NMT-Training-Inference.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Neural Machine Translation\n",
    "\n",
    "Neural machine translation is an approach to machine translation that uses a large artificial neural network to predict the likelihood of a sequence of words, typically modeling entire sentences in a single integrated model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 16556,
     "status": "ok",
     "timestamp": 1561270089395,
     "user": {
      "displayName": "MANDAV",
      "photoUrl": "https://lh4.googleusercontent.com/-Zi5_ERBmNoM/AAAAAAAAAAI/AAAAAAAAA4U/x7o5V3Av1zw/s64/photo.jpg",
      "userId": "06372901974541954018"
     },
     "user_tz": -330
    },
    "id": "ESfnmA5SBUwI",
    "outputId": "0335a13d-b9dd-45ff-c745-73e560484374"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import string\n",
    "from string import digits\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "import re\n",
    "import os\n",
    "from sklearn.utils import shuffle\n",
    "from sklearn.model_selection import train_test_split\n",
    "from keras.layers import Input, LSTM, Embedding, Dense, Bidirectional\n",
    "from keras.layers import Activation, dot, concatenate\n",
    "from keras.models import Model\n",
    "from keras.callbacks import TensorBoard, ModelCheckpoint, ReduceLROnPlateau, EarlyStopping\n",
    "from keras.preprocessing.text import Tokenizer\n",
    "import tensorflow as tf\n",
    "from gensim.models import Word2Vec\n",
    "from tqdm import tqdm_notebook\n",
    "import nltk\n",
    "%matplotlib inline  \n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dataset\n",
    "\n",
    "English - Spanish translation dataset is downloaded from [Tab-delimited Bilingual Sentence Pairs](http://www.manythings.org/anki/).\n",
    "\n",
    "First a CSV file generated using `preprocess.py` for easy management and access.\n",
    "\n",
    "During training, dataset is divided randomly in 7:3 ratio.\n",
    "\n",
    "Dataset from Tab-delimited Bilingual Sentence Pairs is in the format of text file with language1-lagnuage2 seperated by `\\t` (tab). For better management, I have opt to create a `CSV file`. `preprocess.py` takes input the `.txt` (`spa.txt`) and cleans the file by converting non-printable characters to printable format and generates the CSV file.\n",
    "\n",
    "```\n",
    "usage: preprocess.py [-h] --dataset DATASET [--language_1 LANGUAGE_1]\n",
    "                     --language_2 LANGUAGE_2 [--save_file SAVE_FILE]\n",
    "\n",
    "optional arguments:\n",
    "  -h, --help            show this help message and exit\n",
    "  --dataset DATASET, -d DATASET\n",
    "                        Path to .txt file downloaded from\n",
    "                        http://www.manythings.org/anki/\n",
    "  --language_1 LANGUAGE_1, -l1 LANGUAGE_1\n",
    "                        Language-1 name | Default: english\n",
    "  --language_2 LANGUAGE_2, -l2 LANGUAGE_2\n",
    "                        Language-2 name\n",
    "  --save_file SAVE_FILE, -s SAVE_FILE\n",
    "                        Name of CSV file to be generated | Default:\n",
    "                        dataset.csv\n",
    "\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1141,
     "status": "ok",
     "timestamp": 1561288409366,
     "user": {
      "displayName": "MANDAV",
      "photoUrl": "https://lh4.googleusercontent.com/-Zi5_ERBmNoM/AAAAAAAAAAI/AAAAAAAAA4U/x7o5V3Av1zw/s64/photo.jpg",
      "userId": "06372901974541954018"
     },
     "user_tz": -330
    },
    "id": "zRJwXTOGBUwO",
    "outputId": "0814ce91-24f6-44c1-abd3-8796dde7d304"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(120614, 3)"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lines = pd.read_csv('dataset/english-spanish-dataset.csv')\n",
    "lines.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "af5iSs6WBUwV"
   },
   "outputs": [],
   "source": [
    "lines.english = lines.english.apply(lambda x: x.lower())\n",
    "lines.spanish = lines.spanish.apply(lambda x: str(x).lower())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ZoJ3JuzqBUwY"
   },
   "outputs": [],
   "source": [
    "exclude = set(string.punctuation)\n",
    "lines.english = lines.english.apply(lambda x: ''.join(ch for ch in x if ch not in exclude))\n",
    "lines.spanish = lines.spanish.apply(lambda x: ''.join(ch for ch in x if ch not in exclude))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "jxomWPoYBUwb"
   },
   "outputs": [],
   "source": [
    "lines.english = lines.english.apply(lambda x: x.strip())\n",
    "lines.spanish = lines.spanish.apply(lambda x: x.strip())\n",
    "\n",
    "lines.english = lines.english.apply(lambda x: re.sub(\" +\", \" \", x))\n",
    "lines.spanish = lines.spanish.apply(lambda x: re.sub(\" +\", \" \", x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "KKfPrO8JBUwe"
   },
   "outputs": [],
   "source": [
    "lines.english = lines.english.apply(lambda x: re.sub(\"\\?\\?\", '', x))\n",
    "lines.spanish = lines.spanish.apply(lambda x: re.sub(\"\\?\\?\", '', x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "KIcRESM0BUwi"
   },
   "outputs": [],
   "source": [
    "lines.spanish = lines.spanish.apply(lambda x : 'START_ '+ x + '_END')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2383,
     "status": "ok",
     "timestamp": 1561288411651,
     "user": {
      "displayName": "MANDAV",
      "photoUrl": "https://lh4.googleusercontent.com/-Zi5_ERBmNoM/AAAAAAAAAAI/AAAAAAAAA4U/x7o5V3Av1zw/s64/photo.jpg",
      "userId": "06372901974541954018"
     },
     "user_tz": -330
    },
    "id": "AIlTJmu5BUwl",
    "outputId": "10eca49b-a4d6-4f59-dd58-8dd3620feee1"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['Unnamed: 0', 'english', 'spanish'], dtype='object')"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lines.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "-VCepbFNBUwp"
   },
   "outputs": [],
   "source": [
    "del lines['Unnamed: 0']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 359
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2143,
     "status": "ok",
     "timestamp": 1561288411655,
     "user": {
      "displayName": "MANDAV",
      "photoUrl": "https://lh4.googleusercontent.com/-Zi5_ERBmNoM/AAAAAAAAAAI/AAAAAAAAA4U/x7o5V3Av1zw/s64/photo.jpg",
      "userId": "06372901974541954018"
     },
     "user_tz": -330
    },
    "id": "LdGaY2u7BUws",
    "outputId": "4283cfae-3bdb-4806-bf38-49a9cc3a258d"
   },
   "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>english</th>\n",
       "      <th>spanish</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>8867</th>\n",
       "      <td>please go ahead</td>\n",
       "      <td>START_ continua por favor_END</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>112653</th>\n",
       "      <td>around here they roll up the sidewalks at oclock</td>\n",
       "      <td>START_ a las siete de la tarde las calles esta...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59117</th>\n",
       "      <td>that seems to be the problem</td>\n",
       "      <td>START_ ese parece ser el problema_END</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>74883</th>\n",
       "      <td>sometimes history repeats itself</td>\n",
       "      <td>START_ a veces se repite la historia_END</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31308</th>\n",
       "      <td>dont beat yourself up</td>\n",
       "      <td>START_ no te maltrates a ti mismo_END</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>80294</th>\n",
       "      <td>he is nothing but a minor composer</td>\n",
       "      <td>START_ el no es mas que un compositor de segun...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>78561</th>\n",
       "      <td>the airplane is ready for landing</td>\n",
       "      <td>START_ el avion esta listo para aterrizar_END</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>80391</th>\n",
       "      <td>hes not good at remembering names</td>\n",
       "      <td>START_ el no es bueno para recordar nombres_END</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68454</th>\n",
       "      <td>who told you to say that to me</td>\n",
       "      <td>START_ quien te dijo que me dijeras eso_END</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49157</th>\n",
       "      <td>i dont remember my dreams</td>\n",
       "      <td>START_ no recuerdo mis suenos_END</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                 english  \\\n",
       "8867                                     please go ahead   \n",
       "112653  around here they roll up the sidewalks at oclock   \n",
       "59117                       that seems to be the problem   \n",
       "74883                   sometimes history repeats itself   \n",
       "31308                              dont beat yourself up   \n",
       "80294                 he is nothing but a minor composer   \n",
       "78561                  the airplane is ready for landing   \n",
       "80391                  hes not good at remembering names   \n",
       "68454                     who told you to say that to me   \n",
       "49157                          i dont remember my dreams   \n",
       "\n",
       "                                                  spanish  \n",
       "8867                        START_ continua por favor_END  \n",
       "112653  START_ a las siete de la tarde las calles esta...  \n",
       "59117               START_ ese parece ser el problema_END  \n",
       "74883            START_ a veces se repite la historia_END  \n",
       "31308               START_ no te maltrates a ti mismo_END  \n",
       "80294   START_ el no es mas que un compositor de segun...  \n",
       "78561       START_ el avion esta listo para aterrizar_END  \n",
       "80391     START_ el no es bueno para recordar nombres_END  \n",
       "68454         START_ quien te dijo que me dijeras eso_END  \n",
       "49157                   START_ no recuerdo mis suenos_END  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lines.sample(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "b7RIXJS_BUww"
   },
   "outputs": [],
   "source": [
    "vocab_english = set()\n",
    "for sent in lines.english:\n",
    "    for word in sent.split():\n",
    "        if word not in vocab_english:\n",
