_3x (None, 8, 8, 256)    0           Block35_5_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a (Concatenate)          (None, 8, 8, 896)    0           Mixed_6a_Branch_0_Conv2d_1a_3x3_A\n",
      "                                                                 Mixed_6a_Branch_1_Conv2d_1a_3x3_A\n",
      "                                                                 Mixed_6a_Branch_2_MaxPool_1a_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    114688      Mixed_6a[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    384         Block17_1_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    0           Block17_1_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    114688      Block17_1_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    384         Block17_1_Branch_1_Conv2d_0b_1x7[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    0           Block17_1_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_0_Conv2d_1x1 ( (None, 8, 8, 128)    114688      Mixed_6a[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    114688      Block17_1_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_0_Conv2d_1x1_B (None, 8, 8, 128)    384         Block17_1_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    384         Block17_1_Branch_1_Conv2d_0c_7x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_0_Conv2d_1x1_A (None, 8, 8, 128)    0           Block17_1_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    0           Block17_1_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Concatenate (Concaten (None, 8, 8, 256)    0           Block17_1_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block17_1_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Conv2d_1x1 (Conv2D)   (None, 8, 8, 896)    230272      Block17_1_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_ScaleSum (Lambda)     (None, 8, 8, 896)    0           Mixed_6a[0][0]                   \n",
      "                                                                 Block17_1_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Activation (Activatio (None, 8, 8, 896)    0           Block17_1_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    114688      Block17_1_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    384         Block17_2_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    0           Block17_2_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    114688      Block17_2_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    384         Block17_2_Branch_1_Conv2d_0b_1x7[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    0           Block17_2_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_0_Conv2d_1x1 ( (None, 8, 8, 128)    114688      Block17_1_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    114688      Block17_2_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_0_Conv2d_1x1_B (None, 8, 8, 128)    384         Block17_2_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    384         Block17_2_Branch_1_Conv2d_0c_7x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_0_Conv2d_1x1_A (None, 8, 8, 128)    0           Block17_2_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    0           Block17_2_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Concatenate (Concaten (None, 8, 8, 256)    0           Block17_2_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block17_2_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Conv2d_1x1 (Conv2D)   (None, 8, 8, 896)    230272      Block17_2_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_ScaleSum (Lambda)     (None, 8, 8, 896)    0           Block17_1_Activation[0][0]       \n",
      "                                                                 Block17_2_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Activation (Activatio (None, 8, 8, 896)    0           Block17_2_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    114688      Block17_2_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    384         Block17_3_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    0           Block17_3_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    114688      Block17_3_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    384         Block17_3_Branch_1_Conv2d_0b_1x7[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    0           Block17_3_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_0_Conv2d_1x1 ( (None, 8, 8, 128)    114688      Block17_2_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    114688      Block17_3_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_0_Conv2d_1x1_B (None, 8, 8, 128)    384         Block17_3_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    384         Block17_3_Branch_1_Conv2d_0c_7x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_0_Conv2d_1x1_A (None, 8, 8, 128)    0           Block17_3_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    0           Block17_3_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Concatenate (Concaten (None, 8, 8, 256)    0           Block17_3_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block17_3_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Conv2d_1x1 (Conv2D)   (None, 8, 8, 896)    230272      Block17_3_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_ScaleSum (Lambda)     (None, 8, 8, 896)    0           Block17_2_Activation[0][0]       \n",
      "                                                                 Block17_3_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Activation (Activatio (None, 8, 8, 896)    0           Block17_3_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    114688      Block17_3_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    384         Block17_4_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    0           Block17_4_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    114688      Block17_4_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    384         Block17_4_Branch_1_Conv2d_0b_1x7[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    0           Block17_4_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_0_Conv2d_1x1 ( (None, 8, 8, 128)    114688      Block17_3_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    114688      Block17_4_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_0_Conv2d_1x1_B (None, 8, 8, 128)    384         Block17_4_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    384         Block17_4_Branch_1_Conv2d_0c_7x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_0_Conv2d_1x1_A (None, 8, 8, 128)    0           Block17_4_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    0           Block17_4_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Concatenate (Concaten (None, 8, 8, 256)    0           Block17_4_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block17_4_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Conv2d_1x1 (Conv2D)   (None, 8, 8, 896)    230272      Block17_4_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_ScaleSum (Lambda)     (None, 8, 8, 896)    0           Block17_3_Activation[0][0]       \n",
      "                                                                 Block17_4_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Activation (Activatio (None, 8, 8, 896)    0           Block17_4_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    114688      Block17_4_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    384         Block17_5_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    0           Block17_5_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    114688      Block17_5_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    384         Block17_5_Branch_1_Conv2d_0b_1x7[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    0           Block17_5_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_0_Conv2d_1x1 ( (None, 8, 8, 128)    114688      Block17_4_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    114688      Block17_5_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_0_Conv2d_1x1_B (None, 8, 8, 128)    384         Block17_5_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    384         Block17_5_Branch_1_Conv2d_0c_7x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_0_Conv2d_1x1_A (None, 8, 8, 128)    0           Block17_5_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    0           Block17_5_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Concatenate (Concaten (None, 8, 8, 256)    0           Block17_5_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block17_5_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Conv2d_1x1 (Conv2D)   (None, 8, 8, 896)    230272      Block17_5_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_ScaleSum (Lambda)     (None, 8, 8, 896)    0           Block17_4_Activation[0][0]       \n",
      "                                                                 Block17_5_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Activation (Activatio (None, 8, 8, 896)    0           Block17_5_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    114688      Block17_5_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    384         Block17_6_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    0           Block17_6_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    114688      Block17_6_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    384         Block17_6_Branch_1_Conv2d_0b_1x7[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    0           Block17_6_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_0_Conv2d_1x1 ( (None, 8, 8, 128)    114688      Block17_5_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    114688      Block17_6_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_0_Conv2d_1x1_B (None, 8, 8, 128)    384         Block17_6_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    384         Block17_6_Branch_1_Conv2d_0c_7x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_0_Conv2d_1x1_A (None, 8, 8, 128)    0           Block17_6_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    0           Block17_6_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Concatenate (Concaten (None, 8, 8, 256)    0           Block17_6_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block17_6_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Conv2d_1x1 (Conv2D)   (None, 8, 8, 896)    230272      Block17_6_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_ScaleSum (Lambda)     (None, 8, 8, 896)    0           Block17_5_Activation[0][0]       \n",
      "                                                                 Block17_6_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Activation (Activatio (None, 8, 8, 896)    0           Block17_6_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    114688      Block17_6_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    384         Block17_7_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    0           Block17_7_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    114688      Block17_7_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    384         Block17_7_Branch_1_Conv2d_0b_1x7[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    0           Block17_7_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_0_Conv2d_1x1 ( (None, 8, 8, 128)    114688      Block17_6_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    114688      Block17_7_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_0_Conv2d_1x1_B (None, 8, 8, 128)    384         Block17_7_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    384         Block17_7_Branch_1_Conv2d_0c_7x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_0_Conv2d_1x1_A (None, 8, 8, 128)    0           Block17_7_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    0           Block17_7_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Concatenate (Concaten (None, 8, 8, 256)    0           Block17_7_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block17_7_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Conv2d_1x1 (Conv2D)   (None, 8, 8, 896)    230272      Block17_7_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_ScaleSum (Lambda)     (None, 8, 8, 896)    0           Block17_6_Activation[0][0]       \n",
