Day022_Captcha_HW.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 『本次練習內容』\n",
    "#### 搭建一個能識別驗證碼的模型(CNN+CTC)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 『本次練習目的』\n",
    "  #### 了解如何結合CNN與CTC Loss\n",
    "  #### 有時間學員也可以延伸閱讀CNN+RNN+CTC模型\n",
    "  #### 沒有GPU的學員可以使用Colab，可以參考問答區的文章"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "ename": "ModuleNotFoundError",
     "evalue": "No module named 'termios'",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mModuleNotFoundError\u001b[0m                       Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-2-5b3bb6207c55>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m      1\u001b[0m \u001b[1;31m## 使用Colab可先run這裡的code\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      2\u001b[0m \u001b[1;31m#假如要連結自己的雲端硬碟再開啟\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 3\u001b[1;33m \u001b[1;32mfrom\u001b[0m \u001b[0mgoogle\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mcolab\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mdrive\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      4\u001b[0m \u001b[0mdrive\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mmount\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'/content/gdrive'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      5\u001b[0m \u001b[1;31m#檢查GPU\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32mD:\\Anaconda3\\lib\\site-packages\\google\\colab\\__init__.py\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m     21\u001b[0m \u001b[1;32mfrom\u001b[0m \u001b[0mgoogle\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mcolab\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0m_installation_commands\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     22\u001b[0m \u001b[1;32mfrom\u001b[0m \u001b[0mgoogle\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mcolab\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0m_shell_customizations\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 23\u001b[1;33m \u001b[1;32mfrom\u001b[0m \u001b[0mgoogle\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mcolab\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0m_system_commands\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     24\u001b[0m \u001b[1;32mfrom\u001b[0m \u001b[0mgoogle\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mcolab\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0m_tensorflow_magics\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     25\u001b[0m \u001b[1;32mfrom\u001b[0m \u001b[0mgoogle\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mcolab\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mauth\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32mD:\\Anaconda3\\lib\\site-packages\\google\\colab\\_system_commands.py\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m     22\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mlocale\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     23\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mos\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 24\u001b[1;33m \u001b[1;32mimport\u001b[0m \u001b[0mpty\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     25\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mselect\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     26\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0msignal\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32mD:\\Anaconda3\\lib\\pty.py\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m      9\u001b[0m \u001b[1;32mfrom\u001b[0m \u001b[0mselect\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mselect\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     10\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mos\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 11\u001b[1;33m \u001b[1;32mimport\u001b[0m \u001b[0mtty\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     12\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     13\u001b[0m \u001b[0m__all__\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;33m[\u001b[0m\u001b[1;34m\"openpty\"\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m\"fork\"\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m\"spawn\"\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32mD:\\Anaconda3\\lib\\tty.py\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m      3\u001b[0m \u001b[1;31m# Author: Steen Lumholt.\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      4\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 5\u001b[1;33m \u001b[1;32mfrom\u001b[0m \u001b[0mtermios\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[1;33m*\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      6\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      7\u001b[0m \u001b[0m__all__\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;33m[\u001b[0m\u001b[1;34m\"setraw\"\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m\"setcbreak\"\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mModuleNotFoundError\u001b[0m: No module named 'termios'"
     ]
    }
   ],
   "source": [
    "## 使用Colab可先run這裡的code\n",
    "#假如要連結自己的雲端硬碟再開啟\n",
    "from google.colab import drive\n",
    "drive.mount('/content/gdrive')\n",
    "#檢查GPU\n",
    "!nvidia-smi\n",
    "##安裝產生驗證碼的套件"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 記得先 pip install captcha"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 136
    },
    "colab_type": "code",
    "id": "AxGi02yvx1Go",
    "outputId": "e7bf7eb2-6699-441b-d724-6b5635ffa2bc"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting captcha\n",
      "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/90/fe/d4ddf1e6576073b5eaea76e9b2afa022c626212a30c871968480be3ccb7b/captcha-0.3-py3-none-any.whl (101kB)\n",
      "\r",
      "\u001b[K     |███▎                            | 10kB 28.6MB/s eta 0:00:01\r",
      "\u001b[K     |██████▌                         | 20kB 3.0MB/s eta 0:00:01\r",
      "\u001b[K     |█████████▊                      | 30kB 4.4MB/s eta 0:00:01\r",
      "\u001b[K     |█████████████                   | 40kB 2.9MB/s eta 0:00:01\r",
      "\u001b[K     |████████████████▏               | 51kB 3.6MB/s eta 0:00:01\r",
      "\u001b[K     |███████████████████▍            | 61kB 4.3MB/s eta 0:00:01\r",
      "\u001b[K     |██████████████████████▋         | 71kB 4.9MB/s eta 0:00:01\r",
      "\u001b[K     |█████████████████████████▉      | 81kB 5.5MB/s eta 0:00:01\r",
      "\u001b[K     |█████████████████████████████   | 92kB 4.4MB/s eta 0:00:01\r",
      "\u001b[K     |████████████████████████████████| 102kB 3.5MB/s \n",
      "\u001b[?25hRequirement already satisfied: Pillow in /usr/local/lib/python3.6/dist-packages (from captcha) (4.3.0)\n",
      "Requirement already satisfied: olefile in /usr/local/lib/python3.6/dist-packages (from Pillow->captcha) (0.46)\n",
      "Installing collected packages: captcha\n",