    "            vocab_english.add(word)\n",
    "\n",
    "vocab_spanish = set()\n",
    "for sent in lines.spanish:\n",
    "    for word in sent.split():\n",
    "        if word not in vocab_spanish:\n",
    "            vocab_spanish.add(word)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2390,
     "status": "ok",
     "timestamp": 1561288412235,
     "user": {
      "displayName": "MANDAV",
      "photoUrl": "https://lh4.googleusercontent.com/-Zi5_ERBmNoM/AAAAAAAAAAI/AAAAAAAAA4U/x7o5V3Av1zw/s64/photo.jpg",
      "userId": "06372901974541954018"
     },
     "user_tz": -330
    },
    "id": "zjUfamiMBUw3",
    "outputId": "98b9edf4-af84-44fc-fbbb-3c88aea04379"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(54, 7.071260384366657)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "length_list=[]\n",
    "for l in lines.spanish:\n",
    "    length_list.append(len(l.split(' ')))\n",
    "max_length_spanish = np.max(length_list)\n",
    "max_length_spanish, np.average(length_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2272,
     "status": "ok",
     "timestamp": 1561288412236,
     "user": {
      "displayName": "MANDAV",
      "photoUrl": "https://lh4.googleusercontent.com/-Zi5_ERBmNoM/AAAAAAAAAAI/AAAAAAAAA4U/x7o5V3Av1zw/s64/photo.jpg",
      "userId": "06372901974541954018"
     },
     "user_tz": -330
    },
    "id": "ZS0g00QRBUw7",
    "outputId": "7358168d-4e4e-4ba6-fc04-7d84b5cd3a63"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(47, 6.299807650853134)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "length_list=[]\n",
    "for l in lines.english:\n",
    "    length_list.append(len(l.split(' ')))\n",
    "max_length_english = np.max(length_list)\n",
    "max_length_english, np.average(length_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2111,
     "status": "ok",
     "timestamp": 1561288412237,
     "user": {
      "displayName": "MANDAV",
      "photoUrl": "https://lh4.googleusercontent.com/-Zi5_ERBmNoM/AAAAAAAAAAI/AAAAAAAAA4U/x7o5V3Av1zw/s64/photo.jpg",
      "userId": "06372901974541954018"
     },
     "user_tz": -330
    },
    "id": "-0g--tTeBUw_",
    "outputId": "3b2caa3b-9059-4877-ef6f-567e42679d8f"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(13392, 33770)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "input_words = sorted(list(vocab_english))\n",
    "target_words = sorted(list(vocab_spanish))\n",
    "num_encoder_tokens = len(vocab_english)\n",
    "num_decoder_tokens = len(vocab_spanish)\n",
    "num_encoder_tokens, num_decoder_tokens"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1916,
     "status": "ok",
     "timestamp": 1561288412239,
     "user": {
      "displayName": "MANDAV",
      "photoUrl": "https://lh4.googleusercontent.com/-Zi5_ERBmNoM/AAAAAAAAAAI/AAAAAAAAA4U/x7o5V3Av1zw/s64/photo.jpg",
      "userId": "06372901974541954018"
     },
     "user_tz": -330
    },
    "id": "gB_EeRdABUxE",
    "outputId": "53aaed85-229e-47c2-e647-3dc573e3649b"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "33771"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "num_decoder_tokens += 1 # For zero padding\n",
    "num_decoder_tokens"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "5HAw_r07BUxJ"
   },
   "outputs": [],
   "source": [
    "input_token_index = dict([(word, i+1) for i, word in enumerate(input_words)])\n",
    "target_token_index = dict([(word, i+1) for i, word in enumerate(target_words)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ammnby9GBUxM"
   },
   "outputs": [],
   "source": [
    "reverse_input_char_index = dict((i, word) for word, i in input_token_index.items())\n",
    "reverse_target_char_index = dict((i, word) for word, i in target_token_index.items())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 359
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1823,
     "status": "ok",
     "timestamp": 1561288412628,
     "user": {
      "displayName": "MANDAV",
      "photoUrl": "https://lh4.googleusercontent.com/-Zi5_ERBmNoM/AAAAAAAAAAI/AAAAAAAAA4U/x7o5V3Av1zw/s64/photo.jpg",
      "userId": "06372901974541954018"
     },
     "user_tz": -330
    },
    "id": "9nKVhnXtBUxQ",
    "outputId": "4ed7ba96-eab5-4951-9e9f-540e390035c3"
   },
   "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>english</th>\n",
       "      <th>spanish</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>19983</th>\n",
       "      <td>hes a ghost hunter</td>\n",
       "      <td>START_ es un cazafantasmas_END</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28814</th>\n",
       "      <td>it was very beautiful</td>\n",
       "      <td>START_ fue muy bonito_END</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51267</th>\n",
       "      <td>tom is writing a novel now</td>\n",
       "      <td>START_ ahora tom esta escribiendo una novela_END</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>66276</th>\n",
       "      <td>i woke up at five this morning</td>\n",
       "      <td>START_ me desperte a las esta manana_END</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50289</th>\n",
       "      <td>rome wasnt built in a day</td>\n",
       "      <td>START_ no se gano zamora en una hora_END</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44127</th>\n",
       "      <td>do you have a better idea</td>\n",
       "      <td>START_ tienes una idea mejor_END</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86849</th>\n",
       "      <td>i used to listen to english programs</td>\n",
       "      <td>START_ solia escuchar programas en ingles_END</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>55179</th>\n",
       "      <td>this furnitures in the way</td>\n",
       "      <td>START_ este mueble estorba el paso_END</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>119609</th>\n",
       "      <td>in the old days every child could tell the dif...</td>\n",
       "      <td>START_ antiguamente todos los ninos sabian cua...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>65008</th>\n",
       "      <td>dont stay up late every night</td>\n",
       "      <td>START_ no te quedes despierto hasta tarde toda...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  english  \\\n",
       "19983                                  hes a ghost hunter   \n",
       "28814                               it was very beautiful   \n",
       "51267                          tom is writing a novel now   \n",
       "66276                      i woke up at five this morning   \n",
       "50289                           rome wasnt built in a day   \n",
       "44127                           do you have a better idea   \n",
       "86849                i used to listen to english programs   \n",
       "55179                          this furnitures in the way   \n",
       "119609  in the old days every child could tell the dif...   \n",
       "65008                       dont stay up late every night   \n",
       "\n",
       "                                                  spanish  \n",
       "19983                      START_ es un cazafantasmas_END  \n",
       "28814                           START_ fue muy bonito_END  \n",
       "51267    START_ ahora tom esta escribiendo una novela_END  \n",
       "66276            START_ me desperte a las esta manana_END  \n",
       "50289            START_ no se gano zamora en una hora_END  \n",
       "44127                    START_ tienes una idea mejor_END  \n",
       "86849       START_ solia escuchar programas en ingles_END  \n",
       "55179              START_ este mueble estorba el paso_END  \n",
       "119609  START_ antiguamente todos los ninos sabian cua...  \n",
       "65008   START_ no te quedes despierto hasta tarde toda...  "
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lines = shuffle(lines)\n",
    "lines.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2161,
     "status": "ok",
     "timestamp": 1561288413181,
     "user": {
      "displayName": "MANDAV",
      "photoUrl": "https://lh4.googleusercontent.com/-Zi5_ERBmNoM/AAAAAAAAAAI/AAAAAAAAA4U/x7o5V3Av1zw/s64/photo.jpg",
      "userId": "06372901974541954018"
     },
     "user_tz": -330
    },
    "id": "-UPKSoPGBUxU",
    "outputId": "f5165d13-a685-4d37-bf1e-85c0be16bc7b"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(120614, 120614)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "english = list(lines.english)\n",
    "spanish = list(lines.spanish)\n",
    "\n",
    "x = []\n",
    "y = []\n",
    "\n",
    "for i in range(len(english)):\n",
    "    x.append(str(english[i]))\n",
    "    y.append(str(spanish[i]))\n",
    "\n",
    "x = np.array(x)\n",
    "y = np.array(y)\n",
    "len(x), len(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1989,
     "status": "ok",
     "timestamp": 1561288413185,
     "user": {
      "displayName": "MANDAV",
      "photoUrl": "https://lh4.googleusercontent.com/-Zi5_ERBmNoM/AAAAAAAAAAI/AAAAAAAAA4U/x7o5V3Av1zw/s64/photo.jpg",
      "userId": "06372901974541954018"
     },
     "user_tz": -330
    },
    "id": "5IBLLHNjBUxX",
    "outputId": "d5492792-eea8-41ad-90d1-7feb27d6a78a"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((108552,), (12062,))"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(x, y, test_size = 0.1)\n",
    "X_train.shape, X_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "FoCDohIIBUxb"
   },
   "outputs": [],
   "source": [
    "def generate_batch(X = X_train, y = y_train, batch_size = 128):\n",
    "    ''' Generate a batch of data '''\n",
    "    while True:\n",
    "        for j in range(0, len(X), batch_size):\n",
    "            encoder_input_data = np.zeros((batch_size, max_length_english),dtype='float32')\n",
    "            decoder_input_data = np.zeros((batch_size, max_length_spanish),dtype='float32')\n",
    "            decoder_target_data = np.zeros((batch_size, max_length_spanish, num_decoder_tokens),dtype='float32')\n",
    "            for i, (input_text, target_text) in enumerate(zip(X[j:j+batch_size], y[j:j+batch_size])):\n",
    "                for t, word in enumerate(input_text.split()):\n",
    "                    encoder_input_data[i, t] = input_token_index[word] # encoder input seq\n",
    "                for t, word in enumerate(target_text.split()):\n",