      "                                                                 Block17_7_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Activation (Activatio (None, 8, 8, 896)    0           Block17_7_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    114688      Block17_7_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    384         Block17_8_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    0           Block17_8_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    114688      Block17_8_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    384         Block17_8_Branch_1_Conv2d_0b_1x7[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    0           Block17_8_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_0_Conv2d_1x1 ( (None, 8, 8, 128)    114688      Block17_7_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    114688      Block17_8_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_0_Conv2d_1x1_B (None, 8, 8, 128)    384         Block17_8_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    384         Block17_8_Branch_1_Conv2d_0c_7x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_0_Conv2d_1x1_A (None, 8, 8, 128)    0           Block17_8_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    0           Block17_8_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Concatenate (Concaten (None, 8, 8, 256)    0           Block17_8_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block17_8_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Conv2d_1x1 (Conv2D)   (None, 8, 8, 896)    230272      Block17_8_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_ScaleSum (Lambda)     (None, 8, 8, 896)    0           Block17_7_Activation[0][0]       \n",
      "                                                                 Block17_8_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Activation (Activatio (None, 8, 8, 896)    0           Block17_8_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    114688      Block17_8_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    384         Block17_9_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    0           Block17_9_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    114688      Block17_9_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    384         Block17_9_Branch_1_Conv2d_0b_1x7[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    0           Block17_9_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_0_Conv2d_1x1 ( (None, 8, 8, 128)    114688      Block17_8_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    114688      Block17_9_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_0_Conv2d_1x1_B (None, 8, 8, 128)    384         Block17_9_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    384         Block17_9_Branch_1_Conv2d_0c_7x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_0_Conv2d_1x1_A (None, 8, 8, 128)    0           Block17_9_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    0           Block17_9_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Concatenate (Concaten (None, 8, 8, 256)    0           Block17_9_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block17_9_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Conv2d_1x1 (Conv2D)   (None, 8, 8, 896)    230272      Block17_9_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_ScaleSum (Lambda)     (None, 8, 8, 896)    0           Block17_8_Activation[0][0]       \n",
      "                                                                 Block17_9_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Activation (Activatio (None, 8, 8, 896)    0           Block17_9_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_1_Conv2d_0a_1 (None, 8, 8, 128)    114688      Block17_9_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_1_Conv2d_0a_1 (None, 8, 8, 128)    384         Block17_10_Branch_1_Conv2d_0a_1x1\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_1_Conv2d_0a_1 (None, 8, 8, 128)    0           Block17_10_Branch_1_Conv2d_0a_1x1\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_1_Conv2d_0b_1 (None, 8, 8, 128)    114688      Block17_10_Branch_1_Conv2d_0a_1x1\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_1_Conv2d_0b_1 (None, 8, 8, 128)    384         Block17_10_Branch_1_Conv2d_0b_1x7\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_1_Conv2d_0b_1 (None, 8, 8, 128)    0           Block17_10_Branch_1_Conv2d_0b_1x7\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_0_Conv2d_1x1  (None, 8, 8, 128)    114688      Block17_9_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_1_Conv2d_0c_7 (None, 8, 8, 128)    114688      Block17_10_Branch_1_Conv2d_0b_1x7\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_0_Conv2d_1x1_ (None, 8, 8, 128)    384         Block17_10_Branch_0_Conv2d_1x1[0]\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_1_Conv2d_0c_7 (None, 8, 8, 128)    384         Block17_10_Branch_1_Conv2d_0c_7x1\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_0_Conv2d_1x1_ (None, 8, 8, 128)    0           Block17_10_Branch_0_Conv2d_1x1_Ba\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_1_Conv2d_0c_7 (None, 8, 8, 128)    0           Block17_10_Branch_1_Conv2d_0c_7x1\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Concatenate (Concate (None, 8, 8, 256)    0           Block17_10_Branch_0_Conv2d_1x1_Ac\n",
      "                                                                 Block17_10_Branch_1_Conv2d_0c_7x1\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Conv2d_1x1 (Conv2D)  (None, 8, 8, 896)    230272      Block17_10_Concatenate[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_ScaleSum (Lambda)    (None, 8, 8, 896)    0           Block17_9_Activation[0][0]       \n",
      "                                                                 Block17_10_Conv2d_1x1[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Activation (Activati (None, 8, 8, 896)    0           Block17_10_ScaleSum[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_2_Conv2d_0a_1x1 (None, 8, 8, 256)    229376      Block17_10_Activation[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_2_Conv2d_0a_1x1 (None, 8, 8, 256)    768         Mixed_7a_Branch_2_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_2_Conv2d_0a_1x1 (None, 8, 8, 256)    0           Mixed_7a_Branch_2_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_0_Conv2d_0a_1x1 (None, 8, 8, 256)    229376      Block17_10_Activation[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_1_Conv2d_0a_1x1 (None, 8, 8, 256)    229376      Block17_10_Activation[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_2_Conv2d_0b_3x3 (None, 8, 8, 256)    589824      Mixed_7a_Branch_2_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_0_Conv2d_0a_1x1 (None, 8, 8, 256)    768         Mixed_7a_Branch_0_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_1_Conv2d_0a_1x1 (None, 8, 8, 256)    768         Mixed_7a_Branch_1_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_2_Conv2d_0b_3x3 (None, 8, 8, 256)    768         Mixed_7a_Branch_2_Conv2d_0b_3x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_0_Conv2d_0a_1x1 (None, 8, 8, 256)    0           Mixed_7a_Branch_0_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_1_Conv2d_0a_1x1 (None, 8, 8, 256)    0           Mixed_7a_Branch_1_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_2_Conv2d_0b_3x3 (None, 8, 8, 256)    0           Mixed_7a_Branch_2_Conv2d_0b_3x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_0_Conv2d_1a_3x3 (None, 3, 3, 384)    884736      Mixed_7a_Branch_0_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_1_Conv2d_1a_3x3 (None, 3, 3, 256)    589824      Mixed_7a_Branch_1_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_2_Conv2d_1a_3x3 (None, 3, 3, 256)    589824      Mixed_7a_Branch_2_Conv2d_0b_3x3_A\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_0_Conv2d_1a_3x3 (None, 3, 3, 384)    1152        Mixed_7a_Branch_0_Conv2d_1a_3x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_1_Conv2d_1a_3x3 (None, 3, 3, 256)    768         Mixed_7a_Branch_1_Conv2d_1a_3x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_2_Conv2d_1a_3x3 (None, 3, 3, 256)    768         Mixed_7a_Branch_2_Conv2d_1a_3x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_0_Conv2d_1a_3x3 (None, 3, 3, 384)    0           Mixed_7a_Branch_0_Conv2d_1a_3x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_1_Conv2d_1a_3x3 (None, 3, 3, 256)    0           Mixed_7a_Branch_1_Conv2d_1a_3x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_2_Conv2d_1a_3x3 (None, 3, 3, 256)    0           Mixed_7a_Branch_2_Conv2d_1a_3x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_3_MaxPool_1a_3x (None, 3, 3, 896)    0           Block17_10_Activation[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a (Concatenate)          (None, 3, 3, 1792)   0           Mixed_7a_Branch_0_Conv2d_1a_3x3_A\n",
      "                                                                 Mixed_7a_Branch_1_Conv2d_1a_3x3_A\n",
      "                                                                 Mixed_7a_Branch_2_Conv2d_1a_3x3_A\n",
      "                                                                 Mixed_7a_Branch_3_MaxPool_1a_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    344064      Mixed_7a[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    576         Block8_1_Branch_1_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    0           Block8_1_Branch_1_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    110592      Block8_1_Branch_1_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    576         Block8_1_Branch_1_Conv2d_0b_1x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    0           Block8_1_Branch_1_Conv2d_0b_1x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_0_Conv2d_1x1 (C (None, 3, 3, 192)    344064      Mixed_7a[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    110592      Block8_1_Branch_1_Conv2d_0b_1x3_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_0_Conv2d_1x1_Ba (None, 3, 3, 192)    576         Block8_1_Branch_0_Conv2d_1x1[0][0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    576         Block8_1_Branch_1_Conv2d_0c_3x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_0_Conv2d_1x1_Ac (None, 3, 3, 192)    0           Block8_1_Branch_0_Conv2d_1x1_Batc\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    0           Block8_1_Branch_1_Conv2d_0c_3x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Concatenate (Concatena (None, 3, 3, 384)    0           Block8_1_Branch_0_Conv2d_1x1_Acti\n",
      "                                                                 Block8_1_Branch_1_Conv2d_0c_3x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Conv2d_1x1 (Conv2D)    (None, 3, 3, 1792)   689920      Block8_1_Concatenate[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_ScaleSum (Lambda)      (None, 3, 3, 1792)   0           Mixed_7a[0][0]                   \n",