      "Successfully installed captcha-0.3\n"
     ]
    }
   ],
   "source": [
    "##!pip install captcha"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 產生驗證碼"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 254
    },
    "colab_type": "code",
    "id": "h2VVCBKwk6Zv",
    "outputId": "72cda421-1eca-46b1-d5bd-3b3d9e47eb08"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'AB1F')"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 203,
       "width": 369
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from captcha.image import ImageCaptcha\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import random\n",
    "\n",
    "%matplotlib inline\n",
    "%config InlineBackend.figure_format = 'retina'\n",
    "\n",
    "import string\n",
    "##驗證碼包含0-10數字以及26個英文字母\n",
    "characters = string.digits + string.ascii_uppercase\n",
    "print(characters)\n",
    "\n",
    "##設定產生圖片尺寸，以及總類別，n_class之所以要加一是為了留一個位置給Blank\n",
    "width, height, n_len, n_class = 170, 80, 4, len(characters)+1\n",
    "\n",
    "##設定產生驗證碼的generator\n",
    "generator = ImageCaptcha(width=width, height=height)\n",
    "\n",
    "##我們先練習固定長度4個字的驗證碼\n",
    "random_str = ''.join([random.choice(characters) for j in range(4)])\n",
    "img = generator.generate_image(random_str)\n",
    "\n",
    "plt.imshow(img)\n",
    "plt.title(random_str)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 自己搭建一個CNN模型使得output feature Map 高度被壓縮成小於4，8<寬度<20\n",
    "#### 解答的output feature Map 為 (batch_size,1,13,32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 136
    },
    "colab_type": "code",
    "id": "KX3Jq1sClGGn",
    "outputId": "ea40671a-fc3b-4758-8d29-4692252e840d"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:39: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"de...)`\n",
      "D:\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:50: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=[<tf.Tenso..., outputs=[<tf.Tenso...)`\n"
     ]
    }
   ],
   "source": [
    "from keras.models import *\n",
    "from keras.layers import *\n",
    "rnn_size = 128\n",
    "\n",
    "\n",
    "# CTC Loss需要四個資訊，分別是\n",
    "# Label\n",
    "# 預測\n",
    "# CNN OUTPUT寬度\n",
    "# 預測影像所包含文字長度\n",
    "def ctc_lambda_func(args):\n",
    "    y_pred, labels, input_length, label_length = args\n",
    "    return K.ctc_batch_cost(labels, y_pred, input_length, label_length)\n",
    "\n",
    "input_tensor = Input((height,width, 3))\n",
    "x = input_tensor\n",
    "\n",
    "'''自己設計CNN層'''\n",
    "x = Convolution2D(32, (3, 3), activation='relu')(x)\n",
    "x = BatchNormalization()(x)\n",
    "x = Convolution2D(64, (2, 2), activation='relu')(x)\n",
    "x = BatchNormalization()(x)\n",
    "\n",
    "\n",
    "##記錄輸出CNN尺寸，loss部分需要這個資訊\n",
    "##conv_shape=(Batch_size,輸出高度,輸出寬度,輸出深度)\n",
    "conv_shape = x.get_shape()\n",
    "\n",
    "##從(Batch_size,輸出高度,輸出寬度,輸出深度)變成(Batch_size,輸出寬度,輸出深度*輸出高度)，以符合ctc loss需求\n",
    "x = Reshape(target_shape=(int(conv_shape[2]), int(conv_shape[1]*conv_shape[3])))(x)\n",
    "\n",
    "x = Dense(units=32, activation='relu')(x)\n",
    "\n",
    "x = Dropout(0.25)(x)\n",
    "x = Dense(n_class, activation='softmax')(x)\n",
    "\n",
    "\n",
    "##包裝用來預測的model\n",
    "base_model = Model(input=input_tensor, output=x)\n",
    "\n",
    "\n",
    "##設定要給CTC Loss的資訊\n",
    "labels = Input(name='the_labels', shape=[n_len], dtype='float32')\n",
    "input_length = Input(name='input_length', shape=[1], dtype='int64')\n",
    "label_length = Input(name='label_length', shape=[1], dtype='int64')\n",
    "loss_out = Lambda(ctc_lambda_func, output_shape=(1,), \n",
    "                  name='ctc')([x, labels, input_length, label_length])\n",
    "\n",
    "##這裡的model是用來計算loss\n",
    "model = Model(input=[input_tensor, labels, input_length, label_length], output=[loss_out])\n",
    "\n",
    "##之所以要lambda y_true, y_pred: y_pred是因為我們的loss已經包在網路裡，會output:y_true, y_pred，而我們只需要y_pred\n",
    "model.compile(loss={'ctc': lambda y_true, y_pred: y_pred}, optimizer='SGD')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "id": "m1YRv0TVN_FL",
    "outputId": "c5b995c5-43b0-481b-8cd1-0fd5f85e4f03"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_3\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_2 (InputLayer)            (None, 80, 170, 3)   0                                            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_3 (Conv2D)               (None, 78, 168, 32)  896         input_2[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_3 (BatchNor (None, 78, 168, 32)  128         conv2d_3[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_4 (Conv2D)               (None, 77, 167, 64)  8256        batch_normalization_3[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_4 (BatchNor (None, 77, 167, 64)  256         conv2d_4[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "reshape_2 (Reshape)             (None, 167, 4928)    0           batch_normalization_4[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "dense_3 (Dense)                 (None, 167, 32)      157728      reshape_2[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "dropout_2 (Dropout)             (None, 167, 32)      0           dense_3[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "dense_4 (Dense)                 (None, 167, 37)      1221        dropout_2[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "the_labels (InputLayer)         (None, 4)            0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_length (InputLayer)       (None, 1)            0                                            \n",
      "__________________________________________________________________________________________________\n",
      "label_length (InputLayer)       (None, 1)            0                                            \n",
      "__________________________________________________________________________________________________\n",
      "ctc (Lambda)                    (None, 1)            0           dense_4[0][0]                    \n",
      "                                                                 the_labels[0][0]                 \n",
      "                                                                 input_length[0][0]               \n",
      "                                                                 label_length[0][0]               \n",
      "==================================================================================================\n",
      "Total params: 168,485\n",
      "Trainable params: 168,293\n",
      "Non-trainable params: 192\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras import backend as K\n",
    "import keras\n",
    "\n",
    "##CTC Loss需要四個資訊，分別是\n",
    "## Label\n",
    "## 預測\n",
    "## CNN OUTPUT寬度\n",
    "## 預測影像所包含文字長度\n",
    "def ctc_lambda_func(args):\n",
    "    y_pred, labels, input_length, label_length = args\n",
