examples/translate_with_seq2seq.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "lY2BrIyFKbAP"
   },
   "source": [
    "# Neural machine translation with Seq2Seq\n",
    "## Install dependences"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "s7PWEKR7T2dn"
   },
   "outputs": [],
   "source": [
    "# install the latest kashgari\n",
    "!pip uninstall -y kashgari\n",
    "!pip install git+https://github.com/BrikerMan/Kashgari@v2-dev\n",
    "!pip install hanziconv\n",
    "!pip install segtok"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "wfodePkj3jEa"
   },
   "source": [
    "## Download and prepare the dataset\n",
    "\n",
    "We'll use a language dataset provided by http://www.manythings.org/anki/ This dataset contains language translation pairs in the format:\n",
    "\n",
    "```\n",
    "I'm on Tom's side.\t我站在湯姆一邊。\tCC-BY 2.0 (France) Attribution: tatoeba.org #2549046 (CK) & #6065739 (verdastelo9604)\n",
    "```\n",
    "\n",
    "There are a variety of languages available, but we'll use the English-Chinese dataset. After downloading the dataset, here are the steps we'll take to prepare the data:\n",
    "\n",
    "1. Add a start and end token to each sentence.\n",
    "2. Clean the sentences by removing special characters."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "kRVATYOgJs1b"
   },
   "outputs": [],
   "source": [
    "# Setup data path\n",
    "DATA_PATH = 'cmn.txt'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 204
    },
    "colab_type": "code",
    "id": "rd0jw-eC3jEh",
    "outputId": "b16a984b-077e-450d-d4b0-6b16bc38c936"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>en</th>\n",
       "      <th>cn</th>\n",
       "      <th>cc</th>\n",
       "      <th>en_cutted</th>\n",
       "      <th>cn_cutted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Hi.</td>\n",
       "      <td>嗨。</td>\n",
       "      <td>CC-BY 2.0 (France) Attribution: tatoeba.org #5...</td>\n",
       "      <td>[hi, .]</td>\n",
       "      <td>[嗨, 。]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Hi.</td>\n",
       "      <td>你好。</td>\n",
       "      <td>CC-BY 2.0 (France) Attribution: tatoeba.org #5...</td>\n",
       "      <td>[hi, .]</td>\n",
       "      <td>[你, 好, 。]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Run.</td>\n",
       "      <td>你用跑的。</td>\n",
       "      <td>CC-BY 2.0 (France) Attribution: tatoeba.org #4...</td>\n",
       "      <td>[run, .]</td>\n",
       "      <td>[你, 用, 跑, 的, 。]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Wait!</td>\n",
       "      <td>等等！</td>\n",
       "      <td>CC-BY 2.0 (France) Attribution: tatoeba.org #1...</td>\n",
       "      <td>[wait, !]</td>\n",
       "      <td>[等, 等, ！]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Wait!</td>\n",
       "      <td>等一下！</td>\n",
       "      <td>CC-BY 2.0 (France) Attribution: tatoeba.org #1...</td>\n",
       "      <td>[wait, !]</td>\n",
       "      <td>[等, 一, 下, ！]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      en     cn  ...  en_cutted        cn_cutted\n",
       "0    Hi.     嗨。  ...    [hi, .]           [嗨, 。]\n",
       "1    Hi.    你好。  ...    [hi, .]        [你, 好, 。]\n",
       "2   Run.  你用跑的。  ...   [run, .]  [你, 用, 跑, 的, 。]\n",
       "3  Wait!    等等！  ...  [wait, !]        [等, 等, ！]\n",
       "4  Wait!   等一下！  ...  [wait, !]     [等, 一, 下, ！]\n",
       "\n",
       "[5 rows x 5 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from hanziconv import HanziConv\n",
    "from segtok.tokenizer import word_tokenizer\n",
    "\n",
    "import tensorflow as tf\n",
    "\n",
    "df = pd.read_csv(DATA_PATH, header=None, sep='\\t')\n",
    "\n",
    "df.columns = ['en', 'cn', 'cc']\n",
    "\n",
    "df['cn'] = df['cn'].apply(lambda x: HanziConv.toSimplified(x))\n",
    "df['en_cutted'] = df['en'].apply(lambda x: word_tokenizer(x.lower()))\n",
    "df['cn_cutted'] = df['cn'].apply(lambda x: list(x))\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 68
    },
    "colab_type": "code",
    "id": "xrWZxnucKbAc",
    "outputId": "8784d80c-69ec-41ec-b8f0-425373acf7dc"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input sentence samples  : [list(['hi', '.']) list(['hi', '.']) list(['run', '.'])]\n",
      "Output sentence samples : [list(['嗨', '。']) list(['你', '好', '。']) list(['你', '用', '跑', '的', '。'])]\n",
      "Total sentence count    : 8309\n"
     ]
    }
   ],
   "source": [
    "input_datas = df['en_cutted'].values\n",
    "target_datas = df['cn_cutted'].values\n",
    "\n",
    "print(f'Input sentence samples  : {input_datas[:3]}')\n",
    "print(f'Output sentence samples : {target_datas[:3]}')\n",
    "print(f'Total sentence count    : {len(input_datas)}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "WGS-prwHKbAf"
   },