      "                                                                 Block8_1_Conv2d_1x1[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Activation (Activation (None, 3, 3, 1792)   0           Block8_1_ScaleSum[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    344064      Block8_1_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    576         Block8_2_Branch_1_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    0           Block8_2_Branch_1_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    110592      Block8_2_Branch_1_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    576         Block8_2_Branch_1_Conv2d_0b_1x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    0           Block8_2_Branch_1_Conv2d_0b_1x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_0_Conv2d_1x1 (C (None, 3, 3, 192)    344064      Block8_1_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    110592      Block8_2_Branch_1_Conv2d_0b_1x3_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_0_Conv2d_1x1_Ba (None, 3, 3, 192)    576         Block8_2_Branch_0_Conv2d_1x1[0][0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    576         Block8_2_Branch_1_Conv2d_0c_3x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_0_Conv2d_1x1_Ac (None, 3, 3, 192)    0           Block8_2_Branch_0_Conv2d_1x1_Batc\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    0           Block8_2_Branch_1_Conv2d_0c_3x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Concatenate (Concatena (None, 3, 3, 384)    0           Block8_2_Branch_0_Conv2d_1x1_Acti\n",
      "                                                                 Block8_2_Branch_1_Conv2d_0c_3x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Conv2d_1x1 (Conv2D)    (None, 3, 3, 1792)   689920      Block8_2_Concatenate[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_ScaleSum (Lambda)      (None, 3, 3, 1792)   0           Block8_1_Activation[0][0]        \n",
      "                                                                 Block8_2_Conv2d_1x1[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Activation (Activation (None, 3, 3, 1792)   0           Block8_2_ScaleSum[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    344064      Block8_2_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    576         Block8_3_Branch_1_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    0           Block8_3_Branch_1_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    110592      Block8_3_Branch_1_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    576         Block8_3_Branch_1_Conv2d_0b_1x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    0           Block8_3_Branch_1_Conv2d_0b_1x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_0_Conv2d_1x1 (C (None, 3, 3, 192)    344064      Block8_2_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    110592      Block8_3_Branch_1_Conv2d_0b_1x3_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_0_Conv2d_1x1_Ba (None, 3, 3, 192)    576         Block8_3_Branch_0_Conv2d_1x1[0][0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    576         Block8_3_Branch_1_Conv2d_0c_3x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_0_Conv2d_1x1_Ac (None, 3, 3, 192)    0           Block8_3_Branch_0_Conv2d_1x1_Batc\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    0           Block8_3_Branch_1_Conv2d_0c_3x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Concatenate (Concatena (None, 3, 3, 384)    0           Block8_3_Branch_0_Conv2d_1x1_Acti\n",
      "                                                                 Block8_3_Branch_1_Conv2d_0c_3x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Conv2d_1x1 (Conv2D)    (None, 3, 3, 1792)   689920      Block8_3_Concatenate[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_ScaleSum (Lambda)      (None, 3, 3, 1792)   0           Block8_2_Activation[0][0]        \n",
      "                                                                 Block8_3_Conv2d_1x1[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Activation (Activation (None, 3, 3, 1792)   0           Block8_3_ScaleSum[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    344064      Block8_3_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    576         Block8_4_Branch_1_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    0           Block8_4_Branch_1_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    110592      Block8_4_Branch_1_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    576         Block8_4_Branch_1_Conv2d_0b_1x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    0           Block8_4_Branch_1_Conv2d_0b_1x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_0_Conv2d_1x1 (C (None, 3, 3, 192)    344064      Block8_3_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    110592      Block8_4_Branch_1_Conv2d_0b_1x3_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_0_Conv2d_1x1_Ba (None, 3, 3, 192)    576         Block8_4_Branch_0_Conv2d_1x1[0][0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    576         Block8_4_Branch_1_Conv2d_0c_3x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_0_Conv2d_1x1_Ac (None, 3, 3, 192)    0           Block8_4_Branch_0_Conv2d_1x1_Batc\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    0           Block8_4_Branch_1_Conv2d_0c_3x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Concatenate (Concatena (None, 3, 3, 384)    0           Block8_4_Branch_0_Conv2d_1x1_Acti\n",
      "                                                                 Block8_4_Branch_1_Conv2d_0c_3x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Conv2d_1x1 (Conv2D)    (None, 3, 3, 1792)   689920      Block8_4_Concatenate[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_ScaleSum (Lambda)      (None, 3, 3, 1792)   0           Block8_3_Activation[0][0]        \n",
      "                                                                 Block8_4_Conv2d_1x1[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Activation (Activation (None, 3, 3, 1792)   0           Block8_4_ScaleSum[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    344064      Block8_4_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    576         Block8_5_Branch_1_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    0           Block8_5_Branch_1_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    110592      Block8_5_Branch_1_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    576         Block8_5_Branch_1_Conv2d_0b_1x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    0           Block8_5_Branch_1_Conv2d_0b_1x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_0_Conv2d_1x1 (C (None, 3, 3, 192)    344064      Block8_4_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    110592      Block8_5_Branch_1_Conv2d_0b_1x3_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_0_Conv2d_1x1_Ba (None, 3, 3, 192)    576         Block8_5_Branch_0_Conv2d_1x1[0][0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    576         Block8_5_Branch_1_Conv2d_0c_3x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_0_Conv2d_1x1_Ac (None, 3, 3, 192)    0           Block8_5_Branch_0_Conv2d_1x1_Batc\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    0           Block8_5_Branch_1_Conv2d_0c_3x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Concatenate (Concatena (None, 3, 3, 384)    0           Block8_5_Branch_0_Conv2d_1x1_Acti\n",
      "                                                                 Block8_5_Branch_1_Conv2d_0c_3x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Conv2d_1x1 (Conv2D)    (None, 3, 3, 1792)   689920      Block8_5_Concatenate[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_ScaleSum (Lambda)      (None, 3, 3, 1792)   0           Block8_4_Activation[0][0]        \n",
      "                                                                 Block8_5_Conv2d_1x1[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Activation (Activation (None, 3, 3, 1792)   0           Block8_5_ScaleSum[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    344064      Block8_5_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    576         Block8_6_Branch_1_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    0           Block8_6_Branch_1_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    110592      Block8_6_Branch_1_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    576         Block8_6_Branch_1_Conv2d_0b_1x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    0           Block8_6_Branch_1_Conv2d_0b_1x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_0_Conv2d_1x1 (C (None, 3, 3, 192)    344064      Block8_5_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    110592      Block8_6_Branch_1_Conv2d_0b_1x3_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_0_Conv2d_1x1_Ba (None, 3, 3, 192)    576         Block8_6_Branch_0_Conv2d_1x1[0][0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    576         Block8_6_Branch_1_Conv2d_0c_3x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_0_Conv2d_1x1_Ac (None, 3, 3, 192)    0           Block8_6_Branch_0_Conv2d_1x1_Batc\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    0           Block8_6_Branch_1_Conv2d_0c_3x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Concatenate (Concatena (None, 3, 3, 384)    0           Block8_6_Branch_0_Conv2d_1x1_Acti\n",
      "                                                                 Block8_6_Branch_1_Conv2d_0c_3x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Conv2d_1x1 (Conv2D)    (None, 3, 3, 1792)   689920      Block8_6_Concatenate[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_ScaleSum (Lambda)      (None, 3, 3, 1792)   0           Block8_5_Activation[0][0]        \n",
      "                                                                 Block8_6_Conv2d_1x1[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "AvgPool (GlobalAveragePooling2D (None, 1792)         0           Block8_6_ScaleSum[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "Dropout (Dropout)               (None, 1792)         0           AvgPool[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "Bottleneck (Dense)              (None, 128)          229376      Dropout[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "Bottleneck_BatchNorm (BatchNorm (None, 128)          384         Bottleneck[0][0]                 \n",
      "==================================================================================================\n",
      "Total params: 22,808,144\n",
      "Trainable params: 22,779,312\n",
      "Non-trainable params: 28,832\n",
      "__________________________________________________________________________________________________\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/leon/anaconda3/lib/python3.7/site-packages/keras/engine/saving.py:341: UserWarning: No training configuration found in save file: the model was *not* compiled. Compile it manually.\n",
      "  warnings.warn('No training configuration found in save file: '\n"
     ]
    }
   ],
   "source": [
    "from keras.models import load_model\n",
    "# load the model\n",
    "model  =  load_model('./models/facenet_keras.h5')\n",
    "# summarize input and output shape\n",
    "print(model.inputs)\n",