    "    return K.ctc_batch_cost(labels, y_pred, input_length, label_length)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "5Jzb1YnqlJug"
   },
   "outputs": [],
   "source": [
    "##設計generator產生training data\n",
    "## 產生包含要給loss的資訊\n",
    "## X=輸入影像\n",
    "## np.ones(batch_size)*int(conv_shape[2])=CNN輸出feature Map寬度\n",
    "## np.ones(batch_size)*n_len=字串長度(可浮動)\n",
    "\n",
    "def gen(batch_size=128):\n",
    "    X = np.zeros((batch_size,height, width, 3), dtype=np.uint8)\n",
    "    y = np.zeros((batch_size, n_len), dtype=np.uint8)\n",
    "    while True:\n",
    "        generator = ImageCaptcha(width=width, height=height)\n",
    "        for i in range(batch_size):\n",
    "            random_str = ''.join([random.choice(characters) for j in range(4)])\n",
    "            X[i] = np.array(generator.generate_image(random_str))\n",
    "            y[i] = [characters.find(x) for x in random_str]\n",
    "        yield [X, y,np.ones(batch_size)*int(conv_shape[2]), np.ones(batch_size)*n_len], np.ones(batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 225
    },
    "colab_type": "code",
    "id": "dfcFs6InVQaT",
    "outputId": "3f186d49-2e30-4efd-87c0-0670a58bc4c1"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Label:  [[14 24 22 26]]\n",
      "CNN輸出寬度:  [167.]\n",
      "字串長度(可浮動):  [4.]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 191,
       "width": 369
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "next_ge=gen(batch_size=1)\n",
    "test_ge=next(next_ge)\n",
    "plt.imshow(test_ge[0][0][0])\n",
    "print('Label: ',test_ge[0][1])\n",
    "print('CNN輸出寬度: ',test_ge[0][2])\n",
    "print('字串長度(可浮動): ',test_ge[0][3])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 開始訓練"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "id": "fqdKnqollUKj",
    "outputId": "3da34151-a9df-4287-ead6-50c414d30e65"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1\n",
      "300/300 [==============================] - 440s 1s/step - loss: 47.2272\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.callbacks.History at 0x2c974389b00>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit_generator(gen(32), steps_per_epoch=300, epochs=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 預測"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAuIAAAGzCAYAAABq2XLwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAAWJQAAFiUBSVIk8AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAgAElEQVR4nOzdeXxk91nn+++jrUp7726323a39zWJlwTb2UMIJIADSRjCZQnrXPbV3JnXEGZyGS4DdzxsmQFmgCQsAwlJZvBkIBskIYsTQuyY2I6XtN12t91272q1llpU9Zs/qtSq55FU1SWVdLR83q9Xv6SnzqlTR0enTv10+nueYyklAQAAAFhdXVmvAAAAALAZMRAHAAAAMsBAHAAAAMgAA3EAAAAgAwzEAQAAgAwwEAcAAAAywEAcAAAAyAADcQAAACADDMQBAACADDAQBwAAADLAQBwAAADIAANxAAAAIAMMxAEAAIAMMBAHAHSUme0zs2RmKet1AYC1jIE4AAAAkAEG4gAAAEAGGIgDAAAAGWAgDgAAAGSAgTgArEFm9lT9gsdXmdklZvZHZnbYzApmdtDM7jaz0QWe9576895hZjkz+yUz+4qZna0/viXMv8/M3mlmj5nZVH2++8zsX5nZYJP1y5vZL5vZo/V1es7M3mtm163E9gCAjagn6xUAADR1haS/krRT0oSkJGmfpF+Q9EYze0VK6bkFnpeX9GlJL5FUljQVZzCzN0n67/V5JWlaUp+km+v/vtvMviGldDQ8b0jS30n6uvpDJUkDkr5T0rdI+pHFfhgz2yfpYL38gZTSexabFwA2Os6IA8DadrekM5JenlIaljQo6dsknVBtkP4nizzvJyRdJemtkoZSSltUG8BPSpKZvVjSeyX1SvoNSZfWlz0g6TZJ/yjpRkl/usCyf0u1Qfi0pB+oL39U0gslPSLp95fzAwPAZsFAHADWtpyk16eUPitJKaVqSukeSf+iPv0bzOxlCzxvSNJ3ppTel1Iq1Z/7dEqpXJ/+W6oNwn8xpfSvU0qHUk0lpfSPkl4v6Yik15nZrbMLNbNLJf1gvfzxlNJ7ZpeZUvqKpG9U7Qw5AKAFBuIAsLb9VUrpQHwwpfRJSffWy7cs8LyvpJQ+ttACzexySS9V7Yz2Hyw0T0rptKQP18tvaJj0JtU+O45ogbPlKaVTanJGPKX0VErJ6v/es9h8ALAZkBEHgLXtU02m/YOkO1TLc0efb/K8O+pf+yQdNLPF5huqf7244bHZ1/pMSqnaZL0AAC0wEAeAte3Z85i2c4Fpx5s878L6125JF5zHOgw0fD/7WkfOY70AAE0wEAeA9WvRU9mSKk2mzcYSv5xSWuhs+nI1Wy8AQB0ZcQBY2/Y0mTZ7ZrvZ2e+FzLYjvNLM2j0hM/ta57NeAIAmGIgDwNr2yvOYdn+by5zNjw9Jel2bz519rZfZ4uHyZusMAKhjIA4Aa9t3mtll8UEze4VqnU8k6f3tLDCl9KikL9TL32hxB81+M8s1PPQ/JFUlXSTpexaYf6ukH21nfQBgs2IgDgBrW0nSh83sDkkysy4z+1ZJH6hP/3hK6XNLWO5PSSpKukHSZ8zstbMxlfprXG9mb5f0hBqiJimlpyW9q17+gZl9n5n11p93o6SPaO5OnfOY2T4zS/V/37+E9QaADYOBOACsbXdJ2irpc2Z2VrXb3P8v1bqXHJD0tqUsNKX0JUnfrtpdO2+S9HFJk2Z2QlJB0kOS/r1qg/AUnv5zqt15c0C1O3ueNbMxSV+RdL2kH1vKOgHAZsNAHADWtgOSblXtLPQZ1VoOPiXpP0m6NaX03FIXnFL6sKSrJP2qatnvgqQtksZVu1nQv5V0bf0seOPzJiS9qj798frDBUnvk/QSNe9hDgCos5TiiQ4AQNbM7ClJl0p6dUrpU9muDQBgJXBGHAAAAMgAA3EAAAAgAwzEAQAAgAwwEAcAAAAywMWaAAAAQAY4Iw4AAABkgIE4AAAAkAEG4gAAAEAGGIgDAAAAGWAgDgBNmNn3m9k7zOxFWa9LVszsU2aWzOw9Tea50swO1ed7ysz+Zf37pfz7nvoyX3ue838prMsPt/FaH1jhzQcAi+rJegUAYI37fkmvlPSUpAcyXZM1ysyulfT3ki6UdEDS10t6maSjC8zeLWlH/ftTksoLzDO9wGPHJVUXWYUTTVZvoXVodLrFdABYMQzEAQBLZmYvkPR3knZKekTSa1NKRyT9Rf1fnP8KSV+rl29MKX32PF/q5pTSM22uXiWltLvN5wDAqiGaAgBYEjO7RdInVRuEPyjpVfVBOADgPDAQB4AF1LPhSbVYiiS9O2SLn4rzmtmn6vV3m9k/mNnJ+uPfVn/8PfX6HU1edzaP/f2LTO8zs580s8+Y2SkzK5rZ02b2rnpEZFWY2e2qxVG2SbpPtUH4sdV6fQDYCBiIA8DCplXLF89mmMfr9ey/4ws9ycx+V9Kfq5aRNi2ea26bmV0o6YuS3llf/qikoqRLJP2ApPvN7E2LPDe1+iOgjfV4haSP1V//85K+PqV0arnLBYDNhoE4ACwgpfS+er743vpDP5NS2t3w78ULPO0WST8p6d9J2p5S2iZpa8MylszMeiXdI+mFkj4t6RWS+lNKI5J2S/pPkvKS/szMLl/u6zVZj9dK+rCkIUn/IOl1KaUzK/V6ALCRcbEmAHTOkKT/kFL6ldkHUkrjqp1NX663SXqxpH9SbfBbbHiNo5LuMrMBST8m6edU+4Og014i6TtVG/B/XLWLLRfqcLIS7jezxf534bKU0tQCj3eb2fMtlvs9KaW/W+a6AcCSMBAHgM6pSPrNFVr22+pf/0vjIDz4C9UG4t8QJ6SUrAPrMJtBL0n6l6s4CJdqF4Quptn/7l7QYrn5JawLAHQEA3EA6JwDKaVmPa2XxMx6VDsbLUm/aWa/scis3fWvF3d6HeoelHSFpH5Jf2Nmr0opLZiVXwEXL7F9IZ9zANYsMuIA0DkrNSjdJqmv4fsLFvk3e6Oc/hVaj/slvVG1C0Svk/RxM9u6Qq8FABseA3EA6JzKCi238Vj9wpSStfq3QuuhlNLHJb1ZtXjKCyV91MxGVur1AGAjYyAOAKtnpv61WS55dIHHTmpukH9dR9doCVJKfyPprar9PC+W9GEzG8p2rQBg/WEgDgDNzXbq6MRZ5rH6170LTTSzQc1dEHlOSqks6Uv1csE+4astpfQ/JX2van8g3CHpQ2a2UpEYANiQGIgDQHOzrQe3dGBZD9a/vs7MFjor/nOScos89z31r282s1c3e5HVym2nlN4r6QclJUmvkvTXZrbY+gMAAgbiANDcw/WvbzKzhWIj7fiQanfs3CnpT81slySZ2aiZ/ZKkd0ha7OY4fyzpC6odt/+3mf2MmW2bnWhmu8zsu8zsU5J+Jj65k3fWbJRS+lNJ/7dqg/HXSXp//eZDAIAWGIgDQHN/ptqFiS+TdMLMnjWzp8zss+0uqH4b+H9dL79D0lEzOy3plKRflfQrkh5Y5Lll1TqWfE7SgKTfrq/PKTM7K+moan3EX6naoHjVpJT+UHOD/2+V9Bdm1t3kKaul28yeb/Hv81mvJIDNi4E4ADSRUnpUtRvkfES1s9W7JV2qRXLe57G831Xt7pRfkDSl2nH4c5K+vfGOnIs895hqA+3vlvS3ko6pdjdPk/SoamfN3yDp15aybsuRUnqnpF+sl29R7Yz/WviMWazV4+y/ZjcKAoAVZSmt6okTAAAAAOKMOAAAAJAJBuIAAABABhiIAwAAABlgIA4AAABkgIE4AAAAkAEG4gAAAEAGGIgDAAAAGch0IG5me83sXWZ2xMyK9bvV/baZbc1yvQAAAICVltkNfczsckn3Stol6R7V7gr3EkmvlvSYpJemlE5msnIAAADACuvJ8LV/T7VB+E/Xb40sSTKz35T0c5L+P0k/upQFm9lBSSOSnlr+agIAAACL2idpPKW0v90nZnJG3Mwuk/SEagPly1NK1YZpw5Kek2SSdqWUJpew/JP9/f3brr366g6tMQBsDoePf9XVF++8LqM1aV9c91ay/tnW87bezPi9IXrkscc0PT19KqW0vd3nZnVG/DX1rx9rHIRLUkrprJl9TtLrJN0m6e8XW4iZ3bfIpPy1V1+tL37m3o6sLABsFj/ze35Q8Ts/vn6Oo3HdW8n6Z1vP23oz4/eG6CUvv0P3P/DAU0t5blYXa86eqn58kelfq3+9ahXWBQAAAFh1WZ0RH61/PbPI9NnHtzRbSErploUer58pv3lpqwYAAACsvCwv1mzG6l+zaekCABvEZGHM1cfGDrr6nnvvdvX+3Te5+i8/+XZX33n7Xa4ezDc9X9JR8Wf5X59vvu7RG+/w6x6Xt9yfZSW39Wpu581mI71HsP5kFU2ZPeM9usj0kTAfAAAAsKFkNRB/rP51sQz4lfWvi2XIAQAAgHUtq4H4J+tfX2dmbh3q7QtfKmla0hdWe8UAAACA1ZBJRjyl9ISZfUy1FoU/IemdDZP/X0mDkv7rUnqIAwAW94VHPtjW/Ldd++YVWpPli+sWf7Y4fdcWf6+Nlc7ubqRtvZnwe8NqyvJizR9X7Rb3v2tmXy/pEUlfp9ot7h+X9EsZrhsAAACworKKpiil9ISkWyW9R7UB+C9IulzS70q6PaV0Mqt1AwAAAFZapu0LU0qHJf1AlusAAAAAZGGt9hEHAKyAtZ6rbia+dly32L+51fNX2nre1psZvzespsyiKQAAAMBmxkAcAAAAyAADcQAAACADllLKeh06zszuu/lFL7r5i5+5N+tVAYA1ZbIw1tb85F2Xjm29PvF7Q7te8vI7dP8DD9yfUrql3edyRhwAAADIAANxAAAAIAMMxAEAAIAM0EccADYR8qyrh229PvF7w2rijDgAAACQAQbiAAAAQAYYiAMAAAAZICMOYEW004uXTCYAYC1p9RnWqc8tzogDAAAAGWAgDgAAAGSAgTgAAACQATLiADoi5umOjR109Rce+aCrb7v2zee+37Vlv5tGZhwAsJra/Qz7rlf/akdelzPiAAAAQAYYiAMAAAAZYCAOAAAAZICMOIAlaZWnu+feu5s+vzFvd+ftd3VuxQAAaGG5n2E/83vXnfv+8PEjS14PzogDAAAAGWAgDgAAAGSAgTgAAACQATLiADoi9lhtpbGPONauQnHK1VPFM6Eed3U1lVd8nWYN5Ed9nRsJtZ8OAItp9zPsd378q+e+//xf3qHjzz6wpNfljDgAAACQAQbiAAAAQAaIpgDoiBg1aXZLe8nf1p5b2q9dMZpy3yOfdPUzY/6/Y2fS9Iqv06wYPbnlite7esfoJWF+oisAFtbuZ1incEYcAAAAyAADcQAAACADDMQBAACADJARB7AkMdfdmPmWWt+2nlz4+hDbFZ44c9jVh4495OosM+InzhxydcyIkyEHMGu5n2GdwhlxAAAAIAMMxAEAAIAMMBAHAAAAMkBGHMB5mSyMtTU/GfCNoaoZV09Vj/rpyU9fTTG/Pr8ed3W7GfKYER/Ih5pMObBhZPWZxRlxAAAAIAMMxAEAAIAMMBAHAAAAMkBGHMCCYib82NhBV3/hkQ+6+rZr3+zq2JOVzPj6UChOubpV7rqqcoslWptrkNqcf3HLzZDPz4T7mr7kAJaLM+IAAABABhiIAwAAABlgIA4AAABkgIw4AEmtM+H33Ht30+fHzPidt9/VmRXDqiqUpl392MF/dnWx5DPk7epKvf4Bi5nw1PBdNUypqpNaZcg1ftiVMeNNX3IAy8UZcQAAACADDMQBAACADDAQBwAAADJARhzAgmLmO2bIJwunXR37iMfp9BFfH4pFnxEvhsx4IUxvxUIfcVO3q3vSoKv7uvPnvq92Fd202LO8VPF59UoqtbVu7VrtvuSN08mTAxsTZ8QBAACADDAQBwAAADLAQBwAAADIABlxAJLmZ7r3777Z1Q8e/ISrB/NbXX3w+ftdfeP+13Rw7bBqLJadPV/TFTLiuW6fbR7MbT/3/Z5tV7hp41PHXX1i+nFXlyo+v16uFlxdqfrMeSX5erk63Ze8MRc+P0/+BlcP5vw1GHH+/vxwmO4z5lmL22q6MLHovAN5v+79ueFF5gTWPs6IAwAAABlgIA4AAABkgIE4AAAAkAEy4osoFCddHfvDxjxbNc24ur8hfxezejGbZxZCmUAGWmW+d23Z3/T5sY841qvmx6PYF3y5y4/Hw327X3Tu+91bLnHTSqGn+ZYzF7p6ouSPy13dydVjhSdcHY/r0+G4PhP6ks+EjHncEknR/Ef8659/pjx+jpw884yrh/t2uXoov93Vd9zwJld3jfqP/3xuoOm6tqtVj/WpQvOf/csHPubqnjSXA7/9hje6adtH/TnF/pzvTQ+sZZwRBwAAADLAQBwAAADIAANxAAAAIANkxBcxXZxy9Ze++veufnbsK66uyGcXtw9dfO77267x2dn+Pt/zdF4usIXB/JbWMzWYDnn3VsjXQZqf+f7CIx9sOj1myON+2iwzGvOijf2Tpfk54piXxQrq9DUsYXH9+SFXX3np9ee+v2Crz4iXyz6z3fdsv6sn5adfsvMiV8c+4geeftDV06Wzrn5+/GFXz79WyNeV0Ld8Jvm6ldQkUz6vz3bR99k+XT3m6gtGL3f1w0/8o6tfcbPPjLerVQY89kS/78CHXR3vWxCfXyz631XO5q5huf+Rf3DTXn6Tz4zzGYb1hDPiAAAAQAYYiAMAAAAZYCAOAAAAZGDZGXEz2y7p2yV9s6QbJV0kqSTpQUnvlvTulFJ1gefdIentkm6TlJd0QNK7JL0zpVRZ7notVyFkxE+NH3X1M8cfc3U5+WxhZWbuR5gsjLlpE9M+G3fyrM/SLTeLWwi9dk+Hdf/Kgc+6ev/ea1y9fdT35iWfuznE/SjuZ3feflfT58d++DFDeuLMYVffd+Bv56aN+2k7Ri529ctv+C5Xsw+unJaJ8A5HxnO9OVcP5OfyvSPDfp8sl31u+MrLb2y67L6wbCWfwd454jPo83PPL3X1gUM+Uz5V8Lnmo2e/2nR504XwOREy66Xkj93NtrWlMDGUMW++f++1iy9My+/7HTPgMSM+/3PwlKtLM/5nH+zd4eq+3rk+6TNlP6QolvznNbCedOJize+Q9PuSnpP0SUmHJF0g6U2S/kjS683sO1KaOwKa2RslfVBSQdL7JJ2S9K2SfkvSS+vLBAAAADasTgzEH5d0p6S/aTzzbWb/RtIXJb1ZtUH5B+uPj0j6Q0kVSa9KKX2p/vgvS/qEpLeY2VtTSu/twLoBAAAAa9KyM+IppU+klD4U4ycppecl/UG9fFXDpLdI2inpvbOD8Pr8BdWiKpL0Y8tdLwAAAGAtW+k+4uX615mGx15T//qRBeb/tKQpSXeYWS6lVFxgnkxYCOBZ6p43R6Niea5/7Knx58I0n4X7+Jd/v+lrx8x4zOqWZ8quPjvhM+hffMj3QD86/oirv3b0M66OGfFbr/hWV8f8bn/e90Wnh+vG0E4fcGl+JvSRw59z9c5Rn8c9dOyhueeGjHgU86kK+yBWTjz2tRsSt3C+pyscO+P0ZnpD5jvW7cr15V09POSvPSiWfB/weOxrlSl/4tBDfnrMlIdjcVxesaFv+kzIk1ervmd6NVxalc/743A+N9C0ju/Bzzz0l02nz8+M+58t9gmvVv36FULP9t6u/lD79Wvc9qODPj+e5zMHHRCvY2il3Xu6LGbFBuJm1iPp++pl46D76vrXx+NzUkozZnZQ0vWSLpP0SJwnvMZ9i0y6ZpHHAQAAgDVhJdsX/rqkGyT9bUrpow2Pz55yWOx2krOPd+ZPDQAAAGANWpEz4mb205J+QdKjkr633afXvy5+r9/ZGVK6ZZHXv0/SzW2+LgAAALBqOj4QN7OfkPQ7kr4q6etTSqfCLLNnvBdrBDwS5lsjQkZc558RP/jMo27aqekn2nrl2Ec8Koae5wdCLvH5Uz4FdGLyKVeX5bN90yEnNVPwfxPt23WTq198wzc2XT+sTSvdNzh69PC9rm7MHveFPGia8f9ZV62E91fI7sasL5aukvw1J9V5tc8mtxKPlT02FKav3fvKxf2qVaa8VPI57h2jIVMeMuLTIVP+tXDsLhbn9vNKuGTq2NlwL4uS/70M9Gz308O1RDHz3XjNxkJ1q+s4Wn1GzoRro/q6/X4wOrDT1TfuvdPVe3Zcee77Sy+80k3L9/njB3A+Yib82NhBV7e6p8v+3X4stFQdPQKa2c9K+s+SHpL06nrnlGj26HHVAs/vkbRftYs7n+zkugEAAABrSccG4mb2r1S7Ic8Dqg3Cjy0y6yfqX79pgWmvkDQg6d611DEFAAAA6LSODMTrN+P5dUn3qRZHOdFk9g9IOiHprWZ2a8My8pJ+tV427+cHAAAArHPLzoib2dsk/Ypqd8r8jKSfNpvXZ/aplNJ7JCmlNG5mP6LagPxTZvZe1W5xf6dqrQ0/oNpt79eY9jLiheLEue+nS5Nu2gsv85nqp098ydUxh7Rry35Xx96VJ077BNDRU0+5eqLg/3OiIr8+3ep1tUKf31T1f6/tu+g6Ye1rt+93zHzPe36LvsFx/vie6OnyPZ8b+0kP9u5203bkr3X1M8/6dY39nMmIz2m3F25Pl9928fd6fNJfY1JpmRFvfqzsUuwjHo+l61dfX65pPTwY+5T7//jdPrLXT2+41qgS+oafHL/R1Q8f+KKr5/cJf9rVn3/sr/z0cDyI+0HU2+37fsfrPPq6RlydwsdMRf46jxdc+gZXX3/5ba7eseWiudfu9Qvr6V7pW6JgI2iVCb/n3rubPj9mxjuVEe/E3js7SuyW9LOLzPMPkt4zW6SU/trMXinplyS9WVJe0gFJPy/pd1NKLTumAAAAAOvZsgfiKaV3SHrHEp73OUlvaDkjAAAAsAGt3b5RAAAAwAZGsGoR/blBX/f5nqf5nK+LRd8uvZpmzn3f1VNx03Zu8fnWF175qqbrEjPhUXnGZ+1KMz4DXqxMuDppxtVd5rN+2/KXu/rGfa9x9eig7/catxXWhpjx/MxDf+nq2Ce4dea7FZ8NjpnRob5drr5w5Ia5Z3ZV3bQdo5e4evvoRcLCltsLN9/jjy+PP/WAq+N+0CojHq8QanU9zWbKIbbK73eFT+SB3rn8/lTRb/dy1fflHiv53/vkWf97e/yYz6OXK/7500X/OWHh2qGY+e4P+83FW2519XC/72PeG3Lc+bz/DN0yvMPVO7bs8a+X959TwHLFY2Mrre7pslScEQcAAAAywEAcAAAAyAADcQAAACADZMQXEXuwXnf5La4+Mvawq88Ufe6x2pDDnqn4bF+u1y+7P+ezd13W3t9H1ZD5nq6cdHVFPguYQkaz23wf4aG8z4AP5La6uj8/3Nb6IRux7/eJ8cNN64XSvb6KWd+qmunt8pnOS7b6DOlFO6469/3FF/rrEnIhDzqQ8/tcb29f09feyDrdC/f2q7/b1aWyv+akWG73Jsdxv4n7VTy+zbvvxIax3Px+l8295+5/4iNuWqtrPKbn9QEPv5dwv4ge+Wt98j3bXD2c89d47N56tat3jPjrOEaH/OfIhRf46fE93NsTeoOHGui0+H5r9X6M93TpFM6IAwAAABlgIA4AAABkgIE4AAAAkAEy4ouIGfH+kFGNtYW/aRp7dVdT2U0rFH0GM1VDJ90YxQ2KJf/86eJZX5d8Nrixp7kkmflfe698Rr23x2fGR0d8v9jurhYriDXCZ0K7YiY0XBvQG/p+93X7Pr8pdHyuVsN+XfH7XVdoitzT7TOfjb3Bd27zPYP7+nJaS1r1f45a9f7vpGX3wq2G7LD5/aT9BHcKVbiWwFp0Drf1mxlfbn7/kw/8qatfdPk3nvv+8LFH3LT513h4XeEzKWfhOose//7OdY+6Ot/j+4Dv3XaDqy/Y7u+HcdHuS109MuyX1yoTDqy0eFyOme87b7+rred3CmfEAQAAgAwwEAcAAAAywEAcAAAAyAAZ8fPmc4s9lgtT/d801YacZEo+Izm/bp6ZnJnxWdyzkz6L++iTX3b1ZPF0WIJft175rGDsGz484LOBA/0+S9jdTUZ8Peg2n8GMGdDRvr2u3j5whauH+ne42kJ/+7GpZ1393MQDrq6G/dy6/HtoZGRufbq719ahaLn9n2P2cCUz48vthTs97fuEz09ot5fZbpEAX5UlrBXt5vdfuP+bXP3ggXvPfV8oTTV9brf5DHauy/cFH+zxfcD3jL7A1f15f3wY7Pf77MiA7yu+fcsFrt466qfncv4aFGCtWc1reZrhjDgAAACQAQbiAAAAQAYYiAMAAAAZWFvBzDWsK/bWnVeHv2kact/JfFa2K2RlW0UwyyEjfujZJ1x9YuwZV08XJ/3rhb7hfeYz39sGfVb46v0vdHU+1998BbEm9ed85vPibTf7esstrh7o8/N3d/uMZ1evz5yfOOP3ozPlJ109XfCZVuvy75HGaw2yvu5guf2fYxa4VT/a5eh0L9zp6aOunncsWzaf+e7P+exyfz7UOb9frWft5vcr5dD7u3vuep5icbrpa8WM+Pb+q1194eh1rh4Z9NeADA9sdfUley/z6xJ6+/f2+Nfr7mE4ASwFZ8QBAACADDAQBwAAADLAQBwAAADIAKGu8+aD3DF3HfuIexU/77yIuH+gWvWZ8snpcVePTx139cT0KVen5DPlXfJZvtGBPa7eNrLb1VuGfV/xXB8Z8fVoID/i6puvfZWrY3/67q6wT7fICg8M+uzvgeN+P5ma9tcqpKp/H1SrM02Xn6V2+z/HrO9qWm4v3Hj8mTc9HrBain3AfZ0PGfEbL7/D1TFDvp4sN78/MTHh6t7uufdUvC4pbuYu+Ws48j1+XbYMX+jqraP+OH/RBf5aoaFBf7+J3l7/OQKgMzgjDgAAAGSAgTgAAACQAQbiAAAAQAbIiC9RzIRbWrwP8nTZZ2ULZd9fuVjydV+v7988Nn7C1YdPPOTqyfIxV/d2Dbh6ML/d1bu2XurqG6++zdX5nH99rE/5kO2P9XKNTfksb77b9yGvVJ9z9WTppJ8ermVYS9rt/xyzwMvNbWep833EvZgBj5nxjdRHvN39YHLSfxZ0Nfwu4r0s4nVL1RSuwepUXpIAACAASURBVFDJL6vb/14HQv/2vj5/3CcTDqwOzogDAAAAGWAgDgAAAGSAaMqSWagWb/FVCNGTx59+wNV7tl/l6pnQ1u3MpG9XODZ5xNXFir8992Cfv3Xx5Rf46MkLr365q7cM++gK7QqxkHgb+ELRR66mQx2b2BUrvg1npbp2oimdvm38ejKTiq6OkYZK1U9vxRQjFLHV6+IxPniN26orxe0Yoyn+/VSY8a0QK1X/e52Zie1D4zt284rHulbW8/sf2eOMOAAAAJABBuIAAABABhiIAwAAABkgI94h81p+pbn8XqHos3qFUvM6L99W6uDz97u6WDnT9LX7evzzRwcvcPVw/zb/ejnf7hCQ5uckj40ddPXH7/9vri6X/H5YSQVXF0Mbz8a6XPY55N7eXHsr22GbKfMZM+Anph93dcyQtxIz4L0aCXOQET9fLiOuvkWnSVJVvn1hJflrk+J7LKUU6iWv5oqYKvprSqYKc59702HaQM63Tu3Pj4TpcR/0Wh3rNlP