   "source": [
    "## Build model and fit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "W-12T_VzK_Cr",
    "outputId": "8f0e0a93-bf91-418d-ea6c-954a07d04aab"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Seq2Seq API is experimental. It may be changed in the future without notice.\n"
     ]
    }
   ],
   "source": [
    "from kashgari.tasks.seq2seq.model import Seq2Seq\n",
    "model = Seq2Seq()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "oo5EeREsKbAi"
   },
   "outputs": [],
   "source": [
    "class CustomCallback(tf.keras.callbacks.Callback):\n",
    "    def __init__(self, model):\n",
    "        self.model = model\n",
    "        self.sample_count = 5\n",
    "\n",
    "    def on_epoch_end(self, epoch, logs=None):\n",
    "        if epoch % 4 != 0:\n",
    "            return\n",
    "        import random\n",
    "        samples = random.sample(list(input_datas), self.sample_count)\n",
    "        translates, _ = self.model.predict(samples)\n",
    "        print()\n",
    "        for index in range(len(samples)):\n",
    "            print(f\"English: {' '.join(samples[index])}\")\n",
    "            print(f\"Chinese: {''.join(translates[index])}\")\n",
    "            print('------------------------------')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "XPK8ghW0LNdn"
   },
   "outputs": [],
   "source": [
    "his_callback = CustomCallback(model)\n",
    "history = model.fit(input_datas,\n",
    "                    target_datas,\n",
    "                    callbacks=[his_callback],\n",
    "                    epochs=50,\n",
    "                    batch_size=64)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 612
    },
    "colab_type": "code",
    "id": "bF3kv_IlN1rK",
    "outputId": "221f76cd-6e14-427e-ab0a-3a062af91bf5"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "---------------------------\n",
      "input sentence  : i'm glad i was there .\n",
      "input idx       : [  2  19 379   5  27  70   4   3   0]\n",
      "output idx      : [5, 23, 104, 175, 5, 18, 28, 37, 4, 3]\n",
      "output sentence : 我 很 高 兴 我 在 那 里 。\n",
      "\n",
      "---------------------------\n",
      "input sentence  : that young lady is a nurse .\n",
      "input idx       : [   2   28  436  977    8    9 1366    4    3]\n",
      "output idx      : [17, 21, 166, 290, 239, 114, 11, 518, 228, 4, 3]\n",
      "output sentence : 这 个 年 轻 女 孩 是 护 士 。\n",
      "\n",
      "---------------------------\n",
      "input sentence  : give me my beer .\n",
      "input idx       : [  2 107  15  17 495   4   3   0   0]\n",
      "output idx      : [71, 5, 8, 594, 257, 44, 5, 4, 3]\n",
      "output sentence : 把 我 的 啤 酒 给 我 。\n",
      "\n",
      "---------------------------\n",
      "input sentence  : do you know where he lives ?\n",
      "input idx       : [  2  16   7  49  81  13 422   6   3]\n",
      "output idx      : [6, 64, 59, 10, 176, 18, 19, 9, 3]\n",
      "output sentence : 你 知 道 他 住 在 吗 ？\n",
      "\n",
      "---------------------------\n",
      "input sentence  : it rained hard last night .\n",
      "input idx       : [  2  14 987 170 176 205   4   3   0]\n",
      "output idx      : [255, 141, 305, 55, 47, 23, 73, 4, 3]\n",
      "output sentence : 昨 晚 雨 下 得 很 大 。\n",
      "i'm glad i was there .                   -> 我很高兴我在那里。\n",
      "that young lady is a nurse .             -> 这个年轻女孩是护士。\n",
      "give me my beer .                        -> 把我的啤酒给我。\n",
      "do you know where he lives ?             -> 你知道他住在吗？\n",
      "it rained hard last night .              -> 昨晚雨下得很大。\n"
     ]
    }
   ],
   "source": [
    "import random\n",
    "x = random.sample(list(input_datas), 5)\n",
    "transalte, attention = model.predict(x, debug_info=True)\n",
    "\n",
    "for i in range(len(x)):\n",
    "    print(f\"{' '.join(x[i]):<40} -> {''.join(transalte[i])}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ZkBG4u5SKbAt"
   },
   "source": [
    "## Visualize\n",
    "\n",
    "In order to render Chinese, need to install Chinese characters (on Ubuntu / Colab)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 272
    },
    "colab_type": "code",
    "id": "e9AGkfOSKbAt",
    "outputId": "fd720ed9-f980-455f-c547-096f67c879f4"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--2020-05-21 05:25:53--  https://www.wfonts.com/download/data/2014/06/01/simhei/simhei.zip\n",
      "Resolving www.wfonts.com (www.wfonts.com)... 104.225.219.210\n",
      "Connecting to www.wfonts.com (www.wfonts.com)|104.225.219.210|:443... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 10546298 (10M) [application/octetstream]\n",
      "Saving to: ‘simhei.zip’\n",
      "\n",
      "simhei.zip          100%[===================>]  10.06M  5.68MB/s    in 1.8s    \n",
      "\n",