    "print(model.outputs)\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_24\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_1 (InputLayer)            (None, 160, 160, 3)  0                                            \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_1a_3x3 (Conv2D)          (None, 79, 79, 32)   864         input_1[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_1a_3x3_BatchNorm (BatchN (None, 79, 79, 32)   96          Conv2d_1a_3x3[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_1a_3x3_Activation (Activ (None, 79, 79, 32)   0           Conv2d_1a_3x3_BatchNorm[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_2a_3x3 (Conv2D)          (None, 77, 77, 32)   9216        Conv2d_1a_3x3_Activation[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_2a_3x3_BatchNorm (BatchN (None, 77, 77, 32)   96          Conv2d_2a_3x3[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_2a_3x3_Activation (Activ (None, 77, 77, 32)   0           Conv2d_2a_3x3_BatchNorm[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_2b_3x3 (Conv2D)          (None, 77, 77, 64)   18432       Conv2d_2a_3x3_Activation[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_2b_3x3_BatchNorm (BatchN (None, 77, 77, 64)   192         Conv2d_2b_3x3[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_2b_3x3_Activation (Activ (None, 77, 77, 64)   0           Conv2d_2b_3x3_BatchNorm[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "MaxPool_3a_3x3 (MaxPooling2D)   (None, 38, 38, 64)   0           Conv2d_2b_3x3_Activation[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_3b_1x1 (Conv2D)          (None, 38, 38, 80)   5120        MaxPool_3a_3x3[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_3b_1x1_BatchNorm (BatchN (None, 38, 38, 80)   240         Conv2d_3b_1x1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_3b_1x1_Activation (Activ (None, 38, 38, 80)   0           Conv2d_3b_1x1_BatchNorm[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_4a_3x3 (Conv2D)          (None, 36, 36, 192)  138240      Conv2d_3b_1x1_Activation[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_4a_3x3_BatchNorm (BatchN (None, 36, 36, 192)  576         Conv2d_4a_3x3[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_4a_3x3_Activation (Activ (None, 36, 36, 192)  0           Conv2d_4a_3x3_BatchNorm[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_4b_3x3 (Conv2D)          (None, 17, 17, 256)  442368      Conv2d_4a_3x3_Activation[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_4b_3x3_BatchNorm (BatchN (None, 17, 17, 256)  768         Conv2d_4b_3x3[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "Conv2d_4b_3x3_Activation (Activ (None, 17, 17, 256)  0           Conv2d_4b_3x3_BatchNorm[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   8192        Conv2d_4b_3x3_Activation[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_1_Branch_2_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_1_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   8192        Conv2d_4b_3x3_Activation[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_1_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_1_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_1_Branch_2_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_1_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_1_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_0_Conv2d_1x1 ( (None, 17, 17, 32)   8192        Conv2d_4b_3x3_Activation[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_1_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   9216        Block35_1_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_0_Conv2d_1x1_B (None, 17, 17, 32)   96          Block35_1_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_1_Branch_1_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   96          Block35_1_Branch_2_Conv2d_0c_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_0_Conv2d_1x1_A (None, 17, 17, 32)   0           Block35_1_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_1_Branch_1_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   0           Block35_1_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Concatenate (Concaten (None, 17, 17, 96)   0           Block35_1_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block35_1_Branch_1_Conv2d_0b_3x3_\n",
      "                                                                 Block35_1_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Conv2d_1x1 (Conv2D)   (None, 17, 17, 256)  24832       Block35_1_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_ScaleSum (Lambda)     (None, 17, 17, 256)  0           Conv2d_4b_3x3_Activation[0][0]   \n",
      "                                                                 Block35_1_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_1_Activation (Activatio (None, 17, 17, 256)  0           Block35_1_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   8192        Block35_1_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_2_Branch_2_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_2_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   8192        Block35_1_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_2_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_2_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_2_Branch_2_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_2_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_2_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_0_Conv2d_1x1 ( (None, 17, 17, 32)   8192        Block35_1_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_2_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   9216        Block35_2_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_0_Conv2d_1x1_B (None, 17, 17, 32)   96          Block35_2_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_2_Branch_1_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   96          Block35_2_Branch_2_Conv2d_0c_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_0_Conv2d_1x1_A (None, 17, 17, 32)   0           Block35_2_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_2_Branch_1_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   0           Block35_2_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Concatenate (Concaten (None, 17, 17, 96)   0           Block35_2_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block35_2_Branch_1_Conv2d_0b_3x3_\n",
      "                                                                 Block35_2_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Conv2d_1x1 (Conv2D)   (None, 17, 17, 256)  24832       Block35_2_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_ScaleSum (Lambda)     (None, 17, 17, 256)  0           Block35_1_Activation[0][0]       \n",
      "                                                                 Block35_2_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_2_Activation (Activatio (None, 17, 17, 256)  0           Block35_2_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   8192        Block35_2_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_3_Branch_2_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_3_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   8192        Block35_2_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_3_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_3_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_3_Branch_2_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_3_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_3_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_0_Conv2d_1x1 ( (None, 17, 17, 32)   8192        Block35_2_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_3_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   9216        Block35_3_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_0_Conv2d_1x1_B (None, 17, 17, 32)   96          Block35_3_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_3_Branch_1_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   96          Block35_3_Branch_2_Conv2d_0c_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_0_Conv2d_1x1_A (None, 17, 17, 32)   0           Block35_3_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_3_Branch_1_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   0           Block35_3_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Concatenate (Concaten (None, 17, 17, 96)   0           Block35_3_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block35_3_Branch_1_Conv2d_0b_3x3_\n",
      "                                                                 Block35_3_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Conv2d_1x1 (Conv2D)   (None, 17, 17, 256)  24832       Block35_3_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_ScaleSum (Lambda)     (None, 17, 17, 256)  0           Block35_2_Activation[0][0]       \n",
      "                                                                 Block35_3_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_3_Activation (Activatio (None, 17, 17, 256)  0           Block35_3_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   8192        Block35_3_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_4_Branch_2_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_4_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   8192        Block35_3_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_4_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_4_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_4_Branch_2_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_4_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_4_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_0_Conv2d_1x1 ( (None, 17, 17, 32)   8192        Block35_3_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_4_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   9216        Block35_4_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_0_Conv2d_1x1_B (None, 17, 17, 32)   96          Block35_4_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_4_Branch_1_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   96          Block35_4_Branch_2_Conv2d_0c_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_0_Conv2d_1x1_A (None, 17, 17, 32)   0           Block35_4_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_4_Branch_1_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   0           Block35_4_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Concatenate (Concaten (None, 17, 17, 96)   0           Block35_4_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block35_4_Branch_1_Conv2d_0b_3x3_\n",
      "                                                                 Block35_4_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Conv2d_1x1 (Conv2D)   (None, 17, 17, 256)  24832       Block35_4_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_ScaleSum (Lambda)     (None, 17, 17, 256)  0           Block35_3_Activation[0][0]       \n",
      "                                                                 Block35_4_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_4_Activation (Activatio (None, 17, 17, 256)  0           Block35_4_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   8192        Block35_4_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_5_Branch_2_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_2_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_5_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   8192        Block35_4_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_5_Branch_2_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   96          Block35_5_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_5_Branch_2_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_1_Conv2d_0a_1x (None, 17, 17, 32)   0           Block35_5_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_2_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_5_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_0_Conv2d_1x1 ( (None, 17, 17, 32)   8192        Block35_4_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   9216        Block35_5_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   9216        Block35_5_Branch_2_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_0_Conv2d_1x1_B (None, 17, 17, 32)   96          Block35_5_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   96          Block35_5_Branch_1_Conv2d_0b_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   96          Block35_5_Branch_2_Conv2d_0c_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_0_Conv2d_1x1_A (None, 17, 17, 32)   0           Block35_5_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_1_Conv2d_0b_3x (None, 17, 17, 32)   0           Block35_5_Branch_1_Conv2d_0b_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Branch_2_Conv2d_0c_3x (None, 17, 17, 32)   0           Block35_5_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Concatenate (Concaten (None, 17, 17, 96)   0           Block35_5_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block35_5_Branch_1_Conv2d_0b_3x3_\n",