    "                    if t<len(target_text.split())-1:\n",
    "                        decoder_input_data[i, t] = target_token_index[word] # decoder input seq\n",
    "                    if t>0:\n",
    "                        # decoder target sequence (one hot encoded)\n",
    "                        # does not include the START_ token\n",
    "                        # Offset by one timestep\n",
    "                        decoder_target_data[i, t - 1, target_token_index[word]] = 1.\n",
    "            yield([encoder_input_data, decoder_input_data], decoder_target_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Embeddings\n",
    "\n",
    "The embedding layer of the network is initialized first with pre-trained embedding trained using `Word2Vec` (skipgram model) on `english-spanish-dataset.csv` to improve the overall performance of the network.\n",
    "\n",
    "You can find trained embedding under \"RELEASES\" section.\n",
    "\n",
    "You can generate word2vec embeddings using generate_embeddings.py on your own dataset.\n",
    "\n",
    "```\n",
    "usage: generate_embeddings.py [-h] --csv CSV --language LANGUAGE [--sg]\n",
    "                              [-emb_size EMB_SIZE] [--save_file SAVE_FILE]\n",
    "\n",
    "optional arguments:\n",
    "  -h, --help            show this help message and exit\n",
    "  --csv CSV             Path to CSV which is to be read\n",
    "  --language LANGUAGE, -l LANGUAGE\n",
    "                        Language whose embedding is to be generated from CSV\n",
    "                        file. (It should be the column name in CSV)\n",
    "  --sg                  Whether to use SkipGram or CBoW model | Default: CBoW\n",
    "  -emb_size EMB_SIZE, -s EMB_SIZE\n",
    "                        Size of embedding to generate | Defualt: 256\n",
    "  --save_file SAVE_FILE\n",
    "                        Name of embedding file to save | Default:\n",
    "                        [skipgram/cbow]-[language]-[embedding_size].model\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "t38FLk4JBUxd"
   },
   "outputs": [],
   "source": [
    "latent_dim = 256"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 105
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1610,
     "status": "ok",
     "timestamp": 1561288413966,
     "user": {
      "displayName": "MANDAV",
      "photoUrl": "https://lh4.googleusercontent.com/-Zi5_ERBmNoM/AAAAAAAAAAI/AAAAAAAAA4U/x7o5V3Av1zw/s64/photo.jpg",
      "userId": "06372901974541954018"
     },
     "user_tz": -330
    },
    "id": "wuP9BP4xBUxg",
    "outputId": "50c51b44-1380-4741-d79d-5e262af9d05d"
   },
   "outputs": [],
   "source": [
    "english_embedding = Word2Vec.load('embeddings/skipgram-english-256.model')\n",
    "spanish_embedding = Word2Vec.load('embeddings/skipgram-spanish-256.model')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "dN3HieZuBUxi"
   },
   "outputs": [],
   "source": [
    "eng_tok = Tokenizer()\n",
    "ger_tok = Tokenizer()\n",
    "\n",
    "eng_tok.fit_on_texts(english)\n",
    "ger_tok.fit_on_texts(spanish)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 122
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 5034,
     "status": "ok",
     "timestamp": 1561288418110,
     "user": {
      "displayName": "MANDAV",
      "photoUrl": "https://lh4.googleusercontent.com/-Zi5_ERBmNoM/AAAAAAAAAAI/AAAAAAAAA4U/x7o5V3Av1zw/s64/photo.jpg",
      "userId": "06372901974541954018"
     },
     "user_tz": -330
    },
    "id": "iAaqCdvcBUxk",
    "outputId": "b1eadc7b-9b63-444e-c711-88b2c062306e"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:5: DeprecationWarning: Call to deprecated `__getitem__` (Method will be removed in 4.0.0, use self.wv.__getitem__() instead).\n",
      "  \"\"\"\n",
      "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:14: DeprecationWarning: Call to deprecated `__getitem__` (Method will be removed in 4.0.0, use self.wv.__getitem__() instead).\n",
      "  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "index 13392 is out of bounds for axis 0 with size 13392\n"
     ]
    }
   ],
   "source": [
    "# create a weight matrix for words in training docs\n",
    "encoder_embedding_matrix = np.zeros((num_encoder_tokens, latent_dim))\n",
    "for word, i in eng_tok.word_index.items():\n",
    "    try:\n",
    "        embedding_vector = english_embedding[word]\n",
    "        if embedding_vector is not None:\n",
    "            encoder_embedding_matrix[i] = embedding_vector\n",
    "    except Exception as e:\n",
    "        print(e)\n",
    "# create a weight matrix for words in training docs\n",
    "decoder_embedding_matrix = np.zeros((num_decoder_tokens, latent_dim))\n",
    "for word, i in ger_tok.word_index.items():\n",
    "    try:\n",
    "        embedding_vector = spanish_embedding[word]\n",
    "        if embedding_vector is not None:\n",
    "            decoder_embedding_matrix[i] = embedding_vector\n",
    "    except Exception as e:\n",
    "        pass"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LSTM-Attention Model\n",
    "\n",
    "Here a simple model with Luong's Attention is used to achieve the task of language translation although more complex models can be built to improve the model's performance. You can read more about basics and backgrounds of NMTs at [tensorflow/nmt](https://github.com/tensorflow/nmt#basic).\n",
    "\n",
    "I attempt to implement Luong's attention due to its simplicity in understanding as well as implementation.\n",
    "\n",
    "This diagram explains the basic architecture of the model [Source: background-on-the-attention-mechanism]:\n",
    "\n",
    "<p align=\"center\"> <img src=\"results/attention_mechanism.jpg\"/> </p>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 68
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 8662,
     "status": "ok",
     "timestamp": 1561288422093,
     "user": {
      "displayName": "MANDAV",
      "photoUrl": "https://lh4.googleusercontent.com/-Zi5_ERBmNoM/AAAAAAAAAAI/AAAAAAAAA4U/x7o5V3Av1zw/s64/photo.jpg",
      "userId": "06372901974541954018"
     },
     "user_tz": -330
    },
    "id": "OMedXeGwBUxo",
    "outputId": "e92706a3-2db3-406a-988b-9d5b3f0c0c6a"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING: Logging before flag parsing goes to stderr.\n",
      "W0624 21:46:40.619225 140523438131008 deprecation.py:506] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:1188: calling reduce_sum_v1 (from tensorflow.python.ops.math_ops) with keep_dims is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "keep_dims is deprecated, use keepdims instead\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "encoder_outputs:  Tensor(\"lstm_1/transpose_2:0\", shape=(?, ?, 1024), dtype=float32)\n",
      "state_h:  Tensor(\"lstm_1/while/Exit_2:0\", shape=(?, 1024), dtype=float32)\n",
      "state_c:  Tensor(\"lstm_1/while/Exit_3:0\", shape=(?, 1024), dtype=float32)\n"
     ]
    }
   ],
   "source": [
    "encoder_inputs = Input(shape=(None,))\n",
    "enc_emb =  Embedding(num_encoder_tokens, latent_dim, mask_zero = True, weights=[encoder_embedding_matrix])(encoder_inputs)\n",
    "\n",
    "encoder_lstm = LSTM(latent_dim*4, return_sequences=True, return_state=True)\n",
    "encoder_outputs, state_h, state_c = encoder_lstm(enc_emb)\n",
    "encoder_states = [state_h, state_c]\n",
    "\n",
    "print('encoder_outputs: ', encoder_outputs)\n",
    "print('state_h: ', state_h)\n",
    "print('state_c: ', state_c)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "pEitjyBaBUxr"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "W0624 21:46:43.505066 140523438131008 deprecation.py:506] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:1154: calling reduce_max_v1 (from tensorflow.python.ops.math_ops) with keep_dims is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "keep_dims is deprecated, use keepdims instead\n",
      "W0624 21:46:43.515182 140523438131008 deprecation.py:506] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:1320: calling reduce_all_v1 (from tensorflow.python.ops.math_ops) with keep_dims is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "keep_dims is deprecated, use keepdims instead\n"
     ]
    }
   ],
   "source": [
    "decoder_inputs = Input(shape=(None,))\n",
    "dec_emb_layer = Embedding(num_decoder_tokens, latent_dim, mask_zero = True, weights=[decoder_embedding_matrix])\n",
    "dec_emb = dec_emb_layer(decoder_inputs)\n",
    "\n",
    "decoder_lstm = LSTM(latent_dim*4, return_sequences=True, return_state=True)\n",
    "decoder_outputs, _, _ = decoder_lstm(dec_emb,\n",
    "                                     initial_state=encoder_states)\n",
    "\n",
    "attention = dot([decoder_outputs, encoder_outputs], axes=[2, 2])\n",
    "\n",
    "attention = Activation('softmax', name='attention')(attention)\n",
    "\n",
    "context = dot([attention, encoder_outputs], axes=[2, 1], name='context')\n",
    "\n",
    "decoder_combined_context = concatenate([context, decoder_outputs], name='decoder_combined_context')\n",
    "\n",
    "decoder_dense = Dense(num_decoder_tokens, activation='softmax')\n",
    "decoder_outputs = decoder_dense(decoder_combined_context)\n",
    "\n",
    "model = Model([encoder_inputs, decoder_inputs], decoder_outputs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 615
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 11754,
     "status": "ok",
     "timestamp": 1561288426423,
     "user": {
      "displayName": "MANDAV",
      "photoUrl": "https://lh4.googleusercontent.com/-Zi5_ERBmNoM/AAAAAAAAAAI/AAAAAAAAA4U/x7o5V3Av1zw/s64/photo.jpg",
      "userId": "06372901974541954018"
     },
     "user_tz": -330
    },
    "id": "b0yAfcepBUxu",
    "outputId": "1119cad9-2636-4ab0-9d83-65a780dfbc95"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "____________________________________________________________________________________________________\n",
      "Layer (type)                     Output Shape          Param #     Connected to                     \n",
      "====================================================================================================\n",
      "input_1 (InputLayer)             (None, None)          0                                            \n",
      "____________________________________________________________________________________________________\n",