      "2020-05-21 05:25:56 (5.68 MB/s) - ‘simhei.zip’ saved [10546298/10546298]\n",
      "\n",
      "Archive:  simhei.zip\n",
      "  inflating: chinese.simhei.ttf      \n",
      "  inflating: SimHei.ttf              \n",
      "  inflating: sharefonts.net.txt      \n"
     ]
    }
   ],
   "source": [
    "# Download the target font\n",
    "!wget \"https://www.wfonts.com/download/data/2014/06/01/simhei/simhei.zip\"\n",
    "!unzip \"simhei.zip\"\n",
    "!mv SimHei.ttf /usr/share/fonts/truetype/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "V6XidIXbKbAv"
   },
   "outputs": [],
   "source": [
    "# function for plotting the attention weights\n",
    "from matplotlib import font_manager\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.ticker as ticker\n",
    "\n",
    "%matplotlib inline\n",
    "\n",
    "font_name = 'SIMHEI'\n",
    "plt.rcParams['font.family'] = font_name #用来正常显示中文标签\n",
    "plt.rcParams['axes.unicode_minus']=False #用来正常显示负号\n",
    "zhfont = font_manager.FontProperties(fname='/usr/share/fonts/truetype/SimHei.ttf')\n",
    "\n",
    "# 如果绘制图表不清晰，可以设置高 dpi 来提高图表清晰度\n",
    "plt.rcParams['figure.dpi'] = 120"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 565
    },
    "colab_type": "code",
    "id": "FvjxnhWNoSau",
    "outputId": "d298e5b9-258b-4e32-ea0e-5d26736c035a"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:matplotlib.font_manager:findfont: Font family ['SIMHEI'] not found. Falling back to DejaVu Sans.\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAicAAAITCAYAAADcu7xmAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAASdAAAEnQB3mYfeAAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nOzdeZgdVZ3/8ffpPUl3drIvQBKWLEDCvskmoAiigCiogI6K8kMdHVFHx4Wfjo7bzCjICIqIuPxkcUBcQNkh7CQBEkhIgOxLZ+2kk977/P64t5smdJJO0t1V997363nq6dt1q25/qfB0Pjnn1LdCjBFJkqS0KEq6AEmSpI4MJ5IkKVUMJ5IkKVUMJ5IkKVUMJ5IkKVUMJ5IkKVUMJ5IkKVUMJ5IkKVUMJ5IkKVUMJ5IkKVVKki4gV4QQBgAnAcuAxoTLkSQpl5QBY4GHY4w1uzrYcNJ1JwF3JV2EJEk57FzgT7s6yHDSdcsA7rzzTiZOnJh0LZIk5YxFixbxnve8B7J/l+6K4aTrGgEmTpzIlClTkq5FkqRc1KVlES6IlSRJqWI4kSRJqWI4kSRJqWI4kSRJqWI4kSRJqWI4kSRJqWI4kSRJqWI4kSRJqWI4kSRJqWI4kSRJqWI4kSRJqWI4kSRJqWI4kSRJqWI4kSRJqWI4kSRJqWI4SVCMkfW1DUmXIUlSqpQkXUCh+uVjr/ODexfQ0NzCgm+/k9Jic6IkSeDISWL6lRdT19RCa4SVm+qSLkeSpNQwnCRk7OC+7a+XbTCcSJLUxnCSkLGDOoSTjdsSrESSpHQxnCRk5IAKiosCAEs3GE4kSWpjOElISXERowf2AWCZ4USSpHaGkwSNHZwNJxtdcyJJUhvDSYLa1p04ciJJ0hsMJwlqu2Nnw9ZGahuaE65GkqR0MJwk6M23Ezt6IkkSGE4SNXZQn/bXhhNJkjIMJwka13HkxEWxkiQBhpNEDe5XRt+yYsCRE0mS2hhOEhRC8I4dSZK2YzhJWNuiWFvYS5KUYThJWHsjtg11xBgTrkaSpOQZThLWNq1T19TCutrGhKuRJCl5hpOEvfmOHad2JElKbTgJIRwZQrg2hDAvhLA1hLA0hHBrCOGALp4/MIRwQwhhbfb8B0MIM3q67t1lIzZJkt6sJOkCduJLwPHAbcALwAjgSmBWCOGYGOPcHZ0YQigC/gIcCvwAWAdcATwUQjg8xriwp4vvqjE2YpMk6U3SHE7+E7g4xti+ECOE8AfgReDLwId2cu4FwHHA+2KMt2fPvRV4BbgauLinit5d/cpLGFpZxrraRpYaTiRJSu+0Tozx8Y7BJLtvITAPOHgXp18ArAH+2OHctcCtwLkhhPJuLnevjGnvdWKXWEmSUhtOOhNCCMBwMtM0OzMdmBVjbN1u/9NAX2Cn61ZCCMNCCFM6bsCEPa17V8bZ60SSpHY5FU6ADwKjgT/s4riRwKpO9rftG7WL868A5m633dX1MndPW6+TlZvqaGrZPk9JklRYciachBAOAn4KPAHcvIvD+wANneyv7/D+zlwHTN1uO7fLxe6mtl4nrRFWbarfxdGSJOW3NC+IbRdCGEHm7psa4IIYY8suTqkDOltXUtHh/R2KMVYD1dvV0LVi98D2vU7GDem7k6MlScpvqR85CSEMAP4GDATeEWNc2YXTVpGZ2tle276ufEav6djrxDt2JEmFLtXhJIRQAdxNZgHr2THGl7p46hxgRrbfSUdHA9vI3FKcGiMHVFBclBmZsdeJJKnQpTachBCKySx8PZZMv5IndnDcyBDCQSGE0g67bydzV895HY4bCrwPuDvG2Nl6lMSUFBcxamBmxmnZRm8nliQVtjSvOfkR8G4yIyeDQwhvaroWY/xN9uV3gUuB/YDF2X23A08CN4UQJvNGh9hi4Bs9XvkeGDuoL8s21DmtI0kqeGkOJ4dlv56T3bb3m072ARBjbAkhnEWmdf1nyNyd8wxwWYxxQXcX2h0yd+ysZ7nhRJJU4FIbTmKMJ3fxuMuAyzrZvxH4WHZLvbY7dNZvbWRrQzP9ylP7RyNJUo9K7ZqTQvOmBwDaKVaSVMAMJynR8XZin7EjSSpkhpOUGGevE0mSAMNJagzpV0af0mLAXieSpMJmOEmJEEL7AwCXu+ZEklTADCcp0ja147SOJKmQGU5SZEz26cTLNtQRY0y4GkmSkmE4SZG2O3bqmlpYv7Ux4WokSUqG4SRFvGNHkiTDSaq0LYgF79iRJBUuw0mKjB3UsRGb4USSVJgMJynSr7yEIf3KALvESpIKl+EkZcZk1534fB1JUqEynKSMvU4kSYXOcJIyY7NPJ15VU09zS2vC1UiS1PsMJynT1uukpTWyqqY+4WokSep9hpOUsdeJJKnQGU5SxtuJJUmFznCSMiMHVlAUMq+9Y0eSVIgMJylTWlzEqIGZRbFL7XUiSSpAhpMUGtv+dGJHTiRJhcdwkkJtz9hZ7rSOJKkAGU5SqO2OnXW1jWxtaE64GkmSepfhJIXGdridePlG151IkgqL4SSFxgyy14kkqXAZTlKoYyM2F8VKkgqN4SSFhlaW0ae0GLDXiSSp8BhOUiiEwJjsAwAdOZEkFRrDSUq1Te0ssxGbJKnAGE5Squ2OnWUbtxFjTLgaSZJ6j+EkpdrCybbGFtZvbUy4GkmSeo/hJKXGZtecgOtOJEmFxXCSUh0bsS2zEZskqYAYTlJqrL1OJEkFynCSUpXlJQzuVwYYTiRJhcVwkmJttxMvqq5NuBJJknqP4STFDh0zAIAXV9TQ2NyacDWSJPUOw0mKzRg/CICG5lZeWrU54WokSeodhpMUmzFuUPvr55ZsTLASSZJ6j+EkxcYM6sM+VeUAzFpqOJEkFQbDSYqFEDg8O3oyy5ETSVKBMJyk3IzxAwFYVVPPyk02Y5Mk5T/DScodPv6NdSdO7UiSCoHhJOWmjBpAaXEAYNaSTQlXI0lSzzOcpFxFaTFTR2f6nTznyIkkqQAYTnJA2y3F81bUUN/UknA1kiT1LMNJDmhbd9LcGnlxRU3C1UiS1LMMJzmg46JYm7FJkvKd4SQHDO9fweiBfQD7nUiS8p/hJEe0PWdn1tKNxBgTrkaSpJ5jOMkRM8ZlmrGtq21k2QabsUmS8pfhJEe8ad3J0g0JViJJUs8ynOSIg0f2p6I088flolhJUj4znOSI0uIiDhmTmdqxU6wkKZ8ZTnJIWzO2+as3U9vQnHA1kiT1DMNJDmlbd9Ia4YVljp5IkvKT4SSHTM/esQOuO5Ek5S/DSQ4ZWlnOvkP6Apl+J5Ik5SPDSY55oxnbJlpbbcYmSco/hpMc07YotqauidfW1SZcjSRJ3c9wkmM6NmPzlmJJUj4ynOSYA4ZXUVleArgoVpKUnwwnOaa4KHDY2GwzNhfFSpLykOEkB7U9BHBhdS0125oSrkaSpO5lOMlBMzqsO5m9zNETSVJ+MZzkoOljOy6KNZxIkvKL4SQHDehbyqRhlQA857oTSVKeMZzkqLZ+J3OWbqLFZmySpDxiOMlRbf1Otja2sGD1loSrkSSp+xhOclTHRbHeUixJyieGkxy1/9B+DOhTCrgoVpKUXwwnOaqoKLT3O3HkRJKUTwwnOaxtUezi9dtYV9uQcDWSJHUPw0kO6/gQQJ+zI0nKF4aTHHbYuIGUFgcAnnh1fcLVSJLUPQwnOaxvWUn71M6jC9cmXI0kSd3DcJLjTpg4FIBX125lVU1dwtVIkrT3DCc57vhJQ9tfz1zk1I4kKfcZTnLcIaMHUFVeAsDMResSrkaSpL1nOMlxJcVFHDNhCACPLVpHjD5nR5KU2wwneeDE7NTO2i0NLKyuTbgaSZL2juEkDxw/8Y11J48tdGpHkpTbDCd5YP+h/Rg5oALITO1IkpTLDCd5IITQPnry5GvraWppTbgiSZL2nOEkT7T1O9nW2MKcZZsSrkaSpD2X6nASQqgMIVwdQrgnhLAhhBBDCJd18dzLssd3to3o4dJ7netOJEn5ItXhBBgKfB04GHh+Dz/j68CHt9vybmhhn6pyDhpRBdjvRJKU20qSLmAXVgEjY4yrQwhHAM/swWf8Lcb4bDfXlUrHTxzK/NVbmL1sE1vqm6iqKE26JEmSdluqR05ijA0xxtV7+zkhhKoQQnF31JRmbetOWlojT7++IeFqJEnaM6kOJ93kQWAzsC2E8KcQwqSkC+opR+03mJKiAMCjrjuRJOWotE/r7I1twK94I5wcDnweeDyEMCPGuGxHJ4YQhgH7bLd7Qg/V2W36lZcwY9wgnl68wXUnkqSclbfhJMZ4K3Brh113hhDuBR4Bvgp8cienXwF8owfL6zEnTBrK04s3sLC6ljWb6xnevyLpkiRJ2i2FMK3TLsb4GPAU8PZdHHodMHW77dyera57dLyl2NETSVIuKqhwkrUMGLyzA2KM1THGeR034NXeKW/vHDpmAJXlmQExW9lLknJRIYaT/YG1SRfRU0qKizhm/yFAZuQkxphwRZIk7Z68CCchhJEhhINCCKUd9m2/oJUQwllkFsbe05v19bYTJmbCyZrNDSyqrk24GkmSdk/qF8SGEK4EBgKjsrvOCSGMyb6+JsZYA3wXuBTYD1icfe/xEMJs4FmgBpgBfJTMtM53eqf6ZJww6Y1c9tiidUwaXpVgNZIk7Z7UhxPgC8D4Dt+fl90AfkMmeHTmD8C7gDOAvmS6zf4cuDrGuKZnSk2HCfv0Y0T/ClZvrmfmonV85Pj9ki5JkqQuS/20Toxx3xhj2MG2OHvMZR2/z+77txjj9BjjwBhjWYxxfIzxinwPJgAhhPa7dp58bQNNLa0JVyRJUtelPpxoz5wwKbPupLahmReW591zDiVJecxwkqeOn/BGv5PHFq5PsBJJknaP4SRPDetfwQHDKwGbsUmScovhJI+dMDFz186spRupbWhOuBpJkrrGcJLH2tadNLdGnn7dqR1JUm4wnOSxo/YbQklRAFx3IknKHYaTPFZZXsL0cQMBeGxR3nbslyTlGcNJnmtbd/LKmlqWbdiWcDWSJO2a4STPnT55ePvre+etTrASSZK6xnCS5w4eWcXYwX0A+Pu8vG+OK0nKA4aTPBdC4MzJIwB4ZskG1tU2JFyRJEk7ZzgpAGdOzYSTGOG+lxw9kSSlm+GkAMwYN4ihlWWA604kSelnOCkAxUWhfWHszEXr2VLflHBFkiTtmOGkQJwxJTO109jSykML7HkiSUovw0mBOG7CECrLSwCndiRJ6WY4KRDlJcWcctAwAB5asJaG5paEK5IkqXOGkwJy5pTMupPahmYeX+SzdiRJ6WQ4KSAnHziMspLMH7lTO5KktDKcFJDK8hJOmDgUgH+8tIaW1phwRZIkvZXhpMC0Te2s39rIc0s2JlyNJElvZTgpMG8/eDhFIfPaqR1JUhoZTgrMkMpyjtx3MJAJJzE6tSNJShfDSQE6M9uQbfnGOuat3JxwNZIkvZnhpACdkV13AvB3p3YkSSljOClAYwb1Zero/gDcO8+nFEuS0sVwUqDOnJyZ2lmwZguL121NuBpJkt5gOClQZ04d0f7au3YkSWliOClQk4ZVst/QfoDhRJKULoaTAhVCaF8YO2vpJqo31ydckSRJGYaTAtZ2SzHA319yYawkKR0MJwXssDEDGVZVDji1I0lKD8NJASsqemNq54lX11NT15RwRZIkGU4KXtvUTnNr5MH51QlXI0mS4aTgHbP/EPpXlABw9/MrE65GkiTDScErLS7iXYeMBOChV9aydktDwhVJkgqd4URccPgYAFpaI3fNWZFwNZKkQmc4ETPGDWpvyHb7c8uJMSZckSSpkBlORAihffRk/uotzFu5OeGKJEmFzHAiAN47fTQhZF7f/tzyZIuRJBU0w4kAGDWwD8dPGArAXXNW0NjcmnBFkqRCZThRu7apnY3bmnjAnieSpIQYTtTuzCkjqCzP9Dy5Y5ZTO5KkZBhO1K5PWTFnZ3uePDi/mnW19jyRJPU+w4ne5Pzs1E5za+SuOXaMlST1PsOJ3uSI8YMYP6Qv4F07kqRkGE70JiEELpiRGT15edVm5q2sSbgiSVKhMZzoLc47fEx7z5M7nrOdvSSpdxlO9BajB/bh2P2HAHCnPU8kSb3McKJOtfU82bC1kYcW2PNEktR7DCfq1DumjqBfWTFgzxNJUu8ynKhTfctKeFe258n9L1ez3p4nkqReYjjRDp0/442eJ3963p4nkqTeYTjRDh2572DGDbbniSSpdxlOtENFRaF99GTeys28vGpzwhVJkgqB4UQ7dd6M0e2v73D0RJLUCwwn2qmxg/tyzP6DgUzPk6YWe55IknqW4US7dMHhYwFYV9vIP15ak3A1kqR8ZzjRLr1r2kgG9CkF4KaZrydcjSQp3xlOtEt9yor5wFGZ0ZNnFm9k7gofBihJ6jmGE3XJJcfuS3FR5mmAv3T0RJLUgwwn6pLRA/tw5pThAPz5+VWs3WLHWElSz+iWcBJCGBdCOGG7fYeGEH4dQvhDCOE93fFzlKyPHL8fAI0trfz2qSUJVyNJylfdNXLyE+Cbbd+EEIYDDwLnAW8D7gghnNdNP0sJOWL8IKaO7g/Ab55cSkNzS8IVSZLyUXeFk6OAf3T4/hKgD3AoMBq4H/hCN/0sJSSEwEeOy4yerKtt4C8vrEq4IklSPuqucDIYqO7w/dnAwzHGV2OMrcAfgYO66WcpQWcfOpKhleUA3DRzMTHGhCuSJOWb7gona4HxACGEgcAxwL0d3i/Jbspx5SXFfPDocQC8uKKG55ZsTLgiSVK+6a5wch/wmRDC54FfZz/3zg7vTwaWddPPUsI+eMw4SosztxXfNHNxssVIkvJOd4WTLwMvAz8EzgC+EGN8HSCEUA5cSGbdifLAsKoKzjlkFAD3zFvNyk11CVckScon3RJOYoxrYozHA4OA/jHGH2/3M06jw908yn1ttxW3tEZ+/YS3FUuSuk+3NmGLMdbEGBu321cXY3w+xrihO3+WkjVtzACOGD8IgN8/vZS6Rm8rliR1j+5qwnZaCOGq7fZ9NISwNISwJoTwXyGE4u74WUqPj56QGT2pqWvif2evSLgaSVK+6K6Rk2+S6WkCQAhhGnA9mbt4HgI+g31O8s4Zk4czemAfAH71+OveVixJ6hbdFU4OBp7t8P2Hgc3AiTHG9wM/J9OYTXmkpLiIDx87HoBX1tQyc9H6hCuSJOWD7gon/ciEkTbvAO6JMW7Lfv8M2T4oyi8fOHIsFaWZ/41u8mnFkqRu0F3hZBlwJEAIYSIwFfh7h/cHAz7GNg8N7FvGeTPGAPDAgmoWr9uacEWSpFzXXeHkt8AnQgh/ItMZdiNwV4f3Dwde6aafpZT5yHH7AhAj/OrxxYnWIknKfd0VTv4d+A9gLLAUeE+McRNACGEwcDLwp276WUqZScOrOHHSUABufXYZG7Y27uIMSZJ2rLuasDXHGL8aY5weYzwlxvhoh/c2xBhHxBi/2x0/S+l0+dsmALCtsYUbH3st4WokSbmsW5uwAYQQKkMIB2e3yu7+fKXT8ROHMH3cQABufnwJNduaEq5IkpSrui2chBCODCE8SGa9ydzstjGE8EAI4Yju+jlKpxACnzl1EgC1Dc2uPZEk7bHu6hB7NPAIMAP4BfC57PaL7L5HQghHdcfPUnqdfOA+TB3dH4BfznydLfWOnkiSdl93LohdARwYY/xUjPEn2e1TwIHAyuwxymMhBK48JTN6UlPXxC1P+kBASdLu665wcjRwfYxx9fZvxBjXADcAx3TTz1KKnTF5OAcOrwLgF4++zrbG5oQrkiTlmu4KJ61AyU7eL84eozxXVBS48tSJAGzY2sjvnlqacEWSpFzTXeHkceD/hBDe0qI+hDAOuAKY2U0/Syl31rSR7L9PPwCuf+Q16ptaEq5IkpRLuiucfAUYAMwPIfwuhPDN7PZ7YD4wEPjX3f3Q7G3JV4cQ7gkhbAghxBDCZbtx/sAQwg0hhLUhhK0hhAdDCDN2tw7tnuKiwJWnZEZP1m5p4NZnlyVckSQpl3RXE7bZZNad3AO8G/h6djsnu+94YO0efPTQ7OccDDy/OyeGEIqAvwAXA9cCXwSGAQ+FECbtQS3aDe8+dBTjBvcF4GcPvUpjs7N6kqSu6bY+JzHGl2KM7wX6AyOzW/8Y43lkQsqe/PN5FTAyxjgeuGo3z70AOA64LMZ4