      "                                                                 Block35_5_Branch_2_Conv2d_0c_3x3_\n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Conv2d_1x1 (Conv2D)   (None, 17, 17, 256)  24832       Block35_5_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_ScaleSum (Lambda)     (None, 17, 17, 256)  0           Block35_4_Activation[0][0]       \n",
      "                                                                 Block35_5_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block35_5_Activation (Activatio (None, 17, 17, 256)  0           Block35_5_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_1_Conv2d_0a_1x1 (None, 17, 17, 192)  49152       Block35_5_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_1_Conv2d_0a_1x1 (None, 17, 17, 192)  576         Mixed_6a_Branch_1_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_1_Conv2d_0a_1x1 (None, 17, 17, 192)  0           Mixed_6a_Branch_1_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_1_Conv2d_0b_3x3 (None, 17, 17, 192)  331776      Mixed_6a_Branch_1_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_1_Conv2d_0b_3x3 (None, 17, 17, 192)  576         Mixed_6a_Branch_1_Conv2d_0b_3x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_1_Conv2d_0b_3x3 (None, 17, 17, 192)  0           Mixed_6a_Branch_1_Conv2d_0b_3x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_0_Conv2d_1a_3x3 (None, 8, 8, 384)    884736      Block35_5_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_1_Conv2d_1a_3x3 (None, 8, 8, 256)    442368      Mixed_6a_Branch_1_Conv2d_0b_3x3_A\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_0_Conv2d_1a_3x3 (None, 8, 8, 384)    1152        Mixed_6a_Branch_0_Conv2d_1a_3x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_1_Conv2d_1a_3x3 (None, 8, 8, 256)    768         Mixed_6a_Branch_1_Conv2d_1a_3x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_0_Conv2d_1a_3x3 (None, 8, 8, 384)    0           Mixed_6a_Branch_0_Conv2d_1a_3x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_1_Conv2d_1a_3x3 (None, 8, 8, 256)    0           Mixed_6a_Branch_1_Conv2d_1a_3x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a_Branch_2_MaxPool_1a_3x (None, 8, 8, 256)    0           Block35_5_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Mixed_6a (Concatenate)          (None, 8, 8, 896)    0           Mixed_6a_Branch_0_Conv2d_1a_3x3_A\n",
      "                                                                 Mixed_6a_Branch_1_Conv2d_1a_3x3_A\n",
      "                                                                 Mixed_6a_Branch_2_MaxPool_1a_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    114688      Mixed_6a[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    384         Block17_1_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    0           Block17_1_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    114688      Block17_1_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    384         Block17_1_Branch_1_Conv2d_0b_1x7[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    0           Block17_1_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_0_Conv2d_1x1 ( (None, 8, 8, 128)    114688      Mixed_6a[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    114688      Block17_1_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_0_Conv2d_1x1_B (None, 8, 8, 128)    384         Block17_1_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    384         Block17_1_Branch_1_Conv2d_0c_7x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_0_Conv2d_1x1_A (None, 8, 8, 128)    0           Block17_1_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    0           Block17_1_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Concatenate (Concaten (None, 8, 8, 256)    0           Block17_1_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block17_1_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Conv2d_1x1 (Conv2D)   (None, 8, 8, 896)    230272      Block17_1_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_ScaleSum (Lambda)     (None, 8, 8, 896)    0           Mixed_6a[0][0]                   \n",
      "                                                                 Block17_1_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_1_Activation (Activatio (None, 8, 8, 896)    0           Block17_1_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    114688      Block17_1_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    384         Block17_2_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    0           Block17_2_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    114688      Block17_2_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    384         Block17_2_Branch_1_Conv2d_0b_1x7[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    0           Block17_2_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_0_Conv2d_1x1 ( (None, 8, 8, 128)    114688      Block17_1_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    114688      Block17_2_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_0_Conv2d_1x1_B (None, 8, 8, 128)    384         Block17_2_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    384         Block17_2_Branch_1_Conv2d_0c_7x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_0_Conv2d_1x1_A (None, 8, 8, 128)    0           Block17_2_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    0           Block17_2_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Concatenate (Concaten (None, 8, 8, 256)    0           Block17_2_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block17_2_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Conv2d_1x1 (Conv2D)   (None, 8, 8, 896)    230272      Block17_2_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_ScaleSum (Lambda)     (None, 8, 8, 896)    0           Block17_1_Activation[0][0]       \n",
      "                                                                 Block17_2_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_2_Activation (Activatio (None, 8, 8, 896)    0           Block17_2_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    114688      Block17_2_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    384         Block17_3_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    0           Block17_3_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    114688      Block17_3_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    384         Block17_3_Branch_1_Conv2d_0b_1x7[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    0           Block17_3_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_0_Conv2d_1x1 ( (None, 8, 8, 128)    114688      Block17_2_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    114688      Block17_3_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_0_Conv2d_1x1_B (None, 8, 8, 128)    384         Block17_3_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    384         Block17_3_Branch_1_Conv2d_0c_7x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_0_Conv2d_1x1_A (None, 8, 8, 128)    0           Block17_3_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    0           Block17_3_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Concatenate (Concaten (None, 8, 8, 256)    0           Block17_3_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block17_3_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Conv2d_1x1 (Conv2D)   (None, 8, 8, 896)    230272      Block17_3_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_ScaleSum (Lambda)     (None, 8, 8, 896)    0           Block17_2_Activation[0][0]       \n",
      "                                                                 Block17_3_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_3_Activation (Activatio (None, 8, 8, 896)    0           Block17_3_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    114688      Block17_3_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    384         Block17_4_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    0           Block17_4_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    114688      Block17_4_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    384         Block17_4_Branch_1_Conv2d_0b_1x7[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    0           Block17_4_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_0_Conv2d_1x1 ( (None, 8, 8, 128)    114688      Block17_3_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    114688      Block17_4_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_0_Conv2d_1x1_B (None, 8, 8, 128)    384         Block17_4_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    384         Block17_4_Branch_1_Conv2d_0c_7x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_0_Conv2d_1x1_A (None, 8, 8, 128)    0           Block17_4_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    0           Block17_4_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Concatenate (Concaten (None, 8, 8, 256)    0           Block17_4_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block17_4_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Conv2d_1x1 (Conv2D)   (None, 8, 8, 896)    230272      Block17_4_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_ScaleSum (Lambda)     (None, 8, 8, 896)    0           Block17_3_Activation[0][0]       \n",
      "                                                                 Block17_4_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_4_Activation (Activatio (None, 8, 8, 896)    0           Block17_4_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    114688      Block17_4_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    384         Block17_5_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    0           Block17_5_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    114688      Block17_5_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    384         Block17_5_Branch_1_Conv2d_0b_1x7[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    0           Block17_5_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_0_Conv2d_1x1 ( (None, 8, 8, 128)    114688      Block17_4_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    114688      Block17_5_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_0_Conv2d_1x1_B (None, 8, 8, 128)    384         Block17_5_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    384         Block17_5_Branch_1_Conv2d_0c_7x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_0_Conv2d_1x1_A (None, 8, 8, 128)    0           Block17_5_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    0           Block17_5_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Concatenate (Concaten (None, 8, 8, 256)    0           Block17_5_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block17_5_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Conv2d_1x1 (Conv2D)   (None, 8, 8, 896)    230272      Block17_5_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_ScaleSum (Lambda)     (None, 8, 8, 896)    0           Block17_4_Activation[0][0]       \n",