      "input_2 (InputLayer)             (None, None)          0                                            \n",
      "____________________________________________________________________________________________________\n",
      "embedding_1 (Embedding)          (None, None, 256)     3428352     input_1[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "embedding_2 (Embedding)          (None, None, 256)     8645376     input_2[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "lstm_1 (LSTM)                    [(None, None, 1024),  5246976     embedding_1[0][0]                \n",
      "____________________________________________________________________________________________________\n",
      "lstm_2 (LSTM)                    [(None, None, 1024),  5246976     embedding_2[0][0]                \n",
      "                                                                   lstm_1[0][1]                     \n",
      "                                                                   lstm_1[0][2]                     \n",
      "____________________________________________________________________________________________________\n",
      "dot_1 (Dot)                      (None, None, None)    0           lstm_2[0][0]                     \n",
      "                                                                   lstm_1[0][0]                     \n",
      "____________________________________________________________________________________________________\n",
      "attention (Activation)           (None, None, None)    0           dot_1[0][0]                      \n",
      "____________________________________________________________________________________________________\n",
      "context (Dot)                    (None, None, 1024)    0           attention[0][0]                  \n",
      "                                                                   lstm_1[0][0]                     \n",
      "____________________________________________________________________________________________________\n",
      "decoder_combined_context (Concat (None, None, 2048)    0           context[0][0]                    \n",
      "                                                                   lstm_2[0][0]                     \n",
      "____________________________________________________________________________________________________\n",
      "dense_1 (Dense)                  (None, None, 33771)   69196779    decoder_combined_context[0][0]   \n",
      "====================================================================================================\n",
      "Total params: 91,764,459\n",
      "Trainable params: 91,764,459\n",
      "Non-trainable params: 0\n",
      "____________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['acc'])\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "BPxZJJWrBUxx"
   },
   "outputs": [],
   "source": [
    "train_samples = len(X_train)\n",
    "val_samples = len(X_test)\n",
    "batch_size = 16\n",
    "epochs = 20"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training NMT Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "OLyEW_50BUx0"
   },
   "outputs": [],
   "source": [
    "class TrainValTensorBoard(TensorBoard):\n",
    "    def __init__(self, log_dir='./logs', **kwargs):\n",
    "        # Make the original `TensorBoard` log to a subdirectory 'training'\n",
    "        training_log_dir = os.path.join(log_dir, 'training')\n",
    "        super(TrainValTensorBoard, self).__init__(training_log_dir, **kwargs)\n",
    "\n",
    "        # Log the validation metrics to a separate subdirectory\n",
    "        self.val_log_dir = os.path.join(log_dir, 'validation')\n",
    "\n",
    "    def set_model(self, model):\n",
    "        # Setup writer for validation metrics\n",
    "        self.val_writer = tf.summary.FileWriter(self.val_log_dir)\n",
    "        super(TrainValTensorBoard, self).set_model(model)\n",
    "\n",
    "    def on_epoch_end(self, epoch, logs=None):\n",
    "        # Pop the validation logs and handle them separately with\n",
    "        # `self.val_writer`. Also rename the keys so that they can\n",
    "        # be plotted on the same figure with the training metrics\n",
    "        logs = logs or {}\n",
    "        val_logs = {k.replace('val_', ''): v for k, v in logs.items() if k.startswith('val_')}\n",
    "        for name, value in val_logs.items():\n",
    "            summary = tf.Summary()\n",
    "            summary_value = summary.value.add()\n",
    "            summary_value.simple_value = value.item()\n",
    "            summary_value.tag = name\n",
    "            self.val_writer.add_summary(summary, epoch)\n",
    "        self.val_writer.flush()\n",
    "\n",
    "        # Pass the remaining logs to `TensorBoard.on_epoch_end`\n",
    "        logs = {k: v for k, v in logs.items() if not k.startswith('val_')}\n",
    "        super(TrainValTensorBoard, self).on_epoch_end(epoch, logs)\n",
    "\n",
    "    def on_train_end(self, logs=None):\n",
    "        super(TrainValTensorBoard, self).on_train_end(logs)\n",
    "        self.val_writer.close()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "0BMAEEF3BUx2"
   },
   "outputs": [],
   "source": [
    "log_dir = 'eng-spa-weights'\n",
    "logging = TrainValTensorBoard(log_dir=log_dir)\n",
    "checkpoint = ModelCheckpoint(os.path.join(log_dir, 'ep{epoch:03d}-val_loss{val_loss:.3f}-val_acc{val_acc:.3f}.h5'),\n",
    "        monitor='val_loss', save_weights_only=True, save_best_only=True, period=3)\n",
    "reduce_lr = ReduceLROnPlateau(monitor='val_loss', factor=0.1, patience=3, verbose=1)\n",
    "early_stopping = EarlyStopping(monitor='val_loss', min_delta=0, patience=5, verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Dg_xZW8V8lPz"
   },
   "outputs": [],
   "source": [
    "# To retrain first load the trained model!\n",
    "#model.load_weights('eng-spa-weights/ep015-val_loss3.445-val_acc0.592.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 238
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1093473,
     "status": "ok",
     "timestamp": 1561303223221,
     "user": {
      "displayName": "MANDAV",
      "photoUrl": "https://lh4.googleusercontent.com/-Zi5_ERBmNoM/AAAAAAAAAAI/AAAAAAAAA4U/x7o5V3Av1zw/s64/photo.jpg",
      "userId": "06372901974541954018"
     },
     "user_tz": -330
    },
    "id": "zXoQi4TRBUx6",
    "outputId": "4544145f-94dc-45e2-8bd9-4ece51a1ba87"
   },
   "outputs": [],
   "source": [
    "# model.fit_generator(generator = generate_batch(X_train, y_train, batch_size = batch_size),\n",
    "#                     steps_per_epoch = train_samples//batch_size,\n",
    "#                     epochs=epochs,\n",
    "#                     validation_data = generate_batch(X_test, y_test, batch_size = batch_size),\n",
    "#                     validation_steps = val_samples//batch_size,\n",
    "#                     callbacks=[logging, checkpoint, reduce_lr, early_stopping])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "_cX-8VhmS46L"
   },
   "outputs": [],
   "source": [
    "# Save trained Model\n",
    "#model.save_weights('eng-spa-weights/final_weights.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Inferencing NMT Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "xSeN6RUYBUx8"
   },
   "outputs": [],
   "source": [
    "# Loading the model\n",
    "model.load_weights('eng-spa-weights/final_weights.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "PJE-x_8PBUx9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Attention:  Tensor(\"dot_2/MatMul:0\", shape=(?, ?, ?), dtype=float32)\n",
      "Softmax:  Tensor(\"attention_1/truediv:0\", shape=(?, ?, ?), dtype=float32)\n",
      "Context:  Tensor(\"dot_3/MatMul:0\", shape=(?, ?, 1024), dtype=float32)\n",
      "Combined Context:  Tensor(\"concatenate_1/concat:0\", shape=(?, ?, 2048), dtype=float32)\n"
     ]
    }
   ],
   "source": [
    "# Encode the input sequence to get the \"thought vectors\"\n",
    "encoder_model = Model(encoder_inputs, [encoder_outputs] + encoder_states)\n",
    "\n",
    "# Decoder setup\n",
    "# Below tensors will hold the states of the previous time step\n",
    "decoder_state_input_h = Input(shape=(latent_dim*4,))\n",
    "decoder_state_input_c = Input(shape=(latent_dim*4,))\n",
    "\n",
    "encoder_inf_input = Input(shape=(None, latent_dim*4))\n",
    "\n",
    "decoder_states_inputs = [decoder_state_input_h, decoder_state_input_c]\n",
    "\n",
    "dec_emb2= dec_emb_layer(decoder_inputs) # Get the embeddings of the decoder sequence\n",
    "\n",
    "# To predict the next word in the sequence, set the initial states to the states from the previous time step\n",
    "decoder_outputs2, state_h2, state_c2 = decoder_lstm(dec_emb2, initial_state=decoder_states_inputs)\n",
    "decoder_states2 = [state_h2, state_c2]\n",
    "\n",
    "\n",
    "attention = dot([decoder_outputs2, encoder_inf_input], axes=[2, 2])\n",
    "print('Attention: ', attention)\n",
    "\n",
    "attention = Activation('softmax', name='attention')(attention)\n",
    "print('Softmax: ', attention)\n",
    "\n",
    "context = dot([attention, encoder_inf_input], axes=[2, 1])\n",
    "print('Context: ', context)\n",
    "\n",
    "decoder_combined_context = concatenate([context, decoder_outputs2])\n",
    "print('Combined Context: ', decoder_combined_context)\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "decoder_outputs2 = decoder_dense(decoder_combined_context) # A dense softmax layer to generate prob dist. over the target vocabulary\n",
    "\n",
    "# Final decoder model\n",
    "decoder_model = Model(\n",
    "    [decoder_inputs, encoder_inf_input] + decoder_states_inputs,\n",
    "    [decoder_outputs2] + decoder_states2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "9JyYBhHmBUyA"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         (None, None)              0         \n",
      "_________________________________________________________________\n",
      "embedding_1 (Embedding)      (None, None, 256)         3428352   \n",
      "_________________________________________________________________\n",
      "lstm_1 (LSTM)                [(None, None, 1024), (Non 5246976   \n",
      "=================================================================\n",