dYzxp2Ta6LcAV+9BLdoNJcVFXHFypmvsypp67pi1POGKJEm5ots7xMYYW2OMa7LbXv1zOcbY0NkdQF10AbAG+GOHz1sL3AqcG0Io35vatGvnzRjD6IF9ALjuoUU0tTh6IknatW4PJykyHZjVSUB6GugLHND7JRWWspIiPnnS/gAs21DHn+asTLgiSVIuyOdwMpLMtND22vaN2tGJIYRhIYQpHTdgQk8Ume/ed8RYhlVlBql++uAiWlpjwhVJktIun8NJH6Chk/31Hd7fkSt44/lAbdtd3VpdgagoLeYTb8uMnry2bit/ebGzvChJ0ht21jhtp3bzltwdjlL0oDqgs3UlFR3e35HrgNu22zcBA8oe+eDR4/mfh15l/dZGrn1gIWdPG0lRUUi6LElSSu1xOAGeBbo6Rh9249jusorM1M722vbtcAFEjLEaqO64LwT/Mt1TfcqK+diJ+/O9e+bzyppa7p23mndO6+yPRpKkvQsnH+m2KnrGHODEEELRdotijwa2Aa8kU1Zh+vCx47n+kVfZtK2J//zHK5wxZQTFjp5Ikjqxx+EkxnhzdxayN0III8l0qH01xtiU3X07mduJz8u+JoQwFHgfcHeMsbP1KOohleUlfOqkCXz3b/NZWF3LnbNXcP7hY5IuS5KUQqlfEBtCuDKE8G/AR7O7zgkh/Ft2G5Dd913gZWB0h1NvB54EbgohfD2EcAXwEJmHEH6jd6pXR5ccu2/7nTv/dd8rdo2VJHUq9eEE+ALwLeBT2e/Py37/LWDQjk6KMbYAZwF/AD4D/ABYB5waY1zQkwWrc33KivnMaZknByzfWMfvn/aJxZKkt0p9OIkx7htjDDvYFmePuazj9x3O3Rhj/FiMcWiMsV+M8eQY47NJ/Hco48IjxrY/c+eaBxaxrbE54YokSWmT+nCi/FJWUsTnT880511X28BNMxcnW5AkKXUMJ+p17z50FAeNqALg+odfpWZb0y7OkCQVEsOJel1RUeBfzjgQgM31zVz/yKsJVyRJShPDiRLx9oOHMWPcQABumrmY6i31uzhDklQoDCdKRAiBq848CIC6phaufWBRwhVJktLCcKLEHDthCCdOGgrA759eyrIN2xKuSJKUBoYTJeqqMzNrT5paIv91n08UkCQZTpSwQ8YM5B1TRgDwv7NX8MqaLQlXJElKmuFEifvCmQdQFCBG+NHfbd4rSYXOcKLETRxWxXkzMg8BvHfeGuYs25RwRZKkJBlOlAqfPW0SpcUBgB/cOz/haiRJSTKcKBXGDu7LB48eD8DMRet5cH51whVJkpJiOFFqXHnqRKrKSwD41p9forG5NeGKJElJMJwoNYZWlvPZt08C4LV1W/nV468nXJEkKQmGE6XKJcfuy/779APgJ/cvYu2WhoQrkiT1NsOJUqWspIivnz0ZgNqGZhfHSlIBMpwodU4+cBinHTQMgNueW84Ly721WJIKieFEqfRvZ0+mtDgQI3zzT/OIMSZdkiSplxhOlEr7De3HR4/fD4BZSzdx15yVCVckSeothhOl1pWnTmRoZTkA3/3by2xtaE64IklSbzCcKLWqKkr50jsyTy1es7mB6x5alHBFkqTeYDhRqp0/YwyHjhkAwM8ffZ2l67clXJEkqacZTpRqRUWBb7x7CgCNza38+19fSrgiSVJPM5wo9WaMG8R500cDmacWz1y0LuGKJEk9yXCinPCldx5E37JiAK6+ex7NLT53R5LyleFEOWF4/wquPHUiAK+sqeU3Ty5JuCJJUk8xnChnfPT4/Rg3uC8AP/r7K6zZXJ9wRZKknmA4Uc6oKC3m6nMzi2O3NDRz9d3zEq5IktQTDCfKKaccOIx3HTISgL++uJr7X16TcEWSpO5mOFHO+cY5k6mqKAHg63fNs3OsJOUZw4lyzrCqCr78zoMAWLGpjv/8xysJVyRJ6k6GE+Wki44cxxHjBwFw08zXeXF5TcIVSZK6i+FEOamoKPCd86ZRUhRojfCv//uCvU8kKU8YTpSzDhhexeUn7Q/A3BWbufkJe59IUj4wnCinffrUSew7pK33yQJWbKpLuCJJ0t4ynCinVZQW8+/vnQbAtsYWvnHXXGKMCVclSdobhhPlvOMnDm1/MOB9L1dzz9zVCVckSdobhhPlha++62AG9i0F4Bt/msfm+qaEK5Ik7SnDifLCkMpyvnLWwQBUb2ngh/cuSLgiSdKeMpwob7zv8DEcs/9gAG55cgnPLdmYcEWSpD1hOFHeCCHw7++dRllxETHCl+54gfqmlqTLkiTtJsOJ8sqEfSr57NsnAbCoupZrHliYcEWSpN1lOFHe+cTb9mfq6P4A/Ozh15i7wtb2kpRLDCfKO6XFRXz//EMpKQq0tEa+cNvzNDbb2l6ScoXhRHlp8qj+XHHKRADmr97Czx5+NeGKJEldZThR3rrylIkcOLwKgGseWMiC1VsSrkiS1BWGE+WtspIivn/BIRQFaGqJfPH2531ysSTlAMOJ8tqhYwfy8bdlnlz8/PIabnzs9YQrkiTtiuFEee9zbz+A/Yf2A+BH/3iFV9fWJlyRJGlnDCfKexWlxXz/gkMIARqbW/nS7S/Q2uqTiyUprQwnKghH7DuYS4/dF4Bnl2zk108sTrIcSdJOGE5UML74jgMZO7gPAN+7ZwHLNmxLuCJJUmcMJyoYfctK+N55hwBQ19TCVd69I0mpZDhRQTlu4lAuPnocAE++toHv/HV+whVJkrZnOFHB+epZB3PQiExztl/OfJ1bn1mWcEWSpI4MJyo4/cpL+PklRzC4XxkAX73zRZ5dvCHhqiRJbQwnKkhjB/flZx86nJKiQFNL5JO/eY4Vm+qSLkuShOFEBeyo/QbzrfdMBWBdbSMfv/lZtjU2J1yVJMlwooJ20VHjuPTY8QC8tGozV932AjHaoE2SkmQ4UcH72tmTOX7iEH20pLwAACAASURBVAD+8uIqrnlgUcIVSVJhM5yo4JUUF3HtRTMYP6QvAP/5j1e4Z+7qhKuSpMJlOJGAQf3K+MUlR1BZXgLA52+dw8urNidclSQVJsOJlDVpeBU//sBhhADbGlv42M3Psr62IemyJKngGE6kDk47eDhfPPMgAFZsquOz/2+OTzCWpF5mOJG288mT9ufdh44C4LFF6/ifh19NuCJJKiyGE2k7IQS+c9409h/aD4Af/X0BT79uB1lJ6i2GE6kTleUlXHPxdMpKimiN8Jnfz2bD1saky5KkgmA4kXZgyqgBfO3syQCs3lzPF2573vUnktQLDCfSTnzo6HGcNW0EAA/Mr+bGx15PuCJJyn+GE2knQgj8x/mHMHZwHwC+d898Zi/dmHBVkpTfDCfSLvSvKOXai2ZQWhxobo1c+bvZ1GxrSrosScpbhhOpCw4dO5Avv/NgINP/5Et3+IBASeophhOpiz56/L68/eDhANwzbzW3PLkk4YokKT8ZTqQuCiHww/cdwqgBFQB8+88vM3dFTcJVSVL+MZxIu2Fg3zKuuXg6xUWBxpZWrvzdLGrqXH8iSd3JcCLtpsPHD+YLZxwIwOL127jyd7NobmlNuCpJyh+GE2kPXP62/TlzSmb9yaML13H13S8lXJEk5Q/DibQHiooC//X+w5gyqj8Atzy5hJsfX5xsUZKUJwwn0h7qW1bCjZceybCqcgCuvnseDy2oTrgqScp9hhNpL4wYUMEvLj2CitLMAwKv/N1sXlmzJemyJCmnGU6kvXTImIH854WHAVDb0MxHf/UM62sbEq5KknKX4UTqBmdNG8lVZ2bu4Fm+sY7Lb3mOhuaWhKuSpNxkOJG6yRUnT+C86aMBeHbJRv71jhdtcS9Je8BwInWTEALfPX8aR4wfBMAfZ6/guodeTbgqSco9hhOpG5WXFHP9hw9n7OA+APzg3gXc+Njr1DU6xSNJXWU4kbrZkMpybrz0SKrKSwD41p9f4ujv3Me3/vwSr62tTbg6SUq/VIeTEEJ5COF7IYSVIYS6EMJTIYTTu3DeN0MIsZOtvjfqlg4YXsX1lxzO8P6ZHiib65u58bHXOfVHD/PBXzzJ315cRZMt7yWpUyVJF7ALvwIuAP4bWAhcBvw1hHBKjPGxLpz/KaDjP1UdW1evOW7CUB770qnc99IabnlyCY+/uh6AmYvWM3PReob3L+cDR47joqPGMSL7pGNJEoS03k0QQjgKeAq4Ksb4w+y+CmAuUB1jPG4n534T+AawT4xxXTfVMwWYO3fuXKZMmdIdH6kCs6i6lt8+tYTbn1vOlvrm9v1lJUX8zwdncNrBwxOsTpJ6zrx585g6dSrA1BjjvF0dn+ZpnQvIjHTc0LYjxlgP3AgcG0IY24XPCCGE/iGE0EM1Sl02cVgl3zhnCk995TS+d/40po0eAEBjcyufv/V5VmyqS7hCSUqHNIeT6cArMcbN2+1/Ovv1sC58xmtADbAlhPCbEIL/NFXi+paV8P4jx3H3p0/gv9+f+d+4pq6Jz/5+Ns2uQ5GkVIeTkcCqTva37Ru1k3M3AtcCl5MZgfkF8H7g0RBC/1394BDCsBDClI4bMGG3qpe64D3TR/PBo8cBmcZtP75/YcIVSVLy0rwgtg/Q2QNK6ju836kY44+323VHCOFp4LfAFcB/7OJnX0FmzYrU47529mSeXbyRBWu2cO2Dizh2whCOmzA06bIkKTFpHjmpA8o72V/R4f0uizH+DlgNvL0Lh18HTN1uO3d3fp7UVRWlxVx78XQqSouIEf75/83xwYGSClqaw8kqMlM722vbt3IPPnMZMHhXB8UYq2OM8zpugH3I1WMmDa/im+dk7gKr3tLAv9z2PK2t6byTTpJ6WprDyRzggE7WiBzd4f