7Uqw+zogDAAAAGWAgDgAAAGSAgTgAAACQATLi5ynelnn+bZr9bZmLDZmzpKqbVq6G7N5M6Ntb8fNXK/7XVCj7TLlCRnwkv8fVfd0+W9jb0ysgapWTvOfeu/38U2ddHffLfN9ImO5znc8cOXDu+91bL3PTss6Ib2Tlsj/exCxuzN/OVH3Wv5WukF2OWeaV7CO+3vs/58M9HPJ9c9fv9PcNu2mThaPh2f5zIyVfy3xttvi9L7IwVfTXPp0484yrP/fQ+859Px2uu9o5eqmr77jeZ7hjRrzdY10UM+Ot7jMANMMZcQAAACADDMQBAACADDAQBwAAADJARvw8xV7b1132YlcfOf2wq88U5vJ3MSNerc64emraZzLPTkz7ZU0e8ysTon0x/7ZlaJer9198nat7yIjjPMQcZNTd7Q8f1+56raufPnWfq4ul6UXrYsn3OB6ktf2KKYWM+LPPP910evv8+Z2YRO5kMnmj9X/O5fKuvmLfDee+f/rkl900K/jtHPuIJ/OfM1G1WmlaV0Ld3dVetj9mvpv1BV9o/s8+9FeuPjl+5Nz3hYLPiMfrDqaK/vqV7S3WtdWxLor7EbAcnBEHAAAAMsBAHAAAAMgAA3EAAAAgA2TEz1PMiPeHvuEDuVFXN2bWknxWr1zxfXlPjPl+sMdOHnH1xPRpV8fs3kDfVldff6XPrw8P+XXr7fV9xYGFxBxkzFHecMnrXH3k+CFXHzv7NVcXS6F/fqV87vuU0pLXE+2JfcRb1e0Lv8sO9qtuu9d9mD/Wd97+C67OPCMe+og3fs4MhHWz8ZDZTv46i3LV16UZ//6bLvr69Fn/OdPf7z/zWmXEYwb8xJnDrv7cw+9zdcyIT7bIlE8XJ89932V+XYrFcP1J+NmKDc9dSKtj3Vq/tgDrG2fEAQAAgAwwEAcAAAAywEAcAAAAyAAZ8SUy83/DdHf53LU1dMuthszkTMVn986cPenrqedcXSifarou/X2+j3i+z+fXY+4QWEjMOcYc5J233+XqYtFf6zBT9vv5147698R06P071ZDbnA4ZzmLJLzvX5/srYxlCZNssZn872en7PFagDZMFn2P++y//satjZny+/S2mry2NWei+roEwzX98z4Ro/kzFZ/2Pjz3rZ6j699TA4KB//oy/tqm3xf0nYub7cw/5TPih4w81nT9ZvE7EQtX4met/9lLR/6wHnvgnV+8cvdjVoyP+XhutjnXRWsqEx+seWllL644azogDAAAAGWAgDgAAAGSAgTgAAACQATLiSxb/hvG1y5CHHsmlmdDztOzzsdOls37+5Kcn+T7ijT3LZx8BlqtVljDf6/fDraMn/PRunzkdS8ddfWZybv6vPf2gmzY65HvjkxHvHJt3rLKmdevlxTpkfef1iF+5nvGD+a2h9vvw19/0Q67uDsfOU2d8jno69LKuqurqgfywq/vD/STi/SXa1ZgR7+nO+WnzPr7D50zFf87EfH2XfN//PYW9Yf7Qg73o399Toe/3iXHfN/z4uF/+ZNEvL7XYzeLkvq6540lft78uaqB3m6t7uts7Xqy33HTj7yZeF0EP9PWHM+IAAABABhiIAwAAABlgIA4AAABkgIz4EllIsHVb7LHa+DeOz+6VZ3yP5JNnjri6OOMz4lUrh2XHjGW1RQ10Xne3z9f29/m8bD7U3aFf9XhDv/xC0fcYL5Z9r32/JCxHjIB3d3U3nd6ueA1LVb7H80TBXytwfPwpV1fks82NYm75you+ztUnx59x9Uuv/xeuHur3WeLu8BH4pQf/2tVHQs45Wehv3+/v0fCiK97g6oGcz6zHYHSrTX36zNw9Js5M+O1WLPt1qYY+3JWw3c8U/LaJnxPHz/iftavPL/+hpz/u6smSz89Phsx4zJDHT60u+c/M3pDrHsj57PJo/1zv70u33eqmDfdvd/VlF1/j6nzOX6+y3sT9vjEXfs+9dzd9bsyMt+qRjtXHGXEAAAAgAwzEAQAAgAwwEAcAAAAyQEZ8yWLWz/9N053mNu1MmDdmwMcrz7k69g1X8plLhSzgjPl8bcxoAquht8dnPPdsvd7VR0497urCzKlz35erfp+vVH1GfLXFTGYr66sXb6u+4csLiaeQPS7plKvHijOu/uxDf+7q/tzQousSr82JvbH7+4Zc/cCTH3P18MAOv24Fv98dOvaQq4+c9bnpqnxuOh8y4sfGn/bTc6G/ddV/5MbcdOzBPjExd2w/euKom1as+uN+V85fS9Td6+tK8us+XfL7+IEj/+Tqh5/1fcOnZ/zvcSr0WE9h3eNnYsyE53t8j/Vdg9e6eu9On/O+5rKbzn0f328DOd9XvLe3z9c98Rqu9S3mvpuJfcSx9nBGHAAAAMgAA3EAAAAgAwzEAQAAgAysSEbczL5X0p/Wyx9JKf3RAvN8i6S7JN0kqVvSw5J+L6X0JyuxTisunX/OshyyejPJ9xHvSr6vb1f4NaXkM5jTBZ/VOz3xvKuH+n0uciDv83S5Xp/tBZait7s/1H6/6unxuc3p8lxG9cjZr7hpU8XbXF0shf7NfZ3dZ5v16ZXmZzJj7nLXlv2uXsuZ8XJlytWlGV/PVBbv472QmHOO16jEeqLo+2HHvuJOOIxa8ueOks1/9Ub9JX89zkfv+6+u7grnoqbK/lg6o9ir2x97Jwp++WeLvlby1//Mz7w3V5yey3mfPeuvm+ju9us+lMuFZ8driXxf8amKz9cXJ/26F5PPhFfkXz+ue3z/93X53t39Xb7X90DvTldffuGLXX3dZbe4esvo3OdYX8iAbzaNx5+NdGzarDp+RtzMLpb0TkkTTeb5SUkfknSDpD+X9IeS9kh6j5k1704PAAAAbAAdHYhb7fL7d0s6KekPFplnn6S7JZ2SdGtK6SdSSj8n6QWSnpD0C2Z2eyfXCwAAAFhrOn1G/KclvUbSD0iaXGSeH5SUk/SfU0pPzT6YUjot6dfq5Y92eL0AAACANaVjGXEzu1bSr0v6nZTSp83sNYvMOvv4RxaY9uEwzzoSE3Ndi06ryvfRtdB3t8v8r6U3Dfv5w/Tpgs90fvahv3L1VRf6frAvvOJlrt4y7LN6G63nKtaG+A6ZabhWYqp4xk372qEHXb1j5BJXLzcj3ioTfs+9zRNyMZd55+13Lev1m+l0prMcMuDPjT8QpvvMePy9xVR2VJ3XT7rF85sEpec9N2S0o7jsydDrOtYttWypHmZotXHCDK1mt+655VtX6Mtt8Tg9L1AfXsv3FZ+Rz4yX4/0qwvJ6LF4D4uuB3q2u3rf9Ja4eDD3Vh/r9/FtH/efQ8JDf7zdzLjweAxpz362OPWTC176ODMStNjL8M0mHJP2bFrNfXf/6eJyQUnrOzCYl7TWzgZTSVJwnvO59i0y6ZpHHAQAAgDWhU2fE/61q3U9ellJqdcn97O20ziwy/Yykwfp8TQfiAAAAwHq17IG4mb1EtbPg/yml9Pnlr9K5/w9r/Z98Kd2y0OP1M+U3d2BdAAAAgBWxrIF4QyTlcUm/fJ5POyNph2pnvE8uMH22yXWbYb7V1Z8LPVJzQ64eyM316i4VYx7UZ/265fu/juQvdPVQ7x5Xl2d8Z8ixwmFXnzrj+4g/mb7k6uG8z+LdeOWoq8mIYynifpPL+QxpT5P9KvYJn1eXfT0sv88uV8x8txJ79bbSTp/yTvcBLs/4bHDM408VfB0z4ss1r9N3ap7zdmxz33OusVf40LC/LqI07bdjChs6bueK+feQtQjAdyX/fs2bvx/F7iGfAN2z/SpXbxvxn2OX7rnS1X19PvPdGzLgfA4tjtz3xrLco9yQpKskXSupYGZp9p+kf1ef5w/rj/12vX6s/vWqsCyZ2YWqxVKeaZUPBwAAANaz5UZTipL+eJFpN6uWG/+saoPv2djKJyS9VNI3NTw26/UN8wAAAAAb1rIG4vULM394oWlm9g7VBuJ/Em5x/25J/4+knzSzd8/2EjezrZrruLLgzYAAAACAjaJjfcTPV0rpoJn9oqTflfQlM3ufpJKkt0jaq85d9Lmi8rkBV193+a2uPjL21XPfjxVjAsiH+Xrl8+XbBnwm9AVXvsLVx0486+qvPef7v45N+cz4xNQpV5+d9n3Fpws+cz6Q9/l34Hz09PpM545tu1zd2xP7AC+eUS2Xi6EuNa1jvrRdMZcdM+Pt5raX06d8uT3Ko3LZZ8SfP3bETw8Z8k6LWeVq1R+vSjO+0VZfz9y1BaF1tmzFM+PWpGr/+fO17EHgdDf0EU9dYd3Ctpip+N9j/HCfnwn3dZ/5z6F8T+j73esz3ztH/Hvgmv2+d8L2rfH9748Py33PAhvFqg/EJSml9E4ze0rSXZK+T7Ws+lclvT2l9CdZrBMAAACwmlZsIJ5SeoekdzSZ/iFJH1qp1wcAAADWss3dGwoAAADISCbRlI0gZsTzoY94Y1/xmM1Loc6FZW0dvsDVe3b4LN62ET/95LjPfJ6d8hnwQuW0q586+k+uvnzv1a4e6Pfr058bFtBKT7c/nPR2+wxoX7fvK24N5wGqacZNmwrXLTx95ICrh4dC7/s286Yx0x0z361y2e328W2nT3m7PcpbmQkZ8JgJL8/4bT8/59xerrlVJrxQOtv0+Y2Z8XzfUJM5W+vp8vdo6DHfi7vb/PRWOerIWp7Lar7tWm3ZroYceClsR/X4vuAVizerbt6vPa57X7ff1ntGXuDqC7Ze5uptW/3n0NYtvs/4QD/XGgHngzPiAAAAQAYYiAMAAAAZYCAOAAAAZICMeIfEvF2Xel3l+UxmrtfnFvdddI2rhwZ8HjbX57O2MVN+4qyf/3ToaXx68mlXf/af3+/q19zyNldvH+12dczHAwuJ+ds+G1l0eqXq+4KfmnjG1WcmLnF1seTzsUMDftntajfz3a52+pS36lHeaTGnHK9hmZ9j9o/Mz1X740W5POXqmNPuMt9fejC/fW5al/+Iir2z5ye4/SM9Xf61tuWvCOvi99H5WmXGw9ZJYfq8Mm5b//yUfN143YUN+m3xbPVxV09U/edK0mTTde3r8cfx7UP7XH3xBf5zaPvoXlfv3ePfk7k+v60BnB/OiAMAAAAZYCAOAAAAZICBOAAAAJABMuIdErOL3V1zfY27wt871ZjVCz2Q8yEDHjPhXd0+g3n9FS9x9amJw66emHne16VTrj4x7uf/x698xNUvu/nbXN3d7TOdvT2+BiT/HpCkwdwuP93mppervm94ccZf1zBT8Znw2Kt6rVlOn/KVzoRHrTPhrfjjUU/yueuh3tAL3PzvbtfQDa7u75v7+a3LHzt7e/2ycz2hV31P6BseMuKxjtcxRPMy6Ba2VWq+tcya9yGP4n7deKwtl33/94lp/x4pTvvjfNmm5YXPndA3fN8FN7n68kuud/XWUd8nPB73e/gcAJaEM+IAAABABhiIAwAAABlgIA4AAABkgIx4h8zrI24NmzbkxxVyhclCX94WscKYId864rO3l170Alcfm3rU1YWJcVdPFH1m/PhZ32f8yUP++S+4equryYhjIb3duVD7/bavoS5Vz7ppM/L51kLJ96IuFn1mvFQq+mX3teoPvbpWO/fdWS1y0DEjbsOuzvX4euug70d90darwhLnjp3VvtBjfMAva98F+/z03GBcuRZa9AVvvYAO86/fZXPb9vSZ427aqQl/nD5Z9Hn5cot17+/2++RQv7//xOjwNj9/fv3cP6I84/P0lcrMInPWxHx7Y/92YKVxRhwAAADIAANxAAAAIAMMxAEAAIAMEIRaIdaQz+sKGcqYOyzNTIfa513LZV93d/usXj7n6327febywDM+kzk17TPhpdCjeaJw0tWTxdOuni76ns8D/SGXCUjq6vKHl+GBcG1B11xGvCv5eavJZzrHp064+snDj7h6ZNgve61lxNcym9cLu71O4haOZ/F4N9y/3dVX7HmRq7cMXuDqSnXud1/p8/tFzIjv2LrbT19HOeZ29YT7R4wM+Uz34NkRVxemjzVfoK3tXvzNzIQMeMyEj0/4z6znjh5ydT7vP7P2XrjP1UMDflsCK4kz4gAAAEAGGIgDAAAAGWAgDgAAAGSAjPiKmfsbx1LMiHuxR/KhI19z9Z7tPvOdDznI2PN0ZMD3f71m7x2uPjN5xNWnpnx+brrqM+IHj93n6otC797ukF2MmfV838bNbWJxcb/sz/lMa3/fXK57uux721eqPvM5NvWcq6cKvu947CMuLls4p1yZCvV007pVZ+2WwgLyobf33j37Xb1ry8Wubkwup/Dqvb0++9+3ie5hUFXF1RddeImrHzri3xOtfnNV+fdYSiEz3vYvfmU15sAnJ/3x4tCRJ1194vSzrj588quu3ja8x9Xbt+1wNRlxrCbOiAMAAAAZYCAOAAAAZICBOAAAAJABMuIrpLG3rrXoI14IfblPT/o8bKHkp0vb1cxA3vfa3X/Rda5+4nmfySxUxlw9WfT5u7GpZ1z9jw991C9/9w2ufuHVL3M1GfHNycz/nd+f97nLi7fecu77ctnnVc+WfMazmvz0meQz4TFLjDmlkAE/cuYrri5X/fRO6+4J15Dk+109PLxlRV9/o0ih3/vXDn/Z1TFDnhT7hPvPnXK4X8V0adLV8dqlfD7v6u7uzg4f4jGgPFNy9dT03PodPPSYm3b42KOufn7sgH9uyd87oyesennG30sDWE2cEQcAAAAywEAcAAAAyADRlA7pDy268n2Di04rFv1/k1XCf7uXKr4NVWzl1kr8L8OhQX/771uveYOrx7501NVTISozWfK3C+6u+NhAvnfI1fsnrnd1f0NUpq+XW49vFj09sX2hj6aMNtzafM9WH5968oS/pX256vfJ+J6oJv/f8ptZvN33dIiaTRX98WXef8vH/oWKk7tb1OH8jsXYEDGipSgUQxvK0oyri2H6fH67lyr+937kuI9z7N5ymatjS792oykxahKjKI3RE0k6dMSvz/jk3DHhzOQxN