      "Total params: 8,675,328\n",
      "Trainable params: 8,675,328\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[<tf.Tensor 'lstm_1/transpose_2:0' shape=(?, ?, 1024) dtype=float32>,\n",
       " <tf.Tensor 'lstm_1/while/Exit_2:0' shape=(?, 1024) dtype=float32>,\n",
       " <tf.Tensor 'lstm_1/while/Exit_3:0' shape=(?, 1024) dtype=float32>]"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "encoder_model.summary()\n",
    "encoder_model.output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "OSYAXdkeBUyC"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "____________________________________________________________________________________________________\n",
      "Layer (type)                     Output Shape          Param #     Connected to                     \n",
      "====================================================================================================\n",
      "input_2 (InputLayer)             (None, None)          0                                            \n",
      "____________________________________________________________________________________________________\n",
      "embedding_2 (Embedding)          (None, None, 256)     8645376     input_2[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "input_3 (InputLayer)             (None, 1024)          0                                            \n",
      "____________________________________________________________________________________________________\n",
      "input_4 (InputLayer)             (None, 1024)          0                                            \n",
      "____________________________________________________________________________________________________\n",
      "lstm_2 (LSTM)                    [(None, None, 1024),  5246976     embedding_2[1][0]                \n",
      "                                                                   input_3[0][0]                    \n",
      "                                                                   input_4[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "input_5 (InputLayer)             (None, None, 1024)    0                                            \n",
      "____________________________________________________________________________________________________\n",
      "dot_2 (Dot)                      (None, None, None)    0           lstm_2[1][0]                     \n",
      "                                                                   input_5[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "attention (Activation)           (None, None, None)    0           dot_2[0][0]                      \n",
      "____________________________________________________________________________________________________\n",
      "dot_3 (Dot)                      (None, None, 1024)    0           attention[0][0]                  \n",
      "                                                                   input_5[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "concatenate_1 (Concatenate)      (None, None, 2048)    0           dot_3[0][0]                      \n",
      "                                                                   lstm_2[1][0]                     \n",
      "____________________________________________________________________________________________________\n",
      "dense_1 (Dense)                  (None, None, 33771)   69196779    concatenate_1[0][0]              \n",
      "====================================================================================================\n",
      "Total params: 83,089,131\n",
      "Trainable params: 83,089,131\n",
      "Non-trainable params: 0\n",
      "____________________________________________________________________________________________________\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "([<tf.Tensor 'input_2:0' shape=(?, ?) dtype=float32>,\n",
       "  <tf.Tensor 'input_5:0' shape=(?, ?, 1024) dtype=float32>,\n",
       "  <tf.Tensor 'input_3:0' shape=(?, 1024) dtype=float32>,\n",
       "  <tf.Tensor 'input_4:0' shape=(?, 1024) dtype=float32>],\n",
       " [<tf.Tensor 'dense_1_1/truediv:0' shape=(?, ?, 33771) dtype=float32>,\n",
       "  <tf.Tensor 'lstm_2_1/while/Exit_2:0' shape=(?, 1024) dtype=float32>,\n",
       "  <tf.Tensor 'lstm_2_1/while/Exit_3:0' shape=(?, 1024) dtype=float32>])"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "decoder_model.summary()\n",
    "decoder_model.input, decoder_model.output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "JZifboeTBUyE"
   },
   "outputs": [],
   "source": [
    "def decode_sequence(input_seq):\n",
    "    # Encode the input as state vectors.\n",
    "    [encoder_output, h, c] = encoder_model.predict(input_seq)\n",
    "    states_value = [h, c]\n",
    "    \n",
    "    # Generate empty target sequence of length 1.\n",
    "    target_seq = np.zeros((1,1))\n",
    "    # Populate the first character of target sequence with the start character.\n",
    "    target_seq[0, 0] = target_token_index['START_']\n",
    "\n",
    "    # Sampling loop for a batch of sequences\n",
    "    # (to simplify, here we assume a batch of size 1).\n",
    "    \n",
    "    #return\n",
    "    stop_condition = False\n",
    "    decoded_sentence = ''\n",
    "    while not stop_condition:\n",
    "        output_tokens, h, c = decoder_model.predict([target_seq, encoder_output] + states_value)\n",
    "\n",
    "        # Sample a token\n",
    "        sampled_token_index = np.argmax(output_tokens[0, -1, :])\n",
    "        sampled_char = reverse_target_char_index[sampled_token_index]\n",
    "        decoded_sentence += ' '+sampled_char\n",
    "\n",
    "        # Exit condition: either hit max length\n",
    "        # or find stop character.\n",
    "        if (sampled_char == '_END' or\n",
    "           len(decoded_sentence) > 100):\n",
    "            stop_condition = True\n",
    "\n",
    "        # Update the target sequence (of length 1).\n",
    "        target_seq = np.zeros((1,1))\n",
    "        target_seq[0, 0] = sampled_token_index\n",
    "\n",
    "        # Update states\n",
    "        states_value = [h, c]\n",
    "\n",
    "    return decoded_sentence\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "sU-TkMATBUyI"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<keras.engine.topology.InputLayer object at 0x7fcd902813c8>\tname:input_2\n",
      "<keras.layers.embeddings.Embedding object at 0x7fcd90281390>\tname:embedding_2\n",
      "<keras.engine.topology.InputLayer object at 0x7fcd70676c50>\tname:input_3\n",
      "<keras.engine.topology.InputLayer object at 0x7fcd70676d30>\tname:input_4\n",
      "<keras.layers.recurrent.LSTM object at 0x7fcd90281588>\tname:lstm_2\n",
      "<keras.engine.topology.InputLayer object at 0x7fcd70676b38>\tname:input_5\n",
      "<keras.layers.merge.Dot object at 0x7fcd70677048>\tname:dot_2\n",
      "<keras.layers.core.Activation object at 0x7fcd7051ae10>\tname:attention\n",
      "<keras.layers.merge.Dot object at 0x7fcd70676e10>\tname:dot_3\n",
      "<keras.layers.merge.Concatenate object at 0x7fcd703e24a8>\tname:concatenate_1\n",
      "<keras.layers.core.Dense object at 0x7fcd90286b38>\tname:dense_1\n"
     ]
    }
   ],
   "source": [
    "layers = decoder_model.layers\n",
    "for l in layers:\n",
    "    print('%s\\tname:%s' % (str(l), l.name))\n",
    "    \n",
    "assert(model.layers[7] == model.get_layer('attention'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "W1Djk_K4BUyJ"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<keras.engine.training.Model object at 0x7fcd70676a90>\n",
      "[(None, None, 33771), (None, 1024), (None, 1024), (None, None, None)] [(None, None), (None, None, 1024), (None, 1024), (None, 1024)]\n",
      "____________________________________________________________________________________________________\n",
      "Layer (type)                     Output Shape          Param #     Connected to                     \n",
      "====================================================================================================\n",
      "input_2 (InputLayer)             (None, None)          0                                            \n",
      "____________________________________________________________________________________________________\n",
      "embedding_2 (Embedding)          (None, None, 256)     8645376     input_2[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "input_3 (InputLayer)             (None, 1024)          0                                            \n",
      "____________________________________________________________________________________________________\n",
      "input_4 (InputLayer)             (None, 1024)          0                                            \n",
      "____________________________________________________________________________________________________\n",
      "lstm_2 (LSTM)                    [(None, None, 1024),  5246976     embedding_2[1][0]                \n",
      "                                                                   input_3[0][0]                    \n",
      "                                                                   input_4[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "input_5 (InputLayer)             (None, None, 1024)    0                                            \n",
      "____________________________________________________________________________________________________\n",
      "dot_2 (Dot)                      (None, None, None)    0           lstm_2[1][0]                     \n",
      "                                                                   input_5[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "attention (Activation)           (None, None, None)    0           dot_2[0][0]                      \n",
      "____________________________________________________________________________________________________\n",
      "dot_3 (Dot)                      (None, None, 1024)    0           attention[0][0]                  \n",
      "                                                                   input_5[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "concatenate_1 (Concatenate)      (None, None, 2048)    0           dot_3[0][0]                      \n",
      "                                                                   lstm_2[1][0]                     \n",