      "                                                                 Block17_5_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_5_Activation (Activatio (None, 8, 8, 896)    0           Block17_5_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    114688      Block17_5_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    384         Block17_6_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    0           Block17_6_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    114688      Block17_6_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    384         Block17_6_Branch_1_Conv2d_0b_1x7[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    0           Block17_6_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_0_Conv2d_1x1 ( (None, 8, 8, 128)    114688      Block17_5_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    114688      Block17_6_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_0_Conv2d_1x1_B (None, 8, 8, 128)    384         Block17_6_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    384         Block17_6_Branch_1_Conv2d_0c_7x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_0_Conv2d_1x1_A (None, 8, 8, 128)    0           Block17_6_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    0           Block17_6_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Concatenate (Concaten (None, 8, 8, 256)    0           Block17_6_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block17_6_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Conv2d_1x1 (Conv2D)   (None, 8, 8, 896)    230272      Block17_6_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_ScaleSum (Lambda)     (None, 8, 8, 896)    0           Block17_5_Activation[0][0]       \n",
      "                                                                 Block17_6_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_6_Activation (Activatio (None, 8, 8, 896)    0           Block17_6_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    114688      Block17_6_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    384         Block17_7_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    0           Block17_7_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    114688      Block17_7_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    384         Block17_7_Branch_1_Conv2d_0b_1x7[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    0           Block17_7_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_0_Conv2d_1x1 ( (None, 8, 8, 128)    114688      Block17_6_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    114688      Block17_7_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_0_Conv2d_1x1_B (None, 8, 8, 128)    384         Block17_7_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    384         Block17_7_Branch_1_Conv2d_0c_7x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_0_Conv2d_1x1_A (None, 8, 8, 128)    0           Block17_7_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    0           Block17_7_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Concatenate (Concaten (None, 8, 8, 256)    0           Block17_7_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block17_7_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Conv2d_1x1 (Conv2D)   (None, 8, 8, 896)    230272      Block17_7_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_ScaleSum (Lambda)     (None, 8, 8, 896)    0           Block17_6_Activation[0][0]       \n",
      "                                                                 Block17_7_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_7_Activation (Activatio (None, 8, 8, 896)    0           Block17_7_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    114688      Block17_7_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    384         Block17_8_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    0           Block17_8_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    114688      Block17_8_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    384         Block17_8_Branch_1_Conv2d_0b_1x7[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    0           Block17_8_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_0_Conv2d_1x1 ( (None, 8, 8, 128)    114688      Block17_7_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    114688      Block17_8_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_0_Conv2d_1x1_B (None, 8, 8, 128)    384         Block17_8_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    384         Block17_8_Branch_1_Conv2d_0c_7x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_0_Conv2d_1x1_A (None, 8, 8, 128)    0           Block17_8_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    0           Block17_8_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Concatenate (Concaten (None, 8, 8, 256)    0           Block17_8_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block17_8_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Conv2d_1x1 (Conv2D)   (None, 8, 8, 896)    230272      Block17_8_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_ScaleSum (Lambda)     (None, 8, 8, 896)    0           Block17_7_Activation[0][0]       \n",
      "                                                                 Block17_8_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_8_Activation (Activatio (None, 8, 8, 896)    0           Block17_8_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    114688      Block17_8_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    384         Block17_9_Branch_1_Conv2d_0a_1x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_1_Conv2d_0a_1x (None, 8, 8, 128)    0           Block17_9_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    114688      Block17_9_Branch_1_Conv2d_0a_1x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    384         Block17_9_Branch_1_Conv2d_0b_1x7[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_1_Conv2d_0b_1x (None, 8, 8, 128)    0           Block17_9_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_0_Conv2d_1x1 ( (None, 8, 8, 128)    114688      Block17_8_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    114688      Block17_9_Branch_1_Conv2d_0b_1x7_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_0_Conv2d_1x1_B (None, 8, 8, 128)    384         Block17_9_Branch_0_Conv2d_1x1[0][\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    384         Block17_9_Branch_1_Conv2d_0c_7x1[\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_0_Conv2d_1x1_A (None, 8, 8, 128)    0           Block17_9_Branch_0_Conv2d_1x1_Bat\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Branch_1_Conv2d_0c_7x (None, 8, 8, 128)    0           Block17_9_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Concatenate (Concaten (None, 8, 8, 256)    0           Block17_9_Branch_0_Conv2d_1x1_Act\n",
      "                                                                 Block17_9_Branch_1_Conv2d_0c_7x1_\n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Conv2d_1x1 (Conv2D)   (None, 8, 8, 896)    230272      Block17_9_Concatenate[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_ScaleSum (Lambda)     (None, 8, 8, 896)    0           Block17_8_Activation[0][0]       \n",
      "                                                                 Block17_9_Conv2d_1x1[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_9_Activation (Activatio (None, 8, 8, 896)    0           Block17_9_ScaleSum[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_1_Conv2d_0a_1 (None, 8, 8, 128)    114688      Block17_9_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_1_Conv2d_0a_1 (None, 8, 8, 128)    384         Block17_10_Branch_1_Conv2d_0a_1x1\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_1_Conv2d_0a_1 (None, 8, 8, 128)    0           Block17_10_Branch_1_Conv2d_0a_1x1\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_1_Conv2d_0b_1 (None, 8, 8, 128)    114688      Block17_10_Branch_1_Conv2d_0a_1x1\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_1_Conv2d_0b_1 (None, 8, 8, 128)    384         Block17_10_Branch_1_Conv2d_0b_1x7\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_1_Conv2d_0b_1 (None, 8, 8, 128)    0           Block17_10_Branch_1_Conv2d_0b_1x7\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_0_Conv2d_1x1  (None, 8, 8, 128)    114688      Block17_9_Activation[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_1_Conv2d_0c_7 (None, 8, 8, 128)    114688      Block17_10_Branch_1_Conv2d_0b_1x7\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_0_Conv2d_1x1_ (None, 8, 8, 128)    384         Block17_10_Branch_0_Conv2d_1x1[0]\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_1_Conv2d_0c_7 (None, 8, 8, 128)    384         Block17_10_Branch_1_Conv2d_0c_7x1\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_0_Conv2d_1x1_ (None, 8, 8, 128)    0           Block17_10_Branch_0_Conv2d_1x1_Ba\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Branch_1_Conv2d_0c_7 (None, 8, 8, 128)    0           Block17_10_Branch_1_Conv2d_0c_7x1\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Concatenate (Concate (None, 8, 8, 256)    0           Block17_10_Branch_0_Conv2d_1x1_Ac\n",
      "                                                                 Block17_10_Branch_1_Conv2d_0c_7x1\n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Conv2d_1x1 (Conv2D)  (None, 8, 8, 896)    230272      Block17_10_Concatenate[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_ScaleSum (Lambda)    (None, 8, 8, 896)    0           Block17_9_Activation[0][0]       \n",
      "                                                                 Block17_10_Conv2d_1x1[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Block17_10_Activation (Activati (None, 8, 8, 896)    0           Block17_10_ScaleSum[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_2_Conv2d_0a_1x1 (None, 8, 8, 256)    229376      Block17_10_Activation[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_2_Conv2d_0a_1x1 (None, 8, 8, 256)    768         Mixed_7a_Branch_2_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_2_Conv2d_0a_1x1 (None, 8, 8, 256)    0           Mixed_7a_Branch_2_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_0_Conv2d_0a_1x1 (None, 8, 8, 256)    229376      Block17_10_Activation[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_1_Conv2d_0a_1x1 (None, 8, 8, 256)    229376      Block17_10_Activation[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_2_Conv2d_0b_3x3 (None, 8, 8, 256)    589824      Mixed_7a_Branch_2_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_0_Conv2d_0a_1x1 (None, 8, 8, 256)    768         Mixed_7a_Branch_0_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_1_Conv2d_0a_1x1 (None, 8, 8, 256)    768         Mixed_7a_Branch_1_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_2_Conv2d_0b_3x3 (None, 8, 8, 256)    768         Mixed_7a_Branch_2_Conv2d_0b_3x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_0_Conv2d_0a_1x1 (None, 8, 8, 256)    0           Mixed_7a_Branch_0_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_1_Conv2d_0a_1x1 (None, 8, 8, 256)    0           Mixed_7a_Branch_1_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_2_Conv2d_0b_3x3 (None, 8, 8, 256)    0           Mixed_7a_Branch_2_Conv2d_0b_3x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_0_Conv2d_1a_3x3 (None, 3, 3, 384)    884736      Mixed_7a_Branch_0_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_1_Conv2d_1a_3x3 (None, 3, 3, 256)    589824      Mixed_7a_Branch_1_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_2_Conv2d_1a_3x3 (None, 3, 3, 256)    589824      Mixed_7a_Branch_2_Conv2d_0b_3x3_A\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_0_Conv2d_1a_3x3 (None, 3, 3, 384)    1152        Mixed_7a_Branch_0_Conv2d_1a_3x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_1_Conv2d_1a_3x3 (None, 3, 3, 256)    768         Mixed_7a_Branch_1_Conv2d_1a_3x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_2_Conv2d_1a_3x3 (None, 3, 3, 256)    768         Mixed_7a_Branch_2_Conv2d_1a_3x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_0_Conv2d_1a_3x3 (None, 3, 3, 384)    0           Mixed_7a_Branch_0_Conv2d_1a_3x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_1_Conv2d_1a_3x3 (None, 3, 3, 256)    0           Mixed_7a_Branch_1_Conv2d_1a_3x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_2_Conv2d_1a_3x3 (None, 3, 3, 256)    0           Mixed_7a_Branch_2_Conv2d_1a_3x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a_Branch_3_MaxPool_1a_3x (None, 3, 3, 896)    0           Block17_10_Activation[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "Mixed_7a (Concatenate)          (None, 3, 3, 1792)   0           Mixed_7a_Branch_0_Conv2d_1a_3x3_A\n",