+34uG9feKrkWwBXq35bVbXX1YnjBzLEGXEAAAAgAwzEAQAAgAwwEAcAAAAyQEa8Q+Jt3a+7/NZz3x8Ze9hNGyv6v39SaDsV82zz8rBV35aqq6v531MDIZu7ZfBCV28f8recHp/yt7gvVM64erLq2yuOT/l2ig8/cZ+rG29xPTrs5yUzvnHFDOn20W2uPjs+t1/1T/mWmz3dfa4uVXzO+cyUv64h5mdLZd+abTPtZzF7+/yxI66eKbd3zUkUM+A9yf/uzHxmvCvF0HmLEDoW1BXagXaZf391qdfVFrZzbPEZ21pOh2sHJqf9e26m4j+X4jsqTo/7WcyAP3XE57jPjB939fi0rw+d/tK57wtlv66lcA3JTPI/W3/O76P5XC5M99c5AauJM+IAAABABhiIAwAAABlgIA4AAABkgIx4h8SMeGMmbSA36qbFjGVSyIQnXxeKPu8Wb3Xcrr4ev657d/hb1B859birp0NGfKbq+8+emPRZv9ivdvDJuZ//pmtf6tdlE2V3N5ue7nBr85y/tflAw63Oc71+n+wx33s/pTFXTxd8fvXg4a+6euvITldvpv1sZibeKtzXM2V/fJmfJW6uVd/wOF3zajLiSxIy4j1d4Zbz8z7O43YOv9nkrzUqV8PnjPz0Viozfr868vwhVz934hlXnz7rr104dPLLrp4s+etAimkuF16q+Lx5zL/H66K2j17k6ttuuNPV/Xky4rMmC2OtZ2owmN+yQmuyfOvlZ+GMOAAAAJABBuIAAABABhiIAwAAABkgI75i5vJ5sd9rzGRWQ76tkmKfXz99uRnx/rzP3w71b3X1aL/P000Vfc6qJJ/PLVV9PT51zNVnJub6wU5M+Xn78z4bHHtPY+Po7fG/2z279577PvauH+33ve4nS76n8NmC38cmQ2a8VPLXMWjQX6eBOTFf27Z5kW/6hq+E+LnRFfq1x/7tlppn/6vh2qR4rVKrjHi5XHL1xJTv7X3itM94H3zun119cvKgq6dm/Hu6UD0V1mfufhv9IQMer8PaMXqxq19+/VtdHTPjsc/4ZhJz1MfG/O/lC4980NW3XftmV+/a4u9DkmVmfL3+LJwRBwAAADLAQBwAAADIAANxAAAAIAMEcleINfR8tdD/NfaD1bzMd8jmhYil2fIylzGHfdEF+1x9+PnLXT1VPO3qU4Wiq6vJ1+Wq7/F6amKuX2zMDW4d9f2eyYhvXD09va7ub2iDfOlFV7hppyd8z+GjE4+4ulLxPY+LM1Ounij4fTBf8NP7wrr0hnojS9Zef+h2xWteUssMOZbGQhX34XierRLqcG2Swnuq7OvJKf+emu7y76kDTz/s6meP+/tLnJ72fcXPlJ/2r5/86/XnhkI9l+PeMXqJm3bLFa93dZwe+4rHTPlm0ipHfc+9dzd9fsxZ33n7XZ1ZsSVq/HnW68/CGXEAAAAgAwzEAQAAgAwwEAcAAAAyQCB3hTT2fO22XJgW//7xWb1qNfRzTT7bV6n4urvb949tpSfksIcHfa/MG6+63dWThTOuPlv0PZ2LqRRqn0E7W2joIz7t8+aFos8d5vr8tsLG1ZgZ78/53vYDeZ/p7O32+8VUxe9jU0XfR/yJI4+5+mzR51n3777ML38TZcRbaber+Lz7GhABXxH5vn5X9/f5DPVAzh/Hp6Z9X27J358ihcx4peqnx2Pz4we/4pcWrtM4PXnE1QdPf96vT9n3Bc/l/Hs63+fXv1kOnAx458ScdMyQTxb8Z3bsvR2nZ9lHPP4srcSfJSucEQcAAAAywEAcAAAAyAADcQAAACADZMRXjDV850OTXS02e6Hs86wnx064es8On83r6+tbygqekwvZw60jvrf3zq0+jzc+7TPix6YecnXMDp4tz/WEfvK5L7tpe3fvc3U+NxDWLS9sPt3m9+le83nYlHwu8di471ncbT5z3tiDWJJK5YKrB/N+v9tMUrxvQZsp8Xhfg4F8zC7HftD+d4Pzk8v5Y+GV+2509bOnfB/vk9PtnWcrVXwm/Ojpp/zr9/jPoefP+uP+ZNln0ktV/zlWrvrPhdG+ra5++Q3f5epLdt3g6sYcOBnwpYuZ7v27b3b1gwc/4erBvP89HXz+flffuP81HVy75YmZ75gZj9N3bdnv6qzy7ZwRBwAAADLAQBwAAADIAANxAAAAIANkxFdIYw4y3yIjWQp9uotln6V7/tRBV19V8pmu4cHO5uViLvuafTe5erow4erJ4klXj1WecHVxZu7nOzr+qJv24IEvuPr2F3xj03XBxtTT4w9Ffb3+9z7Sf6GrT08/7eqpGX/dwlTB9yyeGPfTq6EX/0YSr9EoVwpNpy+373e8BiZmwm+84g5X94fpOD/xWp74ORI/Z7os3F8ixdI/UJrxGfHnzvgMeEV+PypV/fwxAx5fMOa6Yy/wmAnfMXKx0HmtMt8xNx2tld7bsxpz3XHd77z9rvN+bpY4Iw4AAABkgIE4AAAAkAEG4gAAAEAGyIivkMZ+2Ldc92o3rVD1/VhHi9tcXSrOuLq3O+fq8ozP6nVazGWPDvv127llr6uPnvF9x6crz7m6UBk/9/3Zou81e+TUV1195uytrh7I+1xhX+/yeqZjzmRhrK35VzJP19PT6+qLLvBZv6eefzDM798TxZmzrj499aSrt07tdvV0wedbiyXfZ3w9X5sQM+DPjfmsb6lSXNHXHwzv2ZgJp494h5g/j9bb7TPkFj7ebd55t5ARDxnvUvXZpi8f+8/HDHhj329pfib85de/ten8WB3rtff2QtbSurSDM+IAAABABhiIAwAAABlgIA4AAABkoKMZcTN7uaSflXSHpG2STkl6UNJvp5T+Nsx7h6S3S7pNUl7SAUnvkvTOlNK6b/LbmBHfNuLzqV9/89tcPVUMfcRLPgP+7DGfd21c9mqIuezLL73W1TPJZwv7jvns4OmpQw2V/9sv1+ezvmfGT7v6Qh8/xzLETPixMd+fPstsYG/IiA8N+rzoZRfd6OpiCn3DZ3zf8FTxh7ZKpeTqJw494uqRIZ9v3UgZ8VZ1+33EU4vpFqplNirHgrpCRrzL/D7fY/64Xan6Y23F/HsiqfnHbrsZ8FuueH3T6fH5cflYGfG4vV57b28kHRuIm9nbJf17SSck/W9Jz0naIekmSa+S9LcN875R0gclFSS9T7UB+7dK+i1JL5X0HZ1aLwAAAGAt6shA3My+Q7VB+N9JelNK6WyY3tvw/YikP5RUkfSqlNKX6o//sqRPSHqLmb01pfTeTqwbAAAAsBYtOyNuZl2SfkPSlKT/Kw7CJSmlVG4o3yJpp6T3zg7C6/MUVIuqSNKPLXe9AAAAgLWsE2fE75C0X9IHJJ02s2+WdINqsZMvppQ+H+Z/Tf3rRxZY1qdVG9DfYWa5lNLKNrxdJfOzcLFfqu/LPV2c8FN3Xr0Sq3XeekNGfGRoq6uvv+LFrr7s4qtcPTb1/LnvHz18r5t2E/PLlgAAE8xJREFU46WvdfWO0YuXvJ7wWmXC77n37qbPj5nxVtnBTsr3+Z7Il19ynasv3OXfM6XKlKvLZd+L/+Chr7m6t2cT3ULBfKbbQsa7deK7u6265QLRIT57HzPi3clf52CaCrV/j8SMeMxsx2Nz7ANOBnx9IvOdvU58Gs2Owo5Kul+Su6rKzD4t6S0pnbu6anZU+XhcUEppxswOSrpe0mWSHonzhGXft8ika85v1QEAAIBsdKJ94a761x+V1C/ptZKGVTsr/lFJr5D0/ob5Z/8s9q1C5sw+zp9pAAAA2LA6cUZ89v8lTbUz3/9crx82s29X7cz3K83s9gViKguZ/f+2lv/BmVK6ZcEF1M6U33werwUAAABkohMD8dnGz082DMIlSSmlaTP7qKQfkvQSSZ/X3BnvxQJjs8Gyxc6Yb3j9uaGsV6Gp2PN5dGh703qkob5k1/VNl02OcOXEzHcrsY/4auoJ+9hIT9inBn0dFUu+V/aOkebXHqznvuErzcJ/nPYkf3wyi/+xSkh8NcTrKLYNX+DqPTuvcHW838OMJl1dteYZ8fmZcP+e4tgNLE0noimP1b+OLTJ9dqA+e9SYnf+qOKOZ9ah24eeMpCfjdAAAAGCj6MRA/NOqDZyvNAu38qq5of71qfrXT9S/ftMC875C0oCkezdKxxQAAABgIcseiKeUTqh2d8xRSf+2cZqZfYOkb1QtZjLbrvADqt19861mdmvDvHlJv1ovf3+56wUAAACsZZ1qpvvzkr5O0i+Z2SskfVHSpZK+XbU7aP5ISmlMklJK42b2I6oNyD9lZu9V7Rb3d6rW2vADqg3ssUGQHVwbYuY7Zsbj9F1b9rt6PfWbzYX8bKw3F2tRtxLP14Q6tVheIjO+EnI5f13DjVff5uoriv56nEoqhSXEfvJVVw/k/XGbvuDAyuhENEUppWOqDcR/S9LFkn5atRv3/I2kl6eU3h/m/2tJr1Qt1vJmST8lqazagP6tKXHkBgAAwMbWsdvLpZROqTaQ/vnznP9zkt7QqdcHAAAA1pOOnBEHAAAA0J6OnREHsLbETHfMfN95+11tPR8bg7WdEY9icjDWVWHltboOYuu5m14DWMs4Iw4AAABkgIE4AAAAkAEG4gAAAEAGyIgDmwSZ783J5vX5XuHusDSfBYDzxhlxAAAAIAMMxAEAAIAMMBAHAAAAMkBGHAA2MvOh7apmXJ1ahLpbdR2f15fc4jOW27ccADYuzogDAAAAGWAgDgAAAGSAgTgAAACQATLiALCBxQz4vLbibTf+9vPn8wOu7s+N+Do/2ObyAWDz4Iw4AAAAkAEG4gAAAEAGGIgDAAAAGSAjDgBowmfCY9/w/tywq2+44jZX53NDK7NaALABcEYcAAAAyAADcQAAACADDMQBAACADJARBwAs2WB+i6tjZrw/Rx9xAFgMZ8QBAACADDAQBwAAADLAQBwAAADIABlxANhABnIjrt4+erGrq5pxdbJq0+V1Jf8x0adRV28budDV/fnN2zd8sjDW1vwxXw9g8+GMOAAAAJABBuIAAABABoimAMAG0p/30ZSXXv9mV08Vx8MzkpqxcL7G1O3qgZyPqvRvolvaxyjKsbGDrv7CIx909W3X+t/Fri37XU1UBdh8OCMOAAAAZICBOAAAAJABBuIAAABABsiIA8AGEtsXxhpL1yoTfs+9dzd9fsyM33n7XZ1ZMQDrFmfEAQAAgAwwEAcAAAAywEAcAAAAyAAZcQAAliBmvluJfcQBgDPiAAAAQAYYiAMAAAAZYCAOAAAAZICMOAAASxAz3zEzHqfv2rLf1YP5LSuzYmhL7A/fCr83dBJnxAEAAIAMMBAHAAAAMsBAHAAAAMjApsmIkwEDACxH/FyIme87b7+rredvJOvpMzau67Gxg64m64/VxBlxAAAAIAMMxAEAAIAMMBAHAAAAMrBpMuJkwAAAnbSZPxfWU8661brec+/dTZ8ff5ZW1wIA7eCMOAAAAJABBuIAAABABhiIAwAAABnYNBlxMmAAACzNRspZx3VpJebbgU7ijDgAAACQAQbiAAAAQAYYiAMAAAAZ2LAZ8cPHv6qf+b3rztX7d9/UdH4yYAAAnJ/1nLOO67KWep5j8+GMOAAAAJABBuIAAABABhiIAwAAABnYsBnxi3dep9/58XvP1X/5ybe76WTAAABYmvWUs46vFdelVU9zxgNYSZwRBwAAADLAQBwAAADIAANxAAAAIAMbNiMefderfzXrVQAAYF2KOel4b45W9+pYS8h8Yy3hjDgAAACQAQbiAAAAQAYspZT1OnScmZ3s7+/fdu3VV2e9KgAAANjAHnnsMU1PT59KKW1v97kbdSB+UNKIpHz9oUczXJ316pr6V7Zd+9h2S8N2Wzq23dKx7ZaObbc0bLelW6vbbp+k8ZTS/lYzRhtyID7LzO6TpJTSLVmvy3rDtls6tt3SsN2Wjm23dGy7pWPbLQ3bbek24rYjIw4AAABkgIE4AAAAkAEG4gAAAEAGGIgDAAAAGWAgDgAAAGRgQ3dNAQAAANYqzogDAAAAGWAgDgAAAGSAgTgAAACQAQbiAAAAQAYYiAMAAAAZYCAOAAAAZICBOAAAAJCBDTkQN7O9ZvYuMztiZkUze8rMftvMtma9blkys+1m9sNm9j/N7ICZTZvZGTP7rJn9kJktuD+Y2R1m9rdmdsrMpszsK2b2s2bWvdo/w1pjZt9rZqn+74cXmedbzOxT9W09YWb/aGZvW+11XQvM7OVm9kEze67+3nzOzD5mZm9YYF72uzoz++b6dnqm/r590szeb2a3LzL/ptl2ZvYWM3unmX3GzMbr78U/b/GctrfPRnwft7PtzOxKM/tXZvYJMztsZiUzO2pm95jZq1u8ztvM7Iv17Xamvh2/ZWV+qpW3lH0uPP+PGz43rlhknu76PvmV+nv+VH2fvaNzP8nqW+L71er70Kfq22HazA6a2V+Z2VWLPGf97HMppQ31T9Llko5KSpL+WtKvS/pEvX5U0vas1zHDbfOj9e1wRNJ/l/QfJL1L0lj98Q+ofpOnhue8UdKMpAlJfyzpP9a3Y5L0/qx/poy358X1bXe2vj1+eIF5frI+7YSk/yLptyQdrj92d9Y/wypvr7fXf+7jkt4t6dck/TdJ/yTp/w/zst/NbYvfaNiH/qh+TPuApJKkqqTv2czbTtID9Z/trKRH6t//eZP5294+G/V93M62k/Te+vSHJf3X+ufH/6hvyyTppxd53t316Yfr2+2/SDpZf+wns94Gq7HPhed+a8Nzk6QrFpjHJL1fc+OW/1jfVyfq2/uNWW+D1dp2kvKSPtSwLf5zfd/7E0lPSvqW9b7PZb4CK/BL/mh9Y/9UePw364//QdbrmOG2eU39INAVHt8t6VB9+7y54fERScckFSXd2vB4XtK99fnfmvXPldG2NEl/J+mJ+kFy3kBc0j5JhfoBYF/D41slHag/5/asf5ZV2l7fUf95Py5peIHpvQ3