0uy96xsy+wdu9Lk7rf+48cy9mHZLL3QwvW8suZrydckSQlI83h5HagGPhE244QQjnwEeCpGOOy7L5xIYSDOp4YQtink8/7FLAPcE+PVSzthRAC3zlvWvtzeb53z3xeWL4p4aokqfelNpzEGJ8CbgO+G0L4fgjhE8ADZEY/vtjh0F8DL293+pIQwk0hhM+HEK4IIfyOzN07c4Dre756ac/0ryjlJx+YTklRoKkl8unfz2ZLfVPSZUlSr0ptOMm6hEzr+g8DPwFKgbNjjI/s4rzfAkcB38yefyTwfeBtMcZtPVat1A2mjxvEVWceCMCS9dv46v/OJa2dnCWpJ6T5VuK2jrBXZbcdHXNyJ/s+3oNlST3u4yfuz8xX1/PIK2v50/MrOWHSUC48oitNkSUp96V95EQqSEVFgR+971CGVmbupv/GXfOYv3r7ZsmSlJ8MJ1JK7VNVzn+//zBCgLqmFi6/5Tlqtrn+RFL+M5xIKXbCpKF84Yw31p/88x9m2/9EUt4znEgpd8XJEzhzSuaZlQ8uWMt/3/dKwhVJUs8ynEgpF0Lgh+87lAn79APgJw8s4h8vrUm4KknqOYYTKQdUVZRy/YePoLI8c4Pd5/8wh1fX1iZclST1DMOJlCMmDqvkRxceCsCWhmYuv+U5ahuaE65Kkrqf4UTKIWdOGcGVp0wEYFF1LVfd9rwN2iTlHcOJlGM+d/oBvO2AzOOj/jZ3NT97+LWEK5Kk7mU4kXJMcVHgJx84jHGD+wLwg3vn8+hCH7YtKX8YTqQcNLBvGT/70OFUlBbRGuHTv5/Nsg0+NkpSfjCcSDlq8qj+fO/8QwDYtK2JD9/4FCs31SVclSTtPcOJlMPOPWw0n3jb/gAsXr+N99/whCMoknKe4UTKcf/6zoP4yPH7ArBsQx3vv/4JFq/bmmxRkrQXDCdSjgsh8PWzJ3P5SZkRlJU19bz/hidYVG2TNkm5yXAi5YEQAl9+x0F85tRMD5Q1mxv4wA1PsGD1loQrk6TdZziR8kQIgc+fcSD/cvoBAKyrbeQDNzzBvJU1CVcmSbvHcCLlmU+fNol/fedBAGzc1sTFP3+KF5ZvSrgqSeo6w4mUhy4/aQJfP3syADV1TXzw50/x3JKNCVclSV1jOJHy1EdP2I9vv2cqkHlQ4CU3PsVDC6oTrkqSds1wIuWxDx0znu+ffwghwNbGFv7p5me55cklSZclSTtlOJHy3IVHjuWnF8+gvKSIltbI1+6cy/+9+yVaWn2asaR0MpxIBeCsaSP5w+XHMrSyHIBfznydT/z6WbY2NCdcmSS9leFEKhCHjR3Inf/nOA4cXgXA/fOred/PnmBVjc/jkZQuhhOpgIwZ1JfbP3UsJx2wDwAvrdrMe346kxeX2wtFUnoYTqQCU1VRyo2XHsElx44HMt1kL7z+Cf4+b3XClUlShuFEKkAlxUX833On8s1zJlMUoK6phct/8xw/e/hVYnShrKRkGU6kAnbZ8fvxi0uPoF9ZMTHCf/xtPh//9bNs2taYdGmSCpjhRCpwpx40nNs+eRxjB/cB4L6Xqznrx4/y3JINCVcmqVAZTiQxeVR//vzpE3nHlBEArKyp58Lrn+RnD79Kq/1QJPUyw4kkAAb0KeV/PjSDq989hbLiTMO2//jbfD568zNs2Oo0j6TeYziR1C6EwKXH7csdnzqO8UP6AvDQgrWc9eNHefp1p3kk9Q7DiaS3mDZmAHd/+gTedchIAFZvrueinz/JTx9c5DSPpB5nOJHUqf4VpVx70XS+/Z6plGWfy/ODexfwiVueo9a295J6kOFE0g6FEPjQMeP53yuOY7+h/QC47+U1XPA/j7Nsw7aEq5OUrwwnknZpyqgB3HXl8Zx8YKbt/fzVWzj3pzN5ZrHrUCR1P8OJpC7pX1HKjZceycdO2A+ADVsbufjnT3Lrs8sSrkxSvjGcSOqy4qLAv509me+dP43S4kBTS+SLt7/At//8Ei0ulJXUTQwnknbb+48cx28/dgyD+5UB8IvHXuefbn6GzfVNCVcmKR8YTiTtkaP2G8xd/+d4DhxeBWT6oZx33eMsWb814cok5TrDiaQ9NnZwX+644jjefvAwABZV13LONY9xy5NLnOaRtMcMJ5L2SmV5Cdd/+Ag+edIEADbXN/O1O+dyzjWP+fBASXvEcCJprxUXBb78zoP4+SVHMHpg5unGL63azPn/8wSfv3UO1VvqE65QUi4xnEjqNqdPHs79/3ISnz1tEuUlmV8vf5y1glN/+DC/ePQ1mlpaE65QUi4wnEjqVhWlxXzu9AO47/Mnccbk4QDUNjTz7b+8zFk/fpSZi9YlXKGktDOcSOoRYwf35YZLjuDmjx7F/tnW9wura/ngL57iYzc/wwvLNyVcoaS0MpxI6lEnHbAP9/zz2/jyOw+ib1kxAPe9XM27r53JR256mllLNyZcoaS0MZxI6nFlJUV88qQJPPAvJ3PRUWMpLQ4APJjtjfLhG5/yOT2S2hlOJPWaEQMq+O55h/DgF07mQ8eMo6w48yvo0YXreN/PnuCiG57kiVfXJ1ylpKQZTiT1ujGD+vLt90zj4S+ezGXH7UtZ9s6eJ15bz0U/f5ILf/YEj7/qwlmpUBlOJCVm5IA+fPPdU3jsi6fwTyfsR0Vp5lfS04s3cPHPn+KiG550ukcqQIYTSYkb1r+Cr509mUe/eCqfeNv+9CnNLJx94rX1vO9nT/DhG59itgtnpYJhOJGUGvtUlfOVsw7mkexISlsjt0cXruO91z3OR3/1DC8ur0m4Skk9zXAiKXX2qSrna2dP5pEvnsKlx45vXzj7wPxqzrn2MT7+62d56rX11De1JFyppJ4QYvTJoV0RQpgCzJ07dy5TpkxJuhypoKzYVMdPH1zErc8so7nD045LiwOTR/Zn+rhBTB83kBnjBjFmUB9CCAlWK2l78+bNY+rUqQBTY4zzdnW84aSLDCdS8pZt2MY1DyzkjlkraGnt/HfX0MoyDhv7Rlg5dOwA+paV9HKlkjoynPQQw4mUHtWb63l68QZmL93E7KUbmbtiM407eKhgcVHgoBFVzBg3iBnjM4Fl3OC+jq5IvWh3w4n/nJCUc4b1r+DsQ0Zx9iGjAGhobuHlVVuYvXQjs7KBZfnGOgBaWiPzVm5m3srN3PLkEuCN0ZUTJw3loqPGtfdZkZQOjpx0kSMnUm6p3lLP7KWbmLV0I7OWbOSF5TU0NL91dGXGuIFc98HDGTGgIoEqpcLgyIkkAcOqKjhzygjOnDICgMbmVl5etTkTVpZu4unX17NmcwOzlm7i7Gse5ZqLZnDshCEJVy0JvJVYUoEoKyni0LED+cjx+3HNRdN5+KpTuOiocQCsq23kQzc+xQ2PvIqjyVLyDCeSClJFaTHfPW8a3z//EMpKimhpjXznr/O54rezqG1oTro8qaAZTiQVtAuPHMsfP3UcYwb1AeBvc1fz7msfY+GaLQlXJhUuw4mkgjd19AD+/OkTOOmAfQB4be1Wzv3pTP78wsqEK5MKk+FEkoCBfcu46bIj+expkwgBtjW2cOXvZvPp38/m3nmrqWu0Vb7UW7xbR5KyiooCnzv9AA4bO5B//sMcauqauPv5ldz9/ErKS4o4cdI+nDF5OKcePIyhleVJlyvlLcOJJG3nlIOG8edPn8B3/voyDy6opr6plYbmVu57eQ33vbyGEODwcYM4ffJwTp88nP2G9rPjrNSNbMLWRTZhkwpTXWMLjy5cyz9eWsP986vZsLXxLccMrSznsLEDmT5uINPHDuSQsQOpLPffflIbm7BJUjfqU1bMGVNGcMaUEbS0RmYt3cg/XlrD3+etZvH6bQCsq21oH1UBCAEOHF7VHlgOHTuQiftUUlLsMj+pKxw56SJHTiR1FGNkUXUtjy5cx5xlm5i9bCPLNtTt8PiK0iImj+zPtNEDmDp6ANPGDDCwqGA4ciJJvSCEwKThVUwaXtW+b+2WBuYs28ScZRuZvXQTzy/bxNbsXT71Ta3MWrqJWUs3tR/fFlimjBpARWkR9U2t1De10NCc+Vrf/Mb3JUWBaaMHMH1c5snKYwb1cZ2L8pYjJ13kyImk3dXSGllYvYUXltcwd0UNLyyv4eVVmzt9AOHuGlpZnlnjkg0rh4wZQN+yEuqbWtiwtZENWxvZuK3xjddbG9m4rYm6ppZM8GlqpaH5jdeZMNRC/4pS3nPYaM4/fAyD+5V1w1WQdn/kxHDSRYYTSd2huaWVhdW1vLgiE1heXFHDgtVbiDEzklJRWkx5SfZraTEVJUWUlxZTW9/E3JWbadxBsCkuCpQVF1HX1D39WMqKizhr2gguPno8R+47yFEa7RWndSQpxUqKizh4ZH8OHtmfC48Yu1vnNja38tKqzczOPll59tKNLN+YWefS0hqpa91xMAkBBvQppW9p8RvBp7SIipLi9lBUUVrM88s38drarTS2tHLnnJXcOWclk4ZVcvHR4zhv+hgG9C3dq/9+qSscOekiR04kpVH15npmL9vEnGWbaGxuZXC/Mgb3K2NQ37Ls61IG9ytnQJ9Siot2PfoRY+TJ1zbwu6eXcs/cVTS1vPF3RHlJEeccOorzZozmyH0HU+piXnWR0zo9xHAiqdCsq23g9ueW87unlrJ0w7Y3vVdVUcLbDtiH0w4axskHDnN9inbKcNJDDCeSClVra2Tmq+v47ZNL+cfLa2hpffPfGyHA9LEDOe3g4Zxy4DAOHlnlGhW9iWtOJEndqqgocOKkfThx0j5s2NrIw69Uc//L1Tz8ylq21DcTI+23Sf/g3gUM71/OfkP7MXJAH0YMqGDkgApG9K9o/35IvzKKujDFpOTEGBMNmIYTSVKXDe5Xxnunj+G908fQ1NLKc0s28uD8au6fX82i6loA1mxuYM3mhh1+RmlxYMSACkYP7MPogX0ZM6gPowf1YczAzNeRA/pQVvLm9SwxRppb45v6wLS0Rso7LOgtLylyxGY3NTS3sHjdNhZWb2HhmloWra1l0ZpaJg6v5KcXz0isLsOJJGmPlBYXccz+Qzhm/yH861kHs3T9Nh6Yv4ZnlmxkdU09q2vqWbO5nubtpoGaWiLLNtRlO+pueMvnhgDDqsopCuFNYaR1F6sQQqD9NuyKkmL6lBVTVVHCoL5lDMkuFB5c2fa6vH3xcMkORnHack5JURGVFSX0LS3u9RGfrQ3NrNxUx4pNdayuqaexJXMreYyZwBazrwE6Xp5Apv7M19D+mhBYU1OfCSPVtSxZv+0t03QATS1734tnbxhOJEndYtyQvlx2/H5cdvx+7ftaWiPraxtYVVPPqpp6VtfUsaqmnhXZv3BXbKyjesubR1liZKcjLzsSI9mGcq1A097+57xFCFBZVkJlRQmV5SX0Ky+hKvu6sjyzv6q8hKqK0vZj2vb1KSumpTXS2NxKY0srTS2Z100tma2huZVN2xpZuam+/bqsrKlj07bu/+/YmTGD+jBxWCUHjqhKdGrHcCJJ6jHFRYFh/SsY1r+CQ3fQ1qWhuYVVm+pZvrGOFZu2sWJjJsBkRkIy0zadfS0uCh1a/bdQ3/hGy//6phbqmlrZXNfU3iV3/daGbHDZMzHCloZmtjQ07/FnpEFRgPFD+jFxWCWThlVmv1YxYVg/+palIxakowpJUsEqLylm36H92Hdovx7/Wdsam1lf2/imFv9tsxod717tONHR1NLK1oZmauszwaS2vpnahje2LfXNbM1+rd3L4NKvrJjRg/owamBmG53dRg3sw8gBFfQpK6ZtLCOE0GH6Jrs3ZIqPxMzUD2+e/olE+leUUlFavFd19jTDiSSpYPQtK6Hv4BLGDu7bI5/f2hrZ2pgNLtkws6W+mbrGZkqKiigrKaK0OPO1rLjt+0BpcRH9K0rp36fERb0YTiRJ6jZFRYGqilKqKkphQNLV5C57D0uSpFRJdTgJIZSHEL4XQlgZQqgLITwVQji9i+eODiHcGkLYFELYHEK4K4Swf0/XLEmS9k6qwwnwK+DzwG+BzwItwF9DCCfs7KQQQiXwIHAS8B3gG8B04OEQwpCeLFiSJO2d1K45CSEcBXwAuCrG+MPsvl8Dc4HvA8ft5PQrgEnAUTHGZ7Ln/i177r8AX+nB0iVJ0l5I88jJBWRGSm5o2xFjrAduBI4NIezgjvn2c59pCybZc+cD9wMX9ky5kiSpO6R25ITMNMwrMcbN2+1/Ovv1MGDZ9ieFEIqAQ4BfdvKZTwNnhBCqYoxbdvSDQwjDgH222z2hq4VLkqQ9l+ZwMhJY1cn+tn2jdnDeYKC8C+cu2MnPvoLMOhVJktTL0hxO+gCdPVyhvsP7OzqPPTy3zXXAbdvtmwDctYvzJEnSXkpzOKkjMwKyvYoO7+/oPPbwXABijNVAdcd9duyTJKl3pHlB7CoyUzvba9u3cgfnbSAzarIn50qSpISlOZzMAQ4IIfTfbv/RHd5/ixhjK/AicEQnbx8NvLazxbCSJClZaQ4ntwPFwCfadoQQyoGPAE/FGJdl940LIRzUyblHhhCO6HDugcCpvHUtiSRJSpHUrjmJMT4VQrgN+G721t5FwKXAvsA/dTj012Q6wXZcFHId8HHgLyGEHwJNZDrNrgF+1PPVS5KkPZXacJJ1CfAt4MPAIOAF4OwY4yM7OynGuCWEcDLwX8C/kRkhegj4XIxxbU8WLEmS9k6qw0m2I+xV2W1Hx5y8g/3Lgff1TGWSJKmnpHnNiSRJKkCGE0mSlCqpntZJmTKARYsWJV2HJEk5pcPfnWVdOT7EGHuumjwSQng3tq+XJGlvnBtj/NOuDjKcdFEIYQCZW5aXAY3d9LFtz+s5F3i1mz5TXtee4nXtOV7bnuF17Rl7cl3LgLHAwzHGml0d7LROF2Uv5i7T3u7o8LyeV2OM87rzswuZ17VneF17jte2Z3hde8ZeXNfZXT3QBbGSJClVDCeSJClVDCeSJClVDCfJWgtcnf2q7uN17Rle157jte0ZXtee0ePX1bt1JElSqjhyIkmSUsVwIkmSUsVwIkmSUsVwIkmSUsVwIkmSUsVwkoAQQnkI4XshhJUhhLoQwlMhhNOTritXhBAqQwhXhxDuCSFsCCHEEMJlOzj24Oxxtdljbwkh7NPLJeeEEMKRIYRrQwjzQghbQwhLQwi3hhAO6ORYr2sXhRCmhBBuCyG8FkLYFkJYF0J4JIRwTifHel33Qgjhq9nfB3M7ee+4EMJj2T+D1SGEn4QQKpOoM+1CCCdnr2Nn2zHbHdsj19Vn6yTjV8AFwH8DC4HLgL+GEE6JMT6WYF25YijwdWAp8DxwcmcHhRDGAI8ANcBXgErgC8C0EMJRMcbueoBjvvgScDxwG/ACMAK4EpgVQjgmxjgXvK57YDxQBdwMrAT6AucDfwohXB5jvAG8rnsre/2+Amzt5L3DgPuBl4HPA2PIXNtJwDt7scxc8xPgme32LWp70aPXNcbo1osbcBQQgS902FeR/QN/POn6cmEDyoER2ddHZK/nZZ0cdx2wDRjXYd/bs8d/Iun/jrRtwHFA2Xb7JgH1wG+8rt16rYuBOcB8r2u3XdP/l/2L8iFg7nbv/ZVMMOzfYd/Hstf2jKRrT9tG5h98EbhgF8f12HV1Wqf3XQC0ADe07Ygx1gM3AseGEMYmVViuiDE2xBhXd+HQ84E/xxiXdjj3PuAV4MKeqi9XxRgfj9v96zzGuBCYBxzcYbfXdS/FGFuAZcDADru9rnsohPA2Mr9b/7mT9/oDp5MJ2Js7vPVroBav7U6FEKpCCG+ZZenp62o46X3TgVe2+8MEeDr79bBericvhRBGA8OAZzt5+2kyfw7ahZB5NvpwYF32e6/rHgoh9AshDA0hTAghfI7MsPf92fe8rnsohFAMXAP8Isb4YieHTCOzhOFN1zYbxOfgtd2Zm4DNQH0I4cEQwhEd3uvR6+qak943EljVyf62faN6sZZ8NjL7dUfXenAIoTzG2NCLNeWiDwKjyazxAa/r3vgRcHn2dSvwRzJresDrujc+SWZdz9t38P6uru2JPVFUjmsE7iAzbbMOmExmLcmjIYTjYoyz6eHrajjpfX2Azn7B1Hd4X3uv7Tru6lr7y34HQggHAT8FniCzmBO8rnvjv4HbyfwD5EIy607Ksu95XfdACGEI8H+Bb8UYd/QQul1dW3/nbifG+DjweIddfwoh3E5mofx3gXfQw9fVaZ3eV0dmQef2Kjq8r73Xdh291nsghDAC+AuZO0cuyK6RAK/rHosxzo8x3hdj/HWM8Wwyd+PcnZ0687rumW8DG8hM6+zIrq6t17ULYoyLgLuAU7JTaf+/vfsPtbuu4zj+fEU6y+YMzNoWOMKtRYtCyDWwfzJchYlCWuCP5T+DyPzZEPpDyv4INOyHhjTYP+U/meFCaJVDNk1lI5jI0jJpM9ssbb8iM7P27o/P99Dx7Izu3e6956s+H/DlnPM573vv5/vm3HPe9/P9fD53VvPqyMnce442TD5qMES2dw778no2GGpcOOa5hcB+h8jHS7IA2ESbrPnRqhp+TZrXmXMP8H1gGeZ12pIsBdbSJsEuajUe0D4YT0iyhDZf4v/l1vfcqXuWNtp3MrOcV0dO5t5jwLJupvOwlUPP6zhV1R7gBdpS41FnY57HSnIScB/tA/P8qnpi+HnzOqMGw94LzOsxWUz7DPsusGvoWEl7/e6izZXaCfybkdwmOZG2AMHcTt17aJds/s4s59XiZO7dQ7vWvHbQkGQecCWwraqenVTHXod+Apw/vDw7ybm0N64fT6xXPdUN1f4IWAVcXFWPHiXUvE5DktPHtJ0AXEEb+h4UgOZ1enYCF405fkPboPEiYENVHQI2A5clmT/09ZfTLq2Z2xHjdiVO8kHgAuCXVXV4tvOabtMUzaEkd9N+cb5F23xtDe2vo3Or6sFJ9u21IslVtMsOi4Av0FY+7Oievr2qDnVv8juAg8B3aL8w64A/AR92mPzVknwbuIY2cnL36PNVdVcXZ16nIcm9wCm03V/30HbevRRYDtxQVbd1ceZ1BiTZApxWVSuG2s6iTfB8grbH1LuBG4AHq2r1JPrZZ0keoBXOjwDP01bras3uDAAAA+lJREFUrAVeAVZV1ZNd3OzlddI70b0RD9o10Vtp1+z+SdvHYPWk+/VaOoDdtF0Ixx1LhuLeD/yCtqX1AeAu4J2T7n8fD9rOmkfLaY3Emtep5/VzwP3An7s39/3d4wvGxJrX48/3FkZ2iO3azwEepn3oPg/cAcyfdH/7eABXA9uAfd1rdi/wQ+DMucqrIyeSJKlXnHMiSZJ6xeJEkiT1isWJJEnqFYsTSZLUKxYnkiSpVyxOJElSr1icSJKkXrE4kSRJvWJxIkmSesXiRJKAJJ9PUknG/WdgSXPI4kTSnBkqAI52fGTSfZQ0eW+edAckvSHdBOwa0/70XHdEUv9YnEiahE1V9etJd0JSP3lZR1KvJFnSXeL5cpLrkjyT5KUkW5OsGBP/sSQPJXkxycEkP03yvjFxi5NsSLI3yctJdiW5M8mJI6HzktyW5IXue96b5B2zdsKSjuDIiaRJWJDktJG2qqp9Q4+vAOYD3wNOAq4BHkjygar6C0CSjwObgD8AXwXeAnwJeDjJWVW1u4tbBGwHTgXWA78FFgOfAd4K/Gvo594OHAC+BiwBrgXuAD47A+ctaQosTiRNwuYxbS/TipCBM4GlVbUHIMnPgW3AjcD1XcytwH5gVVXt7+I2AjtoxcWaLu4bwLuAlSOXk25KkpF+7APOq6rqvt+bgKuTLKiqQ8dyspKmx+JE0iR8EXhqpO0/I483DgoTgKranmQb8Cng+iQLgQ8BtwwKky7u8ST3d3GD4uJC4L5x81wGRciQ9SNtDwHXAWcAj0/jHCUdI4sTSZOwfQoTYn8/pu0p4JLu/hnd7e/GxD0JrE5yMvA24BRg5xT79seRxwe627dP8eslHScnxErSq42O4AyMXv6RNEscOZHUV0vHtC0Ddnf3n+lu3zsmbjnw16p6MclLwN+AI1b6SOonR04k9dWFSRYPHiQ5G1hJW51DVT0HPAasSXLqUNwK4DzgZ13cYWAj8OlxW9OPmRAracIcOZE0CZ9MsnxM+yPA4e7+08CvktwJzKMt6d0H3DIUv45WrDyaZAP/W0p8iLa0eOArtIJla5L1tDkpC4GLgXOAgzNzWpJmgsWJpEm4+SjtVwJbuvs/oBUq1wKn0/YpuaobMQGgqjYn+QRt2fDNwCvAVuDGqto1FLcnyUrg68CltAmye2iFzT9m7rQkzYQcuYpOkiYnyRLa/91ZV1XfnGxvJE2Cc04kSVKvWJxIkqResTiRJEm94pwTSZLUK46cSJKkXrE4kSRJvWJxIkmSesXiRJIk9YrFiSRJ6hWLE0mS1CsWJ5IkqVcsTiRJUq9YnEiSpF6xOJEkSb1icSJJknrlv+v9N99XGVIdAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 600x600 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(5,5))\n",
    "plt.plot(history.history['loss'])\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('Loss')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "BzsEF_Cw3ji_"
   },
   "outputs": [],
   "source": [
    "def plot_attention(attention, sentence, predicted_sentence):\n",
    "  fig = plt.figure(figsize=(10,10))\n",
    "  ax = fig.add_subplot(1, 1, 1)\n",
    "  ax.matshow(attention, cmap='viridis')\n",
    "\n",
    "  fontdict = {'fontsize': 14}\n",
    "\n",
    "  ax.set_xticklabels([''] + sentence, fontdict=fontdict, rotation=90)\n",
    "  ax.set_yticklabels([''] + predicted_sentence, fontdict=fontdict, fontproperties=zhfont)\n",
    "\n",
    "  ax.xaxis.set_major_locator(ticker.MultipleLocator(1))\n",
    "  ax.yaxis.set_major_locator(ticker.MultipleLocator(1))\n",
    "\n",
    "  plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "APgut8LW3kGt"
   },
   "outputs": [],
   "source": [
    "def translate(sentence):\n",
    "  predicted_sentences, attention_weights = model.predict([sentence])\n",
    "  sentence = ['<s>'] + sentence + ['</s>']\n",
    "  result = predicted_sentences[0] + ['</s>']\n",
    "  attention_plot = attention_weights[0]\n",
    "\n",
    "  print('Input:', ' '.join(sentence))\n",
    "  print('Predicted translation: {}'.format(' '.join(result)))\n",
    "\n",
    "  attention_plot = attention_plot[:len(result), 1:len(sentence)+1]\n",
    "  plot_attention(attention_plot, sentence, result)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 857
    },
    "colab_type": "code",
    "id": "mPjIwROgKbA4",
    "outputId": "3dc83d36-34a5-4e5d-f0df-6fdfcdb48b5f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input: <s> she isn't afraid of death . </s>\n",
      "Predicted translation: 她 不 怕 死 。 </s>\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x1200 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "translate(\"she isn't afraid of death .\".split(' '))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "id": "_D0W-fz8tBLv",
    "outputId": "7c2842e9-9152-4409-9ff4-9ca4e220b8ea"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input: <s> that young lady is a nurse . </s>\n",
      "Predicted translation: 这 个 年 轻 女 孩 是 护 士 。 </s>\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x1200 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "translate(\"that young lady is a nurse .\".split(' '))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "id": "4dYug4sBKbA7",
    "outputId": "eddb43b4-3b6e-4a84-a54c-99e51ce99ab6"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input: <s> don't bug me . </s>\n",
      "Predicted translation: 别 来 烦 我 。 </s>\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x1200 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input: <s> tom never wanted to try it . </s>\n",
      "Predicted translation: 汤 姆 从 不 想 试 试 。 </s>\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x1200 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input: <s> what's wrong , honey ? </s>\n",
      "Predicted translation: 出 什 么 事 了 ， 宝 贝 ？ </s>\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x1200 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input: <s> i don't see much of him . </s>\n",
      "Predicted translation: 我 不 常 见 到 他 。 </s>\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x1200 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input: <s> i bathe every day . </s>\n",
      "Predicted translation: 我 每 天 都 洗 澡 。 </s>\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x1200 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import random\n",
    "\n",
    "for sentence in random.sample(list(input_datas), 5):\n",
    "    translate(sentence)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Waq890xbub-0"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "kashgari_seq2seq_nmt.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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}