      "                                                                 Mixed_7a_Branch_1_Conv2d_1a_3x3_A\n",
      "                                                                 Mixed_7a_Branch_2_Conv2d_1a_3x3_A\n",
      "                                                                 Mixed_7a_Branch_3_MaxPool_1a_3x3[\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    344064      Mixed_7a[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    576         Block8_1_Branch_1_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    0           Block8_1_Branch_1_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    110592      Block8_1_Branch_1_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    576         Block8_1_Branch_1_Conv2d_0b_1x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    0           Block8_1_Branch_1_Conv2d_0b_1x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_0_Conv2d_1x1 (C (None, 3, 3, 192)    344064      Mixed_7a[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    110592      Block8_1_Branch_1_Conv2d_0b_1x3_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_0_Conv2d_1x1_Ba (None, 3, 3, 192)    576         Block8_1_Branch_0_Conv2d_1x1[0][0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    576         Block8_1_Branch_1_Conv2d_0c_3x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_0_Conv2d_1x1_Ac (None, 3, 3, 192)    0           Block8_1_Branch_0_Conv2d_1x1_Batc\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    0           Block8_1_Branch_1_Conv2d_0c_3x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Concatenate (Concatena (None, 3, 3, 384)    0           Block8_1_Branch_0_Conv2d_1x1_Acti\n",
      "                                                                 Block8_1_Branch_1_Conv2d_0c_3x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Conv2d_1x1 (Conv2D)    (None, 3, 3, 1792)   689920      Block8_1_Concatenate[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_ScaleSum (Lambda)      (None, 3, 3, 1792)   0           Mixed_7a[0][0]                   \n",
      "                                                                 Block8_1_Conv2d_1x1[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_1_Activation (Activation (None, 3, 3, 1792)   0           Block8_1_ScaleSum[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    344064      Block8_1_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    576         Block8_2_Branch_1_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    0           Block8_2_Branch_1_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    110592      Block8_2_Branch_1_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    576         Block8_2_Branch_1_Conv2d_0b_1x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    0           Block8_2_Branch_1_Conv2d_0b_1x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_0_Conv2d_1x1 (C (None, 3, 3, 192)    344064      Block8_1_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    110592      Block8_2_Branch_1_Conv2d_0b_1x3_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_0_Conv2d_1x1_Ba (None, 3, 3, 192)    576         Block8_2_Branch_0_Conv2d_1x1[0][0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    576         Block8_2_Branch_1_Conv2d_0c_3x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_0_Conv2d_1x1_Ac (None, 3, 3, 192)    0           Block8_2_Branch_0_Conv2d_1x1_Batc\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    0           Block8_2_Branch_1_Conv2d_0c_3x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Concatenate (Concatena (None, 3, 3, 384)    0           Block8_2_Branch_0_Conv2d_1x1_Acti\n",
      "                                                                 Block8_2_Branch_1_Conv2d_0c_3x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Conv2d_1x1 (Conv2D)    (None, 3, 3, 1792)   689920      Block8_2_Concatenate[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_ScaleSum (Lambda)      (None, 3, 3, 1792)   0           Block8_1_Activation[0][0]        \n",
      "                                                                 Block8_2_Conv2d_1x1[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_2_Activation (Activation (None, 3, 3, 1792)   0           Block8_2_ScaleSum[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    344064      Block8_2_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    576         Block8_3_Branch_1_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    0           Block8_3_Branch_1_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    110592      Block8_3_Branch_1_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    576         Block8_3_Branch_1_Conv2d_0b_1x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    0           Block8_3_Branch_1_Conv2d_0b_1x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_0_Conv2d_1x1 (C (None, 3, 3, 192)    344064      Block8_2_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    110592      Block8_3_Branch_1_Conv2d_0b_1x3_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_0_Conv2d_1x1_Ba (None, 3, 3, 192)    576         Block8_3_Branch_0_Conv2d_1x1[0][0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    576         Block8_3_Branch_1_Conv2d_0c_3x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_0_Conv2d_1x1_Ac (None, 3, 3, 192)    0           Block8_3_Branch_0_Conv2d_1x1_Batc\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    0           Block8_3_Branch_1_Conv2d_0c_3x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Concatenate (Concatena (None, 3, 3, 384)    0           Block8_3_Branch_0_Conv2d_1x1_Acti\n",
      "                                                                 Block8_3_Branch_1_Conv2d_0c_3x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Conv2d_1x1 (Conv2D)    (None, 3, 3, 1792)   689920      Block8_3_Concatenate[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_ScaleSum (Lambda)      (None, 3, 3, 1792)   0           Block8_2_Activation[0][0]        \n",
      "                                                                 Block8_3_Conv2d_1x1[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_3_Activation (Activation (None, 3, 3, 1792)   0           Block8_3_ScaleSum[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    344064      Block8_3_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    576         Block8_4_Branch_1_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    0           Block8_4_Branch_1_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    110592      Block8_4_Branch_1_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    576         Block8_4_Branch_1_Conv2d_0b_1x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    0           Block8_4_Branch_1_Conv2d_0b_1x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_0_Conv2d_1x1 (C (None, 3, 3, 192)    344064      Block8_3_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    110592      Block8_4_Branch_1_Conv2d_0b_1x3_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_0_Conv2d_1x1_Ba (None, 3, 3, 192)    576         Block8_4_Branch_0_Conv2d_1x1[0][0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    576         Block8_4_Branch_1_Conv2d_0c_3x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_0_Conv2d_1x1_Ac (None, 3, 3, 192)    0           Block8_4_Branch_0_Conv2d_1x1_Batc\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    0           Block8_4_Branch_1_Conv2d_0c_3x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Concatenate (Concatena (None, 3, 3, 384)    0           Block8_4_Branch_0_Conv2d_1x1_Acti\n",
      "                                                                 Block8_4_Branch_1_Conv2d_0c_3x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Conv2d_1x1 (Conv2D)    (None, 3, 3, 1792)   689920      Block8_4_Concatenate[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_ScaleSum (Lambda)      (None, 3, 3, 1792)   0           Block8_3_Activation[0][0]        \n",
      "                                                                 Block8_4_Conv2d_1x1[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_4_Activation (Activation (None, 3, 3, 1792)   0           Block8_4_ScaleSum[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    344064      Block8_4_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    576         Block8_5_Branch_1_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    0           Block8_5_Branch_1_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    110592      Block8_5_Branch_1_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    576         Block8_5_Branch_1_Conv2d_0b_1x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    0           Block8_5_Branch_1_Conv2d_0b_1x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_0_Conv2d_1x1 (C (None, 3, 3, 192)    344064      Block8_4_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    110592      Block8_5_Branch_1_Conv2d_0b_1x3_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_0_Conv2d_1x1_Ba (None, 3, 3, 192)    576         Block8_5_Branch_0_Conv2d_1x1[0][0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    576         Block8_5_Branch_1_Conv2d_0c_3x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_0_Conv2d_1x1_Ac (None, 3, 3, 192)    0           Block8_5_Branch_0_Conv2d_1x1_Batc\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    0           Block8_5_Branch_1_Conv2d_0c_3x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Concatenate (Concatena (None, 3, 3, 384)    0           Block8_5_Branch_0_Conv2d_1x1_Acti\n",
      "                                                                 Block8_5_Branch_1_Conv2d_0c_3x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Conv2d_1x1 (Conv2D)    (None, 3, 3, 1792)   689920      Block8_5_Concatenate[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_ScaleSum (Lambda)      (None, 3, 3, 1792)   0           Block8_4_Activation[0][0]        \n",