fs9/N/cy7JVUkPS9pV5j26vq2eHIzb7v6driy/p58VbMP9qVsn438Pm5z232/pJsWePyVqv1RWJR0YZh2R32ZByRtDdv0ZH277uvUz7MWt1t43s76e/m9kj6lxQfi31Wf9jlJ+YbHX1zfzscWOo6uh3/tbjvVBtFJtRM3XQtM7w31utvnNlQ0xcwuk/Q6/Z/2zjbWjqoKw88KtQ2UtgSFaKghBSqp1B/EWEnBAlFQEjWtNqDEWjSpooRaixG/0MZAokYiVeInSYuAipAAohVIaG+rqQoNxoAk9mJ7DZjWCi1Sblu+uvyx9vQO0/Nxz+H2zDkz75Oc7Jy998yZec+a2Wv2rL03jBB/Xp5vAKPAEjOb2uND6wvcfb273+vuBwv5O4Efp6/n5YoWEzeOX7n7llz9A0TvJsBnjtwR9zXLiQebTxB21YhPAlOAG919JMt09z3ETQXiLUWlSSFP3wb2AZe6+95iHXd/KfdVdjfGyUQI4V/cfVe+wN03EL1KJ+Sya6edu29w92FPrW0butGnstdxJ9q5+1p3/2uD/I2EUzmZcILyZLpcl/TKthkh2ugpxD10oOjQ5vL8NKVXtKmX2eDXkm1mv/swcDthw4s7/O2+oBPtzOxUwoYeBr5a9F3S/l4qZA2czVXKESccI4AHGjibe4mny2OAs3p9YANAZswv5/IyPe9rUH8T4VjNN7MpR/LA+g0zm0OEB6x2900tqrbS7/eFOlVmPjALWAfsSfHOV5vZ55rEOMvuxhgmehvnmdkb8gVmtgCYRryZyZB2relGH13H7WnUfoC0O4SZXQYsBC5392da1JtC3DP3AX9oUKVOun2U8FNvBqab2cfM7Mtm9qlmsfUMoM1NKvsAJpjTU7q1Sfkw0WP+FuDBnhzRAGBmk4CPp695422qp7u/bGbbgTOAU4hYr8qTtLqFCOX5SpvqrfTbYWajwEwzO8bd903skfYV70jpf4BHgLflC81sE7DY3f+bsmR3CXffbWZXE6F1j5vZ3cTr1VOBDxKhPp/ObSLtWtONPrqOW2BmJwPvJhzHTbn8qcBJwPPuvqPBpsMpbTjYrkokjVYTIRh3t6l+GnAUEXJWfLCBGunGWNsxgwgDfX2uzM3sR8TYhFdgcG2uaj3iM1L6vyblWf5xPTiWQeJbwFxgnbvfn8uXnofzdeBM4DJ339+m7nj1m9GkvCqcmNLLgaOB9xA9uXOJMR0LiIFJGbK7HO5+A/AhouNkGfAlIub+SWBtIWRF2rWmG310HTch9d7eRrzuX5UPBUC2CBwKzbuZGGi5fBybSLcxsrbjm8AWohNnGvHg90/gs8A1ufoDqV3VHPF2WEo7jeuqLGa2HLiKGI28pNPNU1oLPc1sHtELfr27/2kidpnSquuXTQlnRM/3g+7+vLv/HVgEPAWc2yRMpRF10Q0AM/siMUvKWqInfCrwdmLGgNvM7Dud7C6ltdCuC7rRp5aapqkebwHOJuKWv9vlrqqu2+eJAa3LCg8q3VIne8vajh3AInd/LLUd64kY+YPASjOb3OF++0q7qjni7Xomphfq1Rozu4J4XfY4cL677y5UkZ6JXEjKVl79BN6K8er33Gs4tEEga3y2ufvf8gXprUL2FmZeSmV3CTM7jxjo+ht3X+nu29x9n7s/QjzE/Bu4Kg1UB2nXjm700XVcIDnhtxJvZn5NTKFZdG7a6dau93LgMbPZwHXAGndfN87NdA2PkbUd9xXfQKe2ZDvRQz4nZQ+kzVXNEf9HSpvF/8xOabMY8tpgZiuI+TgfI5zwnQ2qNdUzOaaziME5247UcfYRxxI6zAEO5BZjcGJGHoCfpbwb0vdW+r2J6Nl8qgZxpZkOzzYpz262Rxfqy+4gW4BiQ7Eg2c1DxH38zJQt7VrTjT66jnMknX4JfAT4BTET0mGxzO4+SjwoHpt0KlKH9vgM0iwd+TYjtRvnpjrDKW9h+v4EMWXpKUnrInXQLaOjtmNQba5qjnjWWF1ohVUizWwa8QptP/DnXh9YP5EGf32PmFj//OK0aDnWp/R9DcoWEDPQbHb3Fyb+KPuOF4gFFRp9sim9/pi+Z2ErrfS7qFCnymwinJvZTV4hzk3pSEpld2Nks3ec0KQ8y38xpdKuNd3oo+s4ka7fO4me8J8DS7KBck2ou3YjNG83ss6vO9L3EYBke5sJW3xXg33WQbeMbFKNucWCND4hc6xHckWDZ3MTNSF5v3zQgj7t9Lkm6bAFOL5N3enEKoi1WRykS01X0XhBn1lUdCGQLjS6NZ3vtYX8C4g4v2eB41Ke7G7snC9O57sTOKlQdlHSbj9pxeC6a8f4FvTpSJ+6XMfj0G4K8LtU5yYaLK7SYJuBW1xlonVrsd0Qr21Bn+lln3sPbG4yMSjzIHBBoezatO3QoNucpQOsDGkC+M3EaNt7iCmo3kms5rQVmO8t5vCsMma2lBjw9QrwAxrHSY24+9rcNguJHpADxGpgu4lp005P+Rd71YyoQ8xsFRGesszdbyqUXQl8n7gB3E70XC4GZhKDPr/Q26MtBzM7kWhUTiPmxn2IWKxmEXHTvNTd78jVl91xaMaF+4mZZvYCdxFO+RwibMWAFe6+OrdNrbRL55u91n8j8F4itCSbg/np/HXWjT5VvY470c7M1hCraz4N/JDGA96G3H2o8BvXAyuJQdl3Es7VJcRUdFe6+40Td0a9oVOba7KPISI8Zba7P1EoMyL2fjExkcK9hF6XEA+NH3b3eybkZHpMF9frOcADhN3cBfyLeCBZQDxUn+Purwo1GTibK/tJ4Ag9Zb0ZWEOMtH2R+ONW06YHuOofxnpuW32GGmx3NmkxFqL37VFiJPhRZZ9TP3xo0iOeK/8AsJFwpEaJVcKWln3cJeh0PPFmanu6Lp8hHpbPalJfdhc6vA5YQYTUPUeE+ewCfgtcWHftxnFfG5kIfap4HXeiHWM9uK0+q5r8ztKk12jSbyPw/rLPv5c212AfmZ6H9Yin8knJJh9NNron2ez8ss+/19oBbyUegHeltuNJ4CfAzBa/MzA2V7kecSGEEEIIIQaBqg3WFEIIIYQQYiCQIy6EEEIIIUQJyBEXQgghhBCiBOSICyGEEEIIUQJyxIUQQgghhCgBOeJCCCGEEEKUgBxxIYQQQgghSkCOuBBCCCGEECUgR1wIIYQQQogSkCMuhBBCCCFECcgRF0IIIYQQogTkiAshhBBCCFECcsSFEEIIIYQoATniQgghhBBClIAccSGEEEIIIUpAjrgQQgghhBAlIEdcCCGEEEKIEvg/6HEWi5Z2M44AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 217,
       "width": 369
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "characters2 = characters + ' '\n",
    "[X_test, y_test, _, _], _  = next(gen(1))\n",
    "y_pred = base_model.predict(X_test)\n",
    "##用ctc_decode得到解答，自己寫可以參考下方\n",
    "out = K.get_value(K.ctc_decode(y_pred, input_length=np.ones(y_pred.shape[0])*y_pred.shape[1], )[0][0])\n",
    "out = ''.join([characters[x] for x in out[0]])\n",
    "y_true = ''.join([characters[x] for x in y_test[0]])\n",
    "\n",
    "plt.imshow(X_test[0])\n",
    "plt.title('pred:' + str(out) + '\\ntrue: ' + str(y_true))\n",
    "\n",
    "argmax = np.argmax(y_pred, axis=2)[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 自己寫 decode CTC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "## 其中0代表預測為空格，如果預測相同字符之間沒有空格要移除\n",
    "word=''\n",
    "n=''\n",
    "for single_result in y_pred[0].argmax(1):\n",
    "    if single_result != 36:\n",
    "        if n != single_result:\n",
    "            word+=characters[single_result]\n",
    "    n=single_result "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 額外參考：加入RNN的神經網路"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:8: UserWarning: Update your `Conv2D` call to the Keras 2 API: `Conv2D(32, (3, 3), activation=\"relu\")`\n",
      "  \n",
      "D:\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:10: UserWarning: Update your `Conv2D` call to the Keras 2 API: `Conv2D(32, (3, 3), activation=\"relu\")`\n",
      "  # Remove the CWD from sys.path while we load stuff.\n",
      "D:\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:23: UserWarning: Update your `GRU` call to the Keras 2 API: `GRU(128, return_sequences=True, name=\"gru1\", kernel_initializer=\"he_normal\")`\n",
      "D:\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:25: UserWarning: Update your `GRU` call to the Keras 2 API: `GRU(128, return_sequences=True, go_backwards=True, name=\"gru1_b\", kernel_initializer=\"he_normal\")`\n",
      "D:\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:28: UserWarning: Update your `GRU` call to the Keras 2 API: `GRU(128, return_sequences=True, name=\"gru2\", kernel_initializer=\"he_normal\")`\n",
      "D:\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:30: UserWarning: Update your `GRU` call to the Keras 2 API: `GRU(128, return_sequences=True, go_backwards=True, name=\"gru2_b\", kernel_initializer=\"he_normal\")`\n",
      "D:\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:34: UserWarning: Update your `Dense` call to the Keras 2 API: `Dense(37, activation=\"softmax\", kernel_initializer=\"he_normal\")`\n",
      "D:\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:36: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"de...)`\n",
      "D:\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:44: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=[<tf.Tenso..., outputs=[<tf.Tenso...)`\n"
     ]
    }
   ],
   "source": [
    "from keras.models import *\n",
    "from keras.layers import *\n",
    "rnn_size = 128\n",
    "\n",
    "input_tensor = Input((height,width, 3))\n",
    "x = input_tensor\n",
    "for i in range(4):\n",
    "    x = Convolution2D(32, 3, 3, activation='relu')(x)\n",
    "    x = keras.layers.BatchNormalization(axis=-1)(x)\n",
    "    x = Convolution2D(32, 3, 3, activation='relu')(x)\n",
    "    x = keras.layers.BatchNormalization(axis=-1)(x)\n",
    "    if i <3:\n",
    "        x = MaxPooling2D(pool_size=(2, 2))(x)\n",
    "    else:\n",
    "        x = MaxPooling2D(pool_size=(2, 1))(x)\n",
    "\n",
    "\n",
    "conv_shape = x.get_shape()\n",