      "____________________________________________________________________________________________________\n",
      "dense_1 (Dense)                  (None, None, 33771)   69196779    concatenate_1[0][0]              \n",
      "====================================================================================================\n",
      "Total params: 83,089,131\n",
      "Trainable params: 83,089,131\n",
      "Non-trainable params: 0\n",
      "____________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "attention_layer = decoder_model.get_layer('attention') # or model.layers[7]\n",
    "attention_model = Model(inputs=decoder_model.inputs, outputs=decoder_model.outputs + [attention_layer.output])\n",
    "\n",
    "print(attention_model)\n",
    "print(attention_model.output_shape, attention_model.input_shape)\n",
    "attention_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "7Xvl-jxrBUyL"
   },
   "outputs": [],
   "source": [
    "def attent_and_generate(input_seq):\n",
    "    decoded_sentence = []\n",
    "    \n",
    "    [encoder_output, h, c] = encoder_model.predict(input_seq)\n",
    "    states_value = [h, c]\n",
    "    \n",
    "    target_seq = np.zeros((1,1))\n",
    "    target_seq[0, 0] = target_token_index['START_']\n",
    "    \n",
    "    stop_condition = False\n",
    "    attention_density = []\n",
    "    index = []\n",
    "    \n",
    "    while not stop_condition:\n",
    "        output_tokens, h, c, attention = attention_model.predict([target_seq, encoder_output] + states_value)\n",
    "        sampled_token_index = np.argmax(output_tokens[0, -1, :])\n",
    "        sampled_char = reverse_target_char_index[sampled_token_index]\n",
    "        decoded_sentence.append(sampled_char)\n",
    "    \n",
    "        if (sampled_char.endswith('_END')) or len(decoded_sentence) > 50:\n",
    "            stop_condition = True\n",
    "            \n",
    "        states_value = [h, c]\n",
    "        target_seq = np.zeros((1,1))\n",
    "        target_seq[0, 0] = sampled_token_index\n",
    "        attention_density.append((sampled_char, attention[0][0]))\n",
    "      \n",
    "    return np.array(attention_density), ' '.join(decoded_sentence)\n",
    "\n",
    "\n",
    "def visualize(text, encoder_input):\n",
    "    attention_weights, decoded_sent = attent_and_generate(encoder_input)\n",
    "    \n",
    "    plt.clf()\n",
    "    plt.figure(figsize=(10,10))\n",
    "    \n",
    "    mats = []\n",
    "    dec_inputs = []\n",
    "    for dec_ind, attn in attention_weights:\n",
    "        mats.append(attn[:len(text[0].split(' '))].reshape(-1))\n",
    "        dec_inputs.append(dec_ind)\n",
    "        \n",
    "    attention_mat = np.transpose(np.array(mats))\n",
    "    fig, ax = plt.subplots(figsize=(8, 8))\n",
    "    ax.imshow(attention_mat)\n",
    "    ax.set_xticks(np.arange(attention_mat.shape[1]))\n",
    "    ax.set_yticks(np.arange(attention_mat.shape[0]))\n",
    "\n",
    "    ax.set_xticklabels([inp for inp in dec_inputs])\n",
    "    ax.set_yticklabels([w for w in str(text[0]).split(' ')])\n",
    "\n",
    "    ax.tick_params(labelsize=15)\n",
    "    ax.tick_params(axis='x', labelrotation=90)\n",
    "\n",
    "    plt.show()\n",
    "    return decoded_sent"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Some Samples generated from the trained model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "8fUASyoHBUyN"
   },
   "outputs": [],
   "source": [
    "train_gen = generate_batch(X_train, y_train, batch_size = 1)\n",
    "k=-1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "6fwqJfEwBUyP"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['cars made in japan are used all over the world']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input English Sentence: cars made in japan are used all over the world\n",
      "Actual spanish:  en todo el mundo se usan autos fabricados en japon\n",
      "Predicted spanish: en todo el mundo se los autos de autos en japon\n"
     ]
    }
   ],
   "source": [
    "k+=1\n",
    "(encoder_input, actual_output), _ = next(train_gen)\n",
    "print(X_train[k:k+1])\n",
    "decoded_sent = visualize(X_train[k:k+1], encoder_input)\n",
    "\n",
    "# Translation of Prediction from Google Translator: all over the world are the cars in japan\n",
    "print('Input English Sentence:', X_train[k:k+1][0])\n",
    "print('Actual spanish:', y_train[k:k+1][0][6:-4])\n",
    "print('Predicted spanish:', decoded_sent[:-4])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Translation of Predicted Spanish (Source: Google Translator): \"all over the world are the cars in japan\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Cy-0X-xs7tZw"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['i did everything for you']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input English Sentence: i did everything for you\n",
      "Actual spanish:  te lo di todo\n",
      "Predicted spanish: lo hice todo por ti\n"
     ]
    }
   ],
   "source": [
    "k+=1\n",
    "(encoder_input, actual_output), _ = next(train_gen)\n",
    "print(X_train[k:k+1])\n",
    "decoded_sent = visualize(X_train[k:k+1], encoder_input)\n",
    "\n",
    "# Translation of Prediction from Google Translator: I did everything for you\n",
    "print('Input English Sentence:', X_train[k:k+1][0])\n",
    "print('Actual spanish:', y_train[k:k+1][0][6:-4])\n",
    "print('Predicted spanish:', decoded_sent[:-4])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Translation of Predicted Spanish (Source: Google Translator): \"I did everything for you\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['tom and mary are not talking to one another']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input English Sentence: tom and mary are not talking to one another\n",
      "Actual spanish:  tom y mary ya no se hablan\n",
      "Predicted spanish: tom y mary ya no se hablan uno al otro\n"
     ]
    }
   ],
   "source": [
    "k+=1\n",
    "(encoder_input, actual_output), _ = next(train_gen)\n",
    "print(X_train[k:k+1])\n",
    "decoded_sent = visualize(X_train[k:k+1], encoder_input)\n",
    "\n",
    "# Translation of Prediction from Google Translator: Tom and Mary no longer talk to each other\n",
    "print('Input English Sentence:', X_train[k:k+1][0])\n",
    "print('Actual spanish:', y_train[k:k+1][0][6:-4])\n",
    "print('Predicted spanish:', decoded_sent[:-4])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Translation of Predicted Spanish (Source: Google Translator): \"Tom and Mary no longer talk to each other\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['tom is the happiest person in the world right now']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input English Sentence: tom is the happiest person in the world right now\n",
      "Actual spanish:  ahora mismo tom es la persona mas feliz del mundo\n",
      "Predicted spanish: tom es la persona mas feliz del mundo en el mundo ahora\n"
     ]
    }
   ],
   "source": [
    "k+=1\n",
    "(encoder_input, actual_output), _ = next(train_gen)\n",
    "print(X_train[k:k+1])\n",
    "decoded_sent = visualize(X_train[k:k+1], encoder_input)\n",
    "\n",
    "# Translation of Prediction from Google Translator: Tom is the happiest person in the world in the world now\n",
    "print('Input English Sentence:', X_train[k:k+1][0])\n",
    "print('Actual spanish:', y_train[k:k+1][0][6:-4])\n",
    "print('Predicted spanish:', decoded_sent[:-4])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Translation of Predicted Spanish (Source: Google Translator): \"Tom is the happiest person in the world in the world now\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['ive been taking dancing lessons']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input English Sentence: ive been taking dancing lessons\n",
      "Actual spanish:  he estado tomando clases de baile\n",
      "Predicted spanish: he estado tomando clases de manzana\n"
     ]
    }
   ],
   "source": [
    "k+=1\n",
    "(encoder_input, actual_output), _ = next(train_gen)\n",
    "print(X_train[k:k+1])\n",
    "decoded_sent = visualize(X_train[k:k+1], encoder_input)\n",
    "\n",
    "# Translation of Prediction from Google Translator: I've been taking apple classes\n",
    "print('Input English Sentence:', X_train[k:k+1][0])\n",
    "print('Actual spanish:', y_train[k:k+1][0][6:-4])\n",
    "print('Predicted spanish:', decoded_sent[:-4])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Translation of Predicted Spanish (Source: Google Translator): \"I've been taking apple classes\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['the dog is beautiful']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input English Sentence: the dog is beautiful\n",
      "Actual spanish:  el perro es hermoso\n",
      "Predicted spanish: el perro es hermoso\n"
     ]
    }
   ],
   "source": [
    "k+=1\n",
    "(encoder_input, actual_output), _ = next(train_gen)\n",
    "print(X_train[k:k+1])\n",
    "decoded_sent = visualize(X_train[k:k+1], encoder_input)\n",
    "\n",
    "# Translation of Prediction from Google Translator: the dog is beautiful\n",
    "print('Input English Sentence:', X_train[k:k+1][0])\n",
    "print('Actual spanish:', y_train[k:k+1][0][6:-4])\n",
    "print('Predicted spanish:', decoded_sent[:-4])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Translation of Predicted Spanish (Source: Google Translator): \"the dog is beautiful\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['tom is able to walk on his hands']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input English Sentence: tom is able to walk on his hands\n",
      "Actual spanish:  tom es capaz de andar con las manos\n",
      "Predicted spanish: tom es capaz de andar sobre las manos\n"
     ]
    }
   ],
   "source": [
    "k+=1\n",
    "(encoder_input, actual_output), _ = next(train_gen)\n",
    "print(X_train[k:k+1])\n",
    "decoded_sent = visualize(X_train[k:k+1], encoder_input)\n",
    "\n",
    "# Translation of Prediction from Google Translator: Tom is able to walk on his hands\n",
    "print('Input English Sentence:', X_train[k:k+1][0])\n",
    "print('Actual spanish:', y_train[k:k+1][0][6:-4])\n",