      "                                                                 Block8_5_Conv2d_1x1[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_5_Activation (Activation (None, 3, 3, 1792)   0           Block8_5_ScaleSum[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    344064      Block8_5_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    576         Block8_6_Branch_1_Conv2d_0a_1x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_1_Conv2d_0a_1x1 (None, 3, 3, 192)    0           Block8_6_Branch_1_Conv2d_0a_1x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    110592      Block8_6_Branch_1_Conv2d_0a_1x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    576         Block8_6_Branch_1_Conv2d_0b_1x3[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_1_Conv2d_0b_1x3 (None, 3, 3, 192)    0           Block8_6_Branch_1_Conv2d_0b_1x3_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_0_Conv2d_1x1 (C (None, 3, 3, 192)    344064      Block8_5_Activation[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    110592      Block8_6_Branch_1_Conv2d_0b_1x3_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_0_Conv2d_1x1_Ba (None, 3, 3, 192)    576         Block8_6_Branch_0_Conv2d_1x1[0][0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    576         Block8_6_Branch_1_Conv2d_0c_3x1[0\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_0_Conv2d_1x1_Ac (None, 3, 3, 192)    0           Block8_6_Branch_0_Conv2d_1x1_Batc\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Branch_1_Conv2d_0c_3x1 (None, 3, 3, 192)    0           Block8_6_Branch_1_Conv2d_0c_3x1_B\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Concatenate (Concatena (None, 3, 3, 384)    0           Block8_6_Branch_0_Conv2d_1x1_Acti\n",
      "                                                                 Block8_6_Branch_1_Conv2d_0c_3x1_A\n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_Conv2d_1x1 (Conv2D)    (None, 3, 3, 1792)   689920      Block8_6_Concatenate[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "Block8_6_ScaleSum (Lambda)      (None, 3, 3, 1792)   0           Block8_5_Activation[0][0]        \n",
      "                                                                 Block8_6_Conv2d_1x1[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "AvgPool (GlobalAveragePooling2D (None, 1792)         0           Block8_6_ScaleSum[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "Dropout (Dropout)               (None, 1792)         0           AvgPool[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "Bottleneck (Dense)              (None, 128)          229376      Dropout[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "Bottleneck_BatchNorm (BatchNorm (None, 128)          384         Bottleneck[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "sequential_5 (Sequential)       (None, 5)            645         Bottleneck_BatchNorm[0][0]       \n",
      "==================================================================================================\n",
      "Total params: 22,808,789\n",
      "Trainable params: 22,779,957\n",
      "Non-trainable params: 28,832\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# TODO：使用 load_model 从`./models/facenet_keras.h5` 加载模型\n",
    "from keras.models import load_model\n",
    "\n",
    "# 模型底层使用 FaceNet 对原始图像进行建模，特征抽取\n",
    "# 加载预训练好的 FaceNet 模型。\n",
    "facenet_model = load_model('./models/facenet_keras.h5')\n",
    "\n",
    "# TODO：自己定义模型顶层，使用抽取后的特征进行人脸识别\n",
    "facenet_face = Sequential()\n",
    "#facenet_face.add(Dense(1024, input_shape=facenet_model.output_shape, activation=\"relu\"))\n",
    "#facenet_face.add(Dropout(0.5))\n",
    "\n",
    "facenet_face.add(Dense(5, activation='softmax'))\n",
    "\n",
    "facenet_face_model = Model(inputs=facenet_model.input, outputs=facenet_face(facenet_model.output))\n",
    "facenet_face_model.summary()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(279, 160, 160, 3)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(160, 160, 3)"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_XX = []\n",
    "for x in train_X:\n",
    "    x_image = Image.fromarray(x)\n",
    "    x_image = x_image.resize((160, 160))\n",
    "    train_XX.append(np.asarray(x_image))\n",
    "train_X2 = np.array(train_XX)\n",
    "print(train_X2.shape)\n",
    "train_X2[0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "279/279 [==============================] - 28s 101ms/step - loss: 0.7489 - accuracy: 0.7384\n",
      "Epoch 2/50\n",
      "279/279 [==============================] - 18s 66ms/step - loss: 0.3368 - accuracy: 0.8674\n",
      "Epoch 3/50\n",
      "279/279 [==============================] - 18s 66ms/step - loss: 0.3809 - accuracy: 0.8817\n",
      "Epoch 4/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.1595 - accuracy: 0.9534\n",
      "Epoch 5/50\n",
      "279/279 [==============================] - 18s 66ms/step - loss: 0.2733 - accuracy: 0.9355\n",
      "Epoch 6/50\n",
      "279/279 [==============================] - 19s 68ms/step - loss: 0.1965 - accuracy: 0.9355\n",
      "Epoch 7/50\n",
      "279/279 [==============================] - 19s 69ms/step - loss: 0.1648 - accuracy: 0.9355\n",
      "Epoch 8/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.1234 - accuracy: 0.9606\n",
      "Epoch 9/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0881 - accuracy: 0.9857\n",
      "Epoch 10/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.1289 - accuracy: 0.9642\n",
      "Epoch 11/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.1824 - accuracy: 0.9498\n",
      "Epoch 12/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.1059 - accuracy: 0.9677\n",
      "Epoch 13/50\n",
      "279/279 [==============================] - 19s 68ms/step - loss: 0.0685 - accuracy: 0.9785\n",
      "Epoch 14/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0801 - accuracy: 0.9677\n",
      "Epoch 15/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0162 - accuracy: 0.9964\n",
      "Epoch 16/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.1168 - accuracy: 0.9713\n",
      "Epoch 17/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.1101 - accuracy: 0.9642\n",
      "Epoch 18/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.1069 - accuracy: 0.9677\n",
      "Epoch 19/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0809 - accuracy: 0.9785\n",
      "Epoch 20/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0262 - accuracy: 0.9928\n",
      "Epoch 21/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0891 - accuracy: 0.9677\n",
      "Epoch 22/50\n",
      "279/279 [==============================] - 19s 68ms/step - loss: 0.1063 - accuracy: 0.9749\n",
      "Epoch 23/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0708 - accuracy: 0.9713\n",
      "Epoch 24/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0410 - accuracy: 0.9892\n",
      "Epoch 25/50\n",
      "279/279 [==============================] - 19s 68ms/step - loss: 0.0221 - accuracy: 0.9928\n",
      "Epoch 26/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0466 - accuracy: 0.9821\n",
      "Epoch 27/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.1531 - accuracy: 0.9606\n",
      "Epoch 28/50\n",
      "279/279 [==============================] - 19s 68ms/step - loss: 0.0205 - accuracy: 0.9892\n",
      "Epoch 29/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0260 - accuracy: 0.9892\n",
      "Epoch 30/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0515 - accuracy: 0.9821\n",
      "Epoch 31/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0344 - accuracy: 0.9928\n",
      "Epoch 32/50\n",
      "279/279 [==============================] - 19s 69ms/step - loss: 0.0880 - accuracy: 0.9821\n",
      "Epoch 33/50\n",
      "279/279 [==============================] - 19s 69ms/step - loss: 0.0597 - accuracy: 0.9821\n",
      "Epoch 34/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0628 - accuracy: 0.9785\n",
      "Epoch 35/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0441 - accuracy: 0.9785\n",
      "Epoch 36/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.1389 - accuracy: 0.9570\n",
      "Epoch 37/50\n",
      "279/279 [==============================] - 19s 68ms/step - loss: 0.0307 - accuracy: 0.9857\n",
      "Epoch 38/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0367 - accuracy: 0.9928\n",
      "Epoch 39/50\n",
      "279/279 [==============================] - 19s 66ms/step - loss: 0.0131 - accuracy: 0.9964\n",
      "Epoch 40/50\n",
      "279/279 [==============================] - 19s 66ms/step - loss: 0.0488 - accuracy: 0.9857\n",
      "Epoch 41/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0533 - accuracy: 0.9785\n",
      "Epoch 42/50\n",
      "279/279 [==============================] - 19s 68ms/step - loss: 0.0207 - accuracy: 0.9964\n",
      "Epoch 43/50\n",
      "279/279 [==============================] - 19s 68ms/step - loss: 0.0431 - accuracy: 0.9892\n",
      "Epoch 44/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0542 - accuracy: 0.9713\n",
      "Epoch 45/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0341 - accuracy: 0.9892\n",
      "Epoch 46/50\n",
      "279/279 [==============================] - 19s 66ms/step - loss: 0.0434 - accuracy: 0.9857\n",
      "Epoch 47/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0488 - accuracy: 0.9785\n",
      "Epoch 48/50\n",
      "279/279 [==============================] - 19s 68ms/step - loss: 0.0135 - accuracy: 0.9964\n",
      "Epoch 49/50\n",
      "279/279 [==============================] - 18s 66ms/step - loss: 0.0911 - accuracy: 0.9677\n",
      "Epoch 50/50\n",
      "279/279 [==============================] - 19s 67ms/step - loss: 0.0454 - accuracy: 0.9892\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.callbacks.History at 0x7f8ebc32b250>"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## 编译模型\n",
    "facenet_face_model.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "# 模型训练\n",
    "facenet_face_model.fit(train_X2, train_Y, batch_size=8, epochs=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "FaceNet Accuracy: 0.920\n"
     ]
    }
   ],
   "source": [
    "test_XX = []\n",
    "for x in test_X:\n",
    "    x_image = Image.fromarray(x)\n",
    "    x_image = x_image.resize((160, 160))\n",
    "    test_XX.append(np.asarray(x_image))\n",
    "test_X2 = np.array(test_XX)\n",
    "test_X2[0].shape\n",
    "\n",
    "preds_Y = facenet_face_model.predict(test_X2)\n",
    "correct = 0.\n",
    "for pr, y in zip(preds_Y, test_Y):\n",
    "    pr_cls = np.argmax(pr)\n",
    "    if y[pr_cls] == 1:\n",
    "        correct += 1\n",
    "accuracy = correct / len(preds_Y)\n",
    "print(\"FaceNet Accuracy: %.3f\" % accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**问题7：**评价 FaceNet 模型的效果，并指出为什么 FaceNet 比上一小节中 ResNet 的效果要好。\n",
    "\n",
    "**回答问题：**\n",
    "\n",
    "FaceNet收敛更快，其模型参数专门针对人脸训练过；"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**问题8：**\n",
    "- 首先你需要在下放画一个表格，将上面所做的实验结果都列出来。\n",
    "- 然后总结此项目，你认为在这个**人脸识别**项目中，哪些技术对识别准确率起比较重要的作用？请结合以上的实验结果分析。\n",
    "- 最后，你再简要说说还有哪些技术对人脸识别任务有较大的帮助？列出你的references\n",
    "\n",
    "**回答问题：**\n",
    "实验结果总结\n",
    "\n",
    "|Models|Accuracy|\n",
    "|---|---|\n",
    "|CNN|#0.44|\n",
    "|ResNet50 no-pretrain|#0.60|\n",
    "|ResNet50 pretrain|#0.56|\n",
    "|FaceNet|#0.96|"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "人脸识别项目中，对识别准确率起比较重要的作用的：\n",
    "\n",
    "1.人脸检测，准确的人脸检测；\n",
    "\n",
    "2.网络模型设计，更深，更复杂的网络，更有利于特征提取；\n",
    "\n",
    "3.训练数据，结果数据增强的数据集，训练效果更好；"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "references：\n",
    "\n",
    "https://arxiv.org/pdf/1503.03832.pdf\n",
    "\n",
    "https://www.cnblogs.com/lijie-blog/p/10168073.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}