    "x = Reshape(target_shape=(int(conv_shape[2]), int(conv_shape[1]*conv_shape[3])))(x)\n",
    "\n",
    "x = Dense(32, activation='relu')(x)\n",
    "\n",
    "gru_1 = GRU(rnn_size, return_sequences=True, init='he_normal', name='gru1')(x)\n",
    "gru_1b = GRU(rnn_size, return_sequences=True, go_backwards=True, \n",
    "             init='he_normal', name='gru1_b')(x)\n",
    "gru1_merged = keras.layers.add([gru_1, gru_1b])\n",
    "\n",
    "gru_2 = GRU(rnn_size, return_sequences=True, init='he_normal', name='gru2')(gru1_merged)\n",
    "gru_2b = GRU(rnn_size, return_sequences=True, go_backwards=True, \n",
    "             init='he_normal', name='gru2_b')(gru1_merged)\n",
    "x = keras.layers.Concatenate()([gru_2, gru_2b])\n",
    "\n",
    "x = Dropout(0.25)(x)\n",
    "x = Dense(n_class, init='he_normal', activation='softmax')(x)\n",
    "\n",
    "base_model = Model(input=input_tensor, output=x)\n",
    "\n",
    "labels = Input(name='the_labels', shape=[n_len], dtype='float32')\n",
    "input_length = Input(name='input_length', shape=[1], dtype='int64')\n",
    "label_length = Input(name='label_length', shape=[1], dtype='int64')\n",
    "loss_out = Lambda(ctc_lambda_func, output_shape=(1,), \n",
    "                  name='ctc')([x, labels, input_length, label_length])\n",
    "\n",
    "model = Model(input=[input_tensor, labels, input_length, label_length], output=[loss_out])\n",
    "model.compile(loss={'ctc': lambda y_true, y_pred: y_pred}, optimizer='Adam')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_5\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_3 (InputLayer)            (None, 80, 170, 3)   0                                            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_5 (Conv2D)               (None, 78, 168, 32)  896         input_3[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_5 (BatchNor (None, 78, 168, 32)  128         conv2d_5[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_6 (Conv2D)               (None, 76, 166, 32)  9248        batch_normalization_5[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_6 (BatchNor (None, 76, 166, 32)  128         conv2d_6[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2D)  (None, 38, 83, 32)   0           batch_normalization_6[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_7 (Conv2D)               (None, 36, 81, 32)   9248        max_pooling2d_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_7 (BatchNor (None, 36, 81, 32)   128         conv2d_7[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_8 (Conv2D)               (None, 34, 79, 32)   9248        batch_normalization_7[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_8 (BatchNor (None, 34, 79, 32)   128         conv2d_8[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2D)  (None, 17, 39, 32)   0           batch_normalization_8[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_9 (Conv2D)               (None, 15, 37, 32)   9248        max_pooling2d_2[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_9 (BatchNor (None, 15, 37, 32)   128         conv2d_9[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_10 (Conv2D)              (None, 13, 35, 32)   9248        batch_normalization_9[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_10 (BatchNo (None, 13, 35, 32)   128         conv2d_10[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_3 (MaxPooling2D)  (None, 6, 17, 32)    0           batch_normalization_10[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_11 (Conv2D)              (None, 4, 15, 32)    9248        max_pooling2d_3[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_11 (BatchNo (None, 4, 15, 32)    128         conv2d_11[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_12 (Conv2D)              (None, 2, 13, 32)    9248        batch_normalization_11[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_12 (BatchNo (None, 2, 13, 32)    128         conv2d_12[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_4 (MaxPooling2D)  (None, 1, 13, 32)    0           batch_normalization_12[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "reshape_3 (Reshape)             (None, 13, 32)       0           max_pooling2d_4[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "dense_5 (Dense)                 (None, 13, 32)       1056        reshape_3[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "gru1 (GRU)                      (None, 13, 128)      61824       dense_5[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "gru1_b (GRU)                    (None, 13, 128)      61824       dense_5[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "add_1 (Add)                     (None, 13, 128)      0           gru1[0][0]                       \n",
      "                                                                 gru1_b[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "gru2 (GRU)                      (None, 13, 128)      98688       add_1[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "gru2_b (GRU)                    (None, 13, 128)      98688       add_1[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_1 (Concatenate)     (None, 13, 256)      0           gru2[0][0]                       \n",
      "                                                                 gru2_b[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "dropout_3 (Dropout)             (None, 13, 256)      0           concatenate_1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "dense_6 (Dense)                 (None, 13, 37)       9509        dropout_3[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "the_labels (InputLayer)         (None, 4)            0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_length (InputLayer)       (None, 1)            0                                            \n",
      "__________________________________________________________________________________________________\n",
      "label_length (InputLayer)       (None, 1)            0                                            \n",
      "__________________________________________________________________________________________________\n",
      "ctc (Lambda)                    (None, 1)            0           dense_6[0][0]                    \n",
      "                                                                 the_labels[0][0]                 \n",
      "                                                                 input_length[0][0]               \n",
      "                                                                 label_length[0][0]               \n",
      "==================================================================================================\n",
      "Total params: 398,245\n",
      "Trainable params: 397,733\n",
      "Non-trainable params: 512\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "Captcha.ipynb",
   "provenance": []
  },
  "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}