    "print('Predicted spanish:', decoded_sent[:-4])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Translation of Predicted Spanish (Source: Google Translator): \"Tom is able to walk on his hands\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['thats my dress']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input English Sentence: thats my dress\n",
      "Actual spanish:  ese es mi vestido\n",
      "Predicted spanish: ese es mi vestido\n"
     ]
    }
   ],
   "source": [
    "k+=1\n",
    "(encoder_input, actual_output), _ = next(train_gen)\n",
    "print(X_train[k:k+1])\n",
    "decoded_sent = visualize(X_train[k:k+1], encoder_input)\n",
    "\n",
    "# Translation of Prediction from Google Translator: that is my dress\n",
    "print('Input English Sentence:', X_train[k:k+1][0])\n",
    "print('Actual spanish:', y_train[k:k+1][0][6:-4])\n",
    "print('Predicted spanish:', decoded_sent[:-4])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Translation of Predicted Spanish (Source: Google Translator): \"that is my dress\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['mary really takes after her mother']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input English Sentence: mary really takes after her mother\n",
      "Actual spanish:  mary se parece mucho a su madre\n",
      "Predicted spanish: mary se parece mucho a su madre\n"
     ]
    }
   ],
   "source": [
    "k+=1\n",
    "(encoder_input, actual_output), _ = next(train_gen)\n",
    "print(X_train[k:k+1])\n",
    "decoded_sent = visualize(X_train[k:k+1], encoder_input)\n",
    "\n",
    "# Translation of Prediction from Google Translator: Mary looks a lot like her mother\n",
    "print('Input English Sentence:', X_train[k:k+1][0])\n",
    "print('Actual spanish:', y_train[k:k+1][0][6:-4])\n",
    "print('Predicted spanish:', decoded_sent[:-4])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Translation of Predicted Spanish (Source: Google Translator): \"Mary looks a lot like her mother\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['tom asked mary to call him after dinner']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input English Sentence: tom asked mary to call him after dinner\n",
      "Actual spanish:  tom le pidio a maria que lo llamara despues de la cena\n",
      "Predicted spanish: tom le pidio a maria que le llamara despues de la cena\n"
     ]
    }
   ],
   "source": [
    "k+=1\n",
    "(encoder_input, actual_output), _ = next(train_gen)\n",
    "print(X_train[k:k+1])\n",
    "decoded_sent = visualize(X_train[k:k+1], encoder_input)\n",
    "\n",
    "# Translation of Prediction from Google Translator: Tom asked Maria to call him after dinner\n",
    "print('Input English Sentence:', X_train[k:k+1][0])\n",
    "print('Actual spanish:', y_train[k:k+1][0][6:-4])\n",
    "print('Predicted spanish:', decoded_sent[:-4])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Translation of Predicted Spanish (Source: Google Translator): \"Tom asked Maria to call him after dinner\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['i have no money to buy the book with']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input English Sentence: i have no money to buy the book with\n",
      "Actual spanish:  no tengo dinero para comprar ese libro\n",
      "Predicted spanish: no tengo dinero para comprar el libro\n"
     ]
    }
   ],
   "source": [
    "k+=1\n",
    "(encoder_input, actual_output), _ = next(train_gen)\n",
    "print(X_train[k:k+1])\n",
    "decoded_sent = visualize(X_train[k:k+1], encoder_input)\n",
    "\n",
    "# Translation of Prediction from Google Translator: I do not have money to buy the book\n",
    "print('Input English Sentence:', X_train[k:k+1][0])\n",
    "print('Actual spanish:', y_train[k:k+1][0][6:-4])\n",
    "print('Predicted spanish:', decoded_sent[:-4])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Translation of Predicted Spanish (Source: Google Translator): \"I do not have money to buy the book\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['his name is known to everyone in this town']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input English Sentence: his name is known to everyone in this town\n",
      "Actual spanish:  su nombre es conocido por todos en esta ciudad\n",
      "Predicted spanish: su nombre es conocido por todos en esta ciudad\n"
     ]
    }
   ],
   "source": [
    "k+=1\n",
    "(encoder_input, actual_output), _ = next(train_gen)\n",
    "print(X_train[k:k+1])\n",
    "decoded_sent = visualize(X_train[k:k+1], encoder_input)\n",
    "\n",
    "# Translation of Prediction from Google Translator: his name is known by everyone in this city\n",
    "print('Input English Sentence:', X_train[k:k+1][0])\n",
    "print('Actual spanish:', y_train[k:k+1][0][6:-4])\n",
    "print('Predicted spanish:', decoded_sent[:-4])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Translation of Predicted Spanish (Source: Google Translator): \"his name is known by everyone in this city\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['what time is tom coming']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input English Sentence: what time is tom coming\n",
      "Actual spanish:  a que hora viene tomas\n",
      "Predicted spanish: a que hora viene tomas\n"
     ]
    }
   ],
   "source": [
    "k+=1\n",
    "(encoder_input, actual_output), _ = next(train_gen)\n",
    "print(X_train[k:k+1])\n",
    "decoded_sent = visualize(X_train[k:k+1], encoder_input)\n",
    "\n",
    "# Translation of Prediction from Google Translator: what time do you come in\n",
    "print('Input English Sentence:', X_train[k:k+1][0])\n",
    "print('Actual spanish:', y_train[k:k+1][0][6:-4])\n",
    "print('Predicted spanish:', decoded_sent[:-4])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Translation of Predicted Spanish (Source: Google Translator): \"what time do you come in\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['i come home from work about seven every day']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 720x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgkAAAF7CAYAAABYYBivAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzt3XmYXFWZ+PFvJwQCigrDIhBJQGZed1EBYWQVd1EGF3D7ITIyrozjMgiKsrjh7igyLggBBxVRARUXNsOiQlAZBR1ejZBIEgwKsgkkIenfH+c2KYrbSXWtvXw/z9NPV917zz2nbveteuusQ8PDw0iSJDWbNugCSJKk8ckgQZIk1TJIkCRJtQwSJElSLYMESZJUyyBBkiTVMkiQJEm1DBIkSVItgwRJklTLIEGSJNVab9AFGJANgJ2Bm4BVAy6LJEm9Nh3YCrgKWN5qoqkaJOwMXDboQkiS1Gd7AJe3evBUDRJuAthjr/1ZsuSmtk5w/YL5bL/DLl0t1ETI2/zN3/y9981/4uW/zTZbcdkl50L1+deqqRokrAJYsuQmFi1a3PZJOknbqUHmbf7mb/7e++Y/YfMfUxO7HRclSVItgwRJklTLIEGSJNUySJAkSbUMEiRJUi2DBEmSVMsgQZIk1TJIkCRJtSZFkBARcyPiF4MuhyRJk8lkmXHxA8CGgy6EJEmTyaQIEjLzj4MugyRJk43NDZIkqdakCBIkSVL3DQ0PDw+6DB2LiLnAEzJzpxaTzAFu6FmBJEkan7YDFrZ68KTok9Cu7XfYpe1lN1etXMr0GVt3uUTjP2/zN3/z9943/4mX/+zZs7h+wfwxp7O5QZIk1TJIkCRJtQwSJElSLYMESZJUyyBBkiTVmhSjGzLzkEGXQZKkycaaBEmSVMsgQZIk1TJIkCRJtQwSJElSLYMESZJUyyBBkiTVMkiQJEm1DBIkSVItgwRJklTLIEGSJNUySJAkSbUMEiRJUi2DBEmSVMsgQZIk1TJIkCRJtQwSJElSLYMESZJUyyBBkiTVMkiQJEm1DBIkSVItgwRJklTLIEGSJNVar92EEbEncBywM7AKuBp4e2ZeHRE7Ap8EdgOWAz8A3pGZy6q0c4AbgFcCzwVeCtwBHJmZ/xMRRwD/AcwATgGOyszVDXk/AfgosGe16UfA4Zn553ZfjyRJeqC2ahIiYm/gImAl8FrgIOAyYJuI2ByYB2wEvAo4HNgLuCAi1m861UeBmyhBwmXAaRHxSWAX4FDgM8ARwIENee8A/BSYCbwGOAR4PPC9iBhq5/VIkqQHGxoeHh5zooj4OeVb/s6ZOdy07wTgjcC2mXlHte3pwBXAqzLz6w01CXMz83XVMQ8DbgEWAo/JzFXV9vnADZl5UPX8q5Qg4omZuaLa9o/AdcCLM/O8Fl7CSP6SJE0l21E+Z1sy5uaGiHgI8HTgbc0BQmUX4PyRAAEgM6+MiIXA7sDXG469qOGYOyLiL8AlIwFCZQGwbcPzZwGnAasjYqT8N1Be9E5AK0ECANvvsAuLFi1u9fAHWLVyKdNnbN1W2k4NMm/zN3/z9943/4mX/+zZs7h+wfwxp2unuWETYIjSTFBnK2BZzfZlwKZN225rer5ilG0zG55vBryb0tTR+LM98Kh1lF2SJLWonY6LfwNWU4KBOjcBW9Rs3xL4ZRv5NbsVOBs4uWbfX7twfkmSRBtBQmb+PSKuBA6OiBNrmhyuBN4UERtn5p0AEbEzpR/A5Z0WmNJE8Xjgl6M0d0iSpC5odwjkkcCFwA8j4kvA3ynDHX8BfAp4E/DjiPgo8FDgBOAa4NsdlxiOBeYD50XEKZTag22AZ1M6Qs7rQh6SJE15bQ2BzMxLKR/KGwH/A5xJGea4ODP/AuwD3EvppPh5yvDGZ4+MRuhEZv4e2BW4G/gS8EPKfA3LKZ0cJUlSF7Q9mVJmXsKayYya910NPHMtaRdSOj82b59Ts+2Qmm3XAS9rubCSJGnMnJZZkiTVMkiQJEm1DBIkSVItgwRJklTLIEGSJNUySJAkSbUMEiRJUi2DBEmSVKvtyZQmg7zyLFh9X9vp7116Wdtpl73wsLbTAix6ymM6Sj932Wjrc7XmuK336Sj977mno/Sv3nrXjtKfsfSKjtJL0lRgTYIkSaplkCBJkmoZJEiSpFoGCZIkqZZBgiRJqmWQIEmSahkkSJKkWgYJkiSplkGCJEmqZZAgSZJqjSlIiIi5EfGLXhVGkiSNH9YkSJKkWgYJkiSpVlurQEbEs4FPAo8GrgbekJm/rfZtBJwAHAg8ArgGeG9mnt+Qfh7wV+A84P3A5sDZwOuBJwMnAo8DfgEcnJl/akg7EzgeeCWwBXAdcFRm/qCd1yJJkuq1U5OwLfBx4EOs+aA+MyKGqv1fBl5X7T8AuBE4LyJ2bzrPrsBrgcOBIyhBxeeq9P8FvAbYHvhSU7pvAYcAHwZeBFwFfDcidmzjtUiSpFEMDQ8Pt3xwRMylfHg/NjP/UG37F0otwGOBIeC3wOsy87Rq/zTgN8CSzHxutW0esCMwOzNvr7Z9E3g5sFdmXlptezPweeAhmXl3ROwLXAjsnZmXNJTrUmBZZr68xZcyB7ih5RcuSdLksB2wsNWD22luWDgSIFR+V/2eBWxNCRTOGtmZmasj4ixKbUGjX4wECJUFwArg8qZtVOddADwL+DPw04hoLPtFlNqFMVl5yyJYfd9YkwEwY/NHs/Ivf2wrLcCyFx7WdtpZ8y9m8S7PbDs9wNxlW7Wd9uhFZ/DB2a/uKP/fc0/baU9f9B0Onv2SjvI/Y+kVbaddtXIp02ds3VH+nTD/qZv/VH7t5t9Z/rNnz+L6BfPHnK6dIOG2pucrqt8zga2AuzLz7qZjlgEbRcQGmbl8Lee5MzNXj3JugM2ARwIra8q1qsXyS5KkFrTVcXEtbgIeGhEbNQUKWwJ3NwQI7boVWAL8S4fnkSRJ69DtIOEqYBh4GXA6QNWh8WU8sBmhXRcB76TUVlzXhfNJkqRRdDVIyMz/i4ivAydGxMbAH4HDgMcAb+pCFhcAPwYuiIiPUjpJPozSCXJmZh7VhTwkSRK9mUzpMOA0yvwH5wKzgf0ys+OahMwcBl4CnAL8ByVg+CKwG92pqZAkSZUx1SRk5iE12xZSRjSMPL+bMvfB4Ws5z941244Fjm3aNq/x3NW25cAx1Y8kSeoRp2WWJEm1DBIkSVItgwRJklTLIEGSJNUySJAkSbUMEiRJUi2DBEmSVMsgQZIk1er22g0TSjz95SxatLittKtWLmXm1nt0uUQt5g3MvrrTpSvaT380cMzSn3SYf/tOp7OlniVJrbEmQZIk1TJIkCRJtQwSJElSLYMESZJUyyBBkiTVMkiQJEm1DBIkSVItgwRJklTLIEGSJNUySJAkSbW6Oi1zRLwfeAOwFXB6Zh7SzfNLkqT+6VqQEBE7AccB7wHmATd369ySJKn/ulmT8Jjq9+cz8466AyJiw8y8p4t5SpKkHulKkBARc4HXVk9vjwiAfYCfAM8D3gI8EzgT+NeI2Ag4ATgQeARwDfDezDy/4ZzzgL8C5wHvBzYHzgZeDzwZOBF4HPAL4ODM/FM3XoskSSq6VZPwAeBGyirCzwTuAbao9n0FOBX4DHBvte3LwIspTRMLgMOA8yJin8y8vOG8uwKbAYcD2wKfrs79dOBjwN+BzwJfogQjkiSpS4aGh4e7cqKIOIQSDGycmXdFxN6UmoTPZObbG457LPBb4HWZeVq1bRrwG2BJZj632jYP2BGYnZm3V9u+Cbwc2CszL622vRn4PPCQzLy7xeLOAW7o5PVKkjQBbQcsbPXgro5uGMV5Tc93BoaAs0Y2ZObqiDgLOKLp2F+MBAiVBcAK4PKmbQBbNzxuyfY77MKiRYvHkuR+q1YuZfqMrdtK26lB5m3+5m/+3vvmP/Hynz17FtcvmD/mdP2YJ2FZ0/OtgLtqvvUvAzaKiA0att3WdMwK4M7MXN20DWBmxyWVJEn360eQ0NyecRPw0KrzYqMtgbszc3kfyiRJktZhEDMuXkUJHF42siEihqrnl4+WSJIk9Vc/+iQ8QGb+X0R8HTgxIjYG/kgZ3fAY4E39Lo8kSarX9yChchjwUcr8ByPzJOzXNPxRkiQNUNeGQE4wc4AbHN1g/uZv/hMp/6n82s2/a6MbxjQE0lUgJUlSLYMESZJUyyBBkiTVMkiQJEm1DBIkSVItgwRJklTLIEGSJNUa1GRKU97pm+8z0PT7Pad53a2xufXVj+so/XqP27aj9Hd86Hkdpf/ciZ0tEfKhrTq7/m+/+K0dpb/7/77dUfqNHvvSjtJ3YtpQ599NOjnH6uHV6z5IEmBNgiRJGoVBgiRJqmWQIEmSahkkSJKkWgYJkiSplkGCJEmqZZAgSZJqGSRIkqRaBgmSJKmWQYIkSaplkCBJkmpNuCAhIhZGxCcGXQ5Jkia7CRckSJKk/pgwQUJEbDjoMkiSNJV0vFR0ROwDXAxsk5lLq20/B3YB/iEzb6u2XQN8NzPfGxE7Ap8EdgOWAz8A3pGZy6pj5wA3AK8Bngu8GPgF8Kya/LcBLgIWAy/OzLs7fU2SJKk7NQlXAiuBPQAiYiPgacAK4BnVtk2BxwOXRcTmwDxgI+BVwOHAXsAFEbF+07k/AdwJvBz4cHPGVTBxKfBHYD8DBEmSumdoeHi445NUNQe/zMy3RsQzgTMp3+4XZuaREfFi4GxgE+A9wBuBbTPzjir904ErgFdl5tcbahLOycwDmvJaCHwL+AKlBuMq4JWZuWIMRR45vyRJU8l2wMJWD+64uaFyKfC86vGewOXAJZTmgpFtv87MOyJiF+D8kQABIDOvrD78dwe+3nDe80bJL6o85wEHZ+Z97RR6+x12YdGixe0kZdXKpUyfsXVbaQFO33yfttO+eukZnLH1q9tOD7Dfc5a1nfbhcy/k9kMe1PIzJus9btu20z7kiFP4+8cO7Sj/z524vO20R/7pDE7YtrPr//aL39p22g122I3lC37eUf4bPfalbaft9H9/2lBnFZgrVyxmxvqz2k6/enh1R/l3+vonat7mP7Hznz17FtcvmD/mdN3quHgZ8ISIeASl2eGy6meniJjZsA1gK6DuE2oZsGnNtjr/XJ3n5HYDBEmStHbdChJ+Wv3eG9iV8i3/t8BdwL7AU1kTJNwEbFFzji2BW5u2jdYWcirwZeCcqmZCkiR1WVeChMz8G3At8HZgFXB1Zg5Tmh2OoDRrjAQJVwLPjYiNR9JHxM6UfgKXjyHbNwLfB34YEU/s9DVIkqQH6uY8CZdR+h78LDNXNW37w8jwRuBT1e8fR8T+EfFq4DvANcC3W80sM1cDB1MCi/MjYocuvAZJklTpdpAApamhedv9NQSZ+RdgH+BeSifFz1fHPXuMIxSo+iMcSAkwLoqIR7VXdEmS1KxboxvIzDMpQx8bt10JDNUcezXwzLWca2FdumrfnKbny4HnjLnAkiRprSbMtMySJKm/DBIkSVItgwRJklTLIEGSJNUySJAkSbUMEiRJUi2DBEmSVKtr8yRobA7+y0/aTvvqDtMDcEb7SVfNhU3P+F1n+dN++lVHwMPe+6OOcv/xprt3lH6n5Ss7St/pKoydpIfOV2LsNL2kicE7XZIk1TJIkCRJtQwSJElSLYMESZJUyyBBkiTVMkiQJEm1DBIkSVItgwRJklTLIEGSJNUySJAkSbUMEiRJUq2urd0QEYcApwIbZ+Zd3TrvWvJbH3gPcE5m/m+v85MkaaqZyDUJ6wPHADsOuiCSJE1GEzlIkCRJPdRyc0NE7AYcBewMPAz4A/DxzGxedPixEfEpYCfgRuDdmXl207neCrwN2LY65vOZ+emG/XOBJ2TmTg3b5gA3AC/KzO8Dd1a7To2IU6vH22XmwlZfkyRJGt1YahJmAz8F/hV4EfBtygf0K5uOOxM4F3gJcA1wVkQ8eWRnRBwGfA74bnWes4BPRsSRYyz7M6vfHwR2q35uGuM5JEnSKFquScjMb4w8jogh4FJgFnAY8PWGQ0/OzE9Ux/0Y+B2lBuIVETENOBaYm5nvrI4/PyIeDhwVEZ/JzHtbLNJV1e8/ZuYVrb6ORtcvmN9OsvutWrm0o/QTNW/zh2ct+2ZH6Vd1mP+gX//KFYundP7e++Y/VfIfS3PDJsBxwP7ANsD0ateSpkPvb1rIzNURcS7w8mrTLGBrSu1BozOBNwFPZM2Hf89tv8MuLFrU3pvNqpVLmT5j6y6XaPznPVny//Gmu7ed9lnLvsmFWx7YUf7PvfXyttN24/VPG2q/O9LKFYuZsf6sjvLvRKf5rx5e3VH+3vvmPxHznz17VltfjMcyBHIusCvwAUrtwB2UD/b9m467ueb5VtXjkd/Lmo4Zeb7pGMojSZJ6qKUgISJmAvsBb8nMLzRsr/s6sgVwS9Pzkb4CNzVsa7Rl9fvW6ve9lCGOjTZppaySJKk7Wq1z3KA6dvnIhojYGHhxzbEHNBwzjVLTMFLHsRhYyprmhxEHUmomrmk4bk4VnIx4TlOaFdXvmUiSpK5rqSYhM2+PiKuA90fEHcBq4EjgdspwyEavj4gVwLXA64EdgFdW51kdEccCX4yIW4ALgL0ozRbvaei0eA5wPHByNRzyKcChTWVaERE3AAdGxLWU2offZOYKJElSx8bSe+lVwPXA6cB/UYZAnl5z3CsotQnnAE8GDsrMq0d2ZuaXKXMkHAB8nxJAvDMzT2g45lpKULAbZajkXsDravJ6I7AZcCGlw+PgepRIkjTJjGUI5AJg35pdx1b751I6NwI8Yx3n+hxlroS1HdN4vhFDTcecDzxpbeeRJEntcVpmSZJUyyBBkiTVMkiQJEm1DBIkSVItgwRJklTLIEGSJNUySJAkSbXGssCTNGkcv94t6z5oFM/qMD3AhjM2GGj6SzbtbHqRK7Z4WttpH3fq8zvKG+CO845uO+2dJ3x93Qetw017/mPbab+4oLMVNI/dep+O0s9ffVtH6V/wyKd0lP6K2xd0lH7TDTfuKP0993U2KW+n916n2s1/5nrNyyG1xpoESZJUyyBBkiTVMkiQJEm1DBIkSVItgwRJklTLIEGSJNUySJAkSbUMEiRJUi2DBEmSVMsgQZIk1TJIkCRJtQwSJElSLYMESZJUq6VVICPi8cAngV2ADYA/ASdm5uer/fsD7wOeANwGnA68NzNXRsTewE+AJ2TmbxvOuQmwDHhzZp5cbdsD+CCwM3AP8B3gHZl5Z7X/EOBU4EnAp4B/Bm4E3pOZ32n7KkiSpAdptSbhe8Aq4DXAi4HPARsDRMSBlA/z+dW+44B/Az5Spb0UuAk4sOmcB1S/v12d5xnAhcCfgZcB/wG8gBIUNPsa8N3qHH8AvhERna2/KkmSHmCdNQkRsRmwHbB/Zl5Tbb6o2jcEfBw4PTPf3JBmOfD5iPhIZt4SEWcBBwHHNJz6IOD8zPxb9fwE4GeZeVDDeZYAF0XEEzLz2oa0n87MU6pjfkmpkdgP+MIYXrskSVqLoeHh4bUeEBHTgIWUav3PAj/JzJurfQFcR/nGf0FDslnADcDemXlJRPwz8FNgx8z8dRV43AQcmplfjYiNgDuAw4EvN5xnGnAXcFhmntbQ3PCozFzcUMalwCmZeXSLr3tOVT5JkqaS7Sif6S1ZZ01CZq6OiOcAHwJOATaMiJ8C/w5sVB32g1GSP6r6/XNKP4aDgF8DLwXuA86p9m8CTAdOqn5GO8+I25qerwBmruu1NNt+h11YtGjxug+ssWrlUqbP2LqttJ0aZN6TJf9nbPHYttNeuuQi9txm347y/9Xfrm877V1338BDN9quo/wv2fRJbad92uJz+eWs/dtO/7hTn992WoANn/1G7rmg/UrDO0/4ekf5b3HRJdy8715tp//igvZbRt+36Aw+MPvVbacHmL+6+e2zdd+78Txe9KgXdpT/FbcvaDvtX+5INn9YdJT/PfetaDttN+69TnSS/7bbbsPvrrt8zOla6riYmdcBL42IGcAewEeB84BnV4f8G3B1TdIbqvTDEfFNSr+E91CChR+OdEikfOgPA8dSH3AsbaWckiSpe1oKEkZk5krg4oj4FKXz4E3AEmBOZn55rYnhG8C7ImI/YC/glQ3n/XtEXAFEZh4/ljJJkqTeaKXj4pOATwBnAtdTmgbeDfw6M2+NiHcCX42IhwE/pFT9bw/8C/CyzLwbIDN/GRELgC9Rhjd+vymrIyidFFcD3wLuBLYFXkgZTvn7Tl+sJElqXStDIP9MGT3wXkoQcBLwf5ThjmTmmcD+wI7AWZThkG8GfkUJGBqdCWwFfG8keBiRmZcDewKbA1+lDLs8gtJhctnYX5okSepEKx0Xbwb+3zqO+SElgFjXuY4GRh2BkJlXAs9by/65wNya7XPWlbckSRobp2WWJEm1DBIkSVItgwRJklTLIEGSJNUySJAkSbUMEiRJUi2DBEmSVMsgQZIk1VrnUtGT1BzgBleBnLr5rz99Rttp77l3ERvOnN1R/itWrWw7bTde//Rp09tOu2L5jay/QfPCrGPIe6iz7yadXv9Orj1M/Ht/aGio7bT3rVjCeutv01H+nXzmTIb3nkHlP3v2LK5fMB/GuFS0NQmSJKmWQYIkSaplkCBJkmoZJEiSpFoGCZIkqZZBgiRJqmWQIEmSahkkSJKkWgYJkiSplkGCJEmqZZAgSZJqGSRIkqRaEyZIiIgNB10GSZKmkvXGmiAi9gA+COwM3AN8B3gHsBlwPbBfZp7XcPx0YAlwcmYeXW17AvBRYM/qsB8Bh2fmn6v9ewM/AZ4HvAV4JnBmRGwMbJGZezeV6VjgTcCszOxsiTdJkgSMsSYhIp4BXAj8GXgZ8B/AC4BTM/MGYD5wYFOyvYAtgW9U59gB+CkwE3gNcAjweOB7EdG8hulXgF8DL64efwXYMyK2ayjTEPBa4H8MECRJ6p6x1iScAPwsMw8a2RARS4CLqtqBbwDHRMQGmbm8OuQg4LeZeW31/BhKkPH8zFxRneM3wHWUgOP+WgjgrMx8X0Ne04DFlMDimGrzPsAc4NQxvhZJkrQWLQcJEbERsBtweEQ0prscWAk8Dfgm8ElKM8G51XEvAT7bcPyzgNOA1Q3nuQFYCOzEA4OExsdk5uqIOBU4OCKOzcxhSsDwi4YgpGXXL5g/1iQPsGrl0o7ST9S8zR/uuXfRQPMf9OtfsfzGgeY/la//oP/2961YMtD8B/36p1r+Y6lJ2ASYDpxU/TR7VGYuiYjLKbUH5wL7UvoqfKPhuM2Ad1c/DzpH0/NlNcecCrwP2CcirgJeCrxrDK/jftvvsAuLFi1uJymrVi5l+oyt20rbqUHmPVnyX3/6jLbT3nPvIjacObuj/Fesar9lrBuvf/q06W2nXbH8RtbfoPlWHUPeQ531l+70+ndy7WHi3/tDQ82tuq27b8US1lt/m47yHx4ebjvtZHjvGVT+s2fPauuL8ViChNuAYeBY4Ac1+0fCmzOBE6rRCAcBV2fmHxqOuxU4Gzi55hx/bXr+oP+mzFwYERdSahC2o/Sr+HrLr0KSJLWk5SAhM/8eEVcAkZnHr+XQs4D/Ag6ofj7StP8iSkfFX1bNBe34CnBKdZ5zMvO2Ns8jSZJGMdaOi0dQOimuBr4F3AlsC7wQeG9m/j4zb46IecAngEdQ+ik0OpYyCuK8iDiFUnuwDfBsYG5mzmuhHOdQmjyeChw1xtcgSZJaMKbGwcy8nDK3webAV4HvUQKHG3lg/4FvAFsBV2TmwqZz/B7YFbgb+BLwQ+A4YDmwoMVyLK/S3UgZkilJkrpszJMpZeaVlNELazvmZOr7HIzsv44yz8Jo++cBo/auqUZF7AOckpmr11FkSZLUhjEHCYMUEesDTwZeBfwD8MXBlkiSpMlrQgUJwNaU/gw3A2/IzPbGL0qSpHWaUEFC1b+h/UG+kiSpZRNmFUhJktRfBgmSJKmWQYIkSaplkCBJkmoZJEiSpFoTanSD1C2drgTYafqprBvXrpNzTOtwFcpOzzF9Wmf5z5je2dv2UIcDxGZM6zD/DlahBNhgvfU7St+piZr/+tPbS2dNgiRJqmWQIEmSahkkSJKkWgYJkiSplkGCJEmqZZAgSZJqGSRIkqRaBgmSJKmWQYIkSaplkCBJkmoNdFrmiHgCcA2wT2bOG2RZJEnSA1mTIEmSahkkSJKkWn1tboiINwNHAZsCFwOfbdr/TuAVwD8B9wLzgbdn5oKG9B8DHpmZdzWk2xv4CbBjZv66969EkqTJr281CRGxP/B54PvASyh9EU5pOmwWcCKwP3AYMB34WUQ8vNr/tWrby5rSvQ74lQGCJEndMzQ8PNyXjCJiPnBLZj6/YduXgddT03ExIqYD6wM3A2/JzNOr7f8DPCoz96qePxT4M3BkZp7YYnHmADd09IIkSZp4tgMWtnpwX5obImI94KnAW5t2fYcSJIwctyvwgerYTRuO+6eGx18BLoqI7TPzeuBAyuv42ljLtf0Ou7Bo0eKxJgNg1cqlTJ+xdVtpOzXIvM1/cuQ/fdr0ttOuWH4j62/wqLbTr1q9qu200PnrnzbUWQXqyhWLmbH+rLbTT5/Wfv733vsnZs7ctu30AEMMtZ32nnsXseHM2Z3lP9R+/nffs5CNNpzTUf6dmMj5b7vtLK7Ly8ecrl/NDZtRmglubtp+//OI2BY4HxgC3gA8A9i5OmZmQ5p5wPXAIdXz1wHnZuatPSi3JElTVr86Lv4VWAVs0bS98fnzgI2A/TPz73B/DURjjQKZORwRpwD/VjU97A48H0mS1FV9qUnIzPuAqykdEhu9pOHxhsBq4L6GbSNNCc3mUjo5fgVYAlzQrbJKkqSin0MgPwx8JyL+Gzgb2ItSezDiYkqTxKkR8RXg8cC7gNuaT5SZSyPiR8ALgY9kZmeNnJIk6UH6NgQyM88GDgdeBJwDPAX414b911D6GTydMkzyVcDLgdtHOeU51e9Te1NiSZKmtr5OplQNUWwepjjUsP+rwFeb9s8Z5XTPAS7PzD90rYCSJOl+A13gqR0R8URgJ0p/hlcMuDiSJE1aEy5IAL5HGVJ5UmZ+a9CFkSRpsppwQUJmzhl0GSRJmgpcBVKSJNUySJAkSbUMEiRJUi2DBEmSVGvajmK9AAAUQklEQVTCdVzskukA22yzVUcnmT27/ZXgOjXIvM1/4uffySqQnebf6SqQnebf6SqQnebfySqQneZdtL8KI8C2HebfySqUUFYzHKSJmv822zxy5OGYbv6h4eHhtjKc4HYHLht0ISRJ6rM9gJbXjJ6qQcIGlGWob6KsTilJ0mQ2HdgKuApY3mqiqRokSJKkdbDjoiRJqmWQIEmSahkkSJKkWgYJkiSplkGCJEmqZZAgSZJqGSRIkqRaBgmSJKmWQYIkSaplkCBJkmpN1VUgJ5yIOAg4DPgnYGbz/szcou+F6rOImMXor/8H/S9Rf0TE9sB/UhYm2xS4lbJA2Scy8/o+5D+l//ciYgPgUGAn4FHAWzLzD9V1+U1m/l+fyjFEmXv/5sy8rx95jgfj5fpPVQYJYxARmwBPoPyj/jAz/xYRM4EVmbm6h/m+CjgFmAs8s3o8DXgxcBtweq/ybirHesC21H9Q/K6H+W4MfBN4TrVpZK3ZxoVHOlv7+IH5fXMMhw9n5kHdyrumLE8DfgLcC3wfWAZsCbwUeHVE7JOZv+ph/gP/34uILYCHZOYN1fMhStDyOOCizPxeD/P+J+AC4OHAL4G9gY2r3XsALwQO7lX+VRleABwD7Ej5P98F+FVEfAm4NDP/p4d5D+zaV/n1/fqPh/s/IvYE3gjsSrnfodz7PwO+kJktr+LYKZsbWhAR0yPiY8Bi4BLgq8B21e5vU27gXvpP4APAW6rnJ2XmoVUZ/grc3cvMI2JGRPw3cAfwB+Camp9e+gglONmDEiAcQHmz+ApwA+VG6qbNx/DT62/RnwCuBuZk5qGZeVTD3/7qan8vDfR/rzIXeHvD8+OBk4DnAWdHxCE9zPuzwJ+AOcBzWROgQnkv2L2HeRMRBwPfBa4D/o0Hvmf/AfjXXubPYK89DOb6D/T+j4j3AfOAPSlLOv8X5TpcTnnfuyQiju52vqOxJqE1H6ZEz2+lfKtrrOI9lxLx9TJQ+Efgp5m5KiJWAQ8DyMw7I+KjwKfp7YfF+4H9KG9IZ1A+MP4OvAZ4NHB4D/MGeAFwNHBl9XxpZl4FXBoRn6R8kB3Yrcwyc59unasLdgEOzMwHfBhn5t0R8QngzB7nP+j/PYCnAl8CiIhplPvtPZn5sYg4DvgPyodZL+wBvDwzb4uI5tqqZZTq/156L/DxzDyqyv/Uhn2/Bd7V4/wHee1hANd/kPd/ROwOHEcJzI9rrqGursExwHERcXFm/qzXZbImoTUHA0dm5qnAjU37/ghs3+P87wA2qB4vAR7bsG8I+Ice538gcCylyh9gfmaenpnPoUS3+/c4/y2BGzNzFSU42bRh3w9Y0wzRUxExFBFbV80u/XIPo/99N6U0Q/TSoP/3oFQ131I9fhrldZ9RPb8Y2KGHed8LbDjKvm0oTS69NJtS3V7nXqqgrYcGee1h8Nf/fn26/98I/Cgzj6lrws7MVZn5fuDHwJt6WI77GSS05hGUYKDO+nSxPXwUVwFPqh5/F3h/RBwWEa8FPg5c0eP8HwX8vvqQvhfYpGHfGZT28V66EdisevwHSq3GiKfT4w/KiHhBRFxZ5XMj1d8iIr4cEa/pZd7AecAJ1TeMxjLtTmmG6WmbMIP/34PSzPe46vELgesyc0n1/OH09u9/AfCeiHh4w7bhqjPd4ZQgtZduBJ4yyr6dgAU9zn+Q1x4Gf/37ff/vCnyjheO+AezW5bxr2dzQmmsp35YvrNn3fKBnHccqH6F8o4BS9T8b+G9KkHcV8IYe538TJVCC0gdgT9Zci0f3OG8obxTPAs6mVG+fVnXoW16V5ZO9yrhqEz6FEgydxAOre39PaYLpWccx4B2UJq1LIuJm4GZKO+gWwM+Bd/Ywbxj8/x6U6/+xiHgW5YPqqIZ9uwK97N3+n8BPKR/GF1A6y74feDzlC8JLepg3lH43x0TEMuCcattQROwLHEHpI9BLg7z2MODrP4D7/5GU99h1uYHeN3UBBgmt+iDw7YjYEDiL8o+6Y0QcQHmTfHEvM8/MK6i+sWXmbcD+VSS9QWbe0cu8K/MobYPfA74MfDwidqB8SB8EfL3H+b8b2AggM78aEXcBL6NUQ74V+GIP8x5om3Bm3gLsHhHPA3amvDHcBFyZmef3Mu8q/0H/75GZH4mIJZTXfzjlTXvEpsDJPcz7xoh4MiVY25dSo7gV5X3gU9Xfp5c+SqnJOw1YVW37GaX28ouZ+dleZj7Ia1/lP+jr3+/7fyPK++q6rKBmlFkvDA0PD6/7KBERBwIfo/SyH7EEeGdmjmXIzIQTEY8ENsvMa6vnb2fNh/QFwPGZ+fcBFrFnIuJe4AWZeXH1JrES2CkzfxURzwTOy8zR2kylroiIR1M+JDejzJNxcWb+frClmvz6ff9HxGpKQLSuv+1jKMFLr5u6rUloVRUIfLMatztyo2ZmTvooKzP/DPy54fmnKdX+fTHgsdojbcIX1+zrSZtwRGw0MpohIjZax+HDmXlPt8swHg1ino6IuB44IDN/XbPvCcB3M7OrHZerMfJ1rmt4/MgqeCczL+1m/qOUaVBzpPT9+jfp+/0PfKrF4/ry2WOQMEZV9D7pI/iIuIox/BNm5i49LM5cys3479Xz4yltowuAt0bE6zNzbo/yHkSb8J0RsVtmzgfuYh1/h4hYDvwCOLzuzXSii4gZlHHir2XNSItmvfpGNWcteW4EzOpBnvMof/O6ScPq9Ozb5ICvPQzm+jfq9/2/3boP6S+DhBZUEymtVWYe0Y+y9NFv6VOk2oJBjtUeRJvwoawZTXMo6/47PIwyZ8WXKfMqTDZ9nacjIh7Gmo66UL61b9t02EzgFZQmx257YsPjrSj9AH4EfIc1HVdfSplc6NAe5N+o73OkjIPr36iv939mLurm+brBIKE1L6/Ztgnlzfl24G+UqHLSyMxDBl2GBusaq92zHv5Vc9JbIuJT9KlNODNPa3g8t5U0EXEt5YNkMmqcp+MMyjwdvwROj4jTKCOPujkU7u2UCWuGq5+zRzluiB7872Xmb0ceR8SHgdMzs3mGvR9FxAcpAXLdqKtu6fe1hwFf/0aDuP/HG4OEFmRmbRVQRDyd8g33jf0t0ZQzMlb7MvowVns8tgm34BJ6P0X0oNw/T0fVkax5no6v0d2hmF+jNN8MUeaGeBeQTcesoPRJ+lMX862zL3DiKPsuoQQJvdTvaw8Dvv6DvP8j4gZar8EdzsyeD0E3SOhAZl4ZER+n3MRPG3R5JrF+j9WexzhpE4bWVmGsZme7vZflGKC+ztORmX+gTNpFROwD/Coz7+x2Pi26lfJtvW7WxQOq/b3U9zlSxsH1n8fg7v9vt5DfU4F9WjiuKwwSOncLEIMuxGQ2gLHa46ZNOMbBKozjwDwGNE9HZl4CEBHPZ81SxR/MzD9V3zgXZObSXuUPnACcGBFzKN+qR/7/9qdM5PbWHuYNA54jZUDXf2D3f2aOOu9CROwKvI8SIPyWsqZQzxkktGCUYWjrU+axP57yB1MPZebp1HwgZmbXm3rGWZvwyCqMJ1BWATypGqO9MeXbZT9WYRy091JNy52Zn6mGwI7M0/E5ejjrYERsSflwfhqwkNL7/AuUlQlfR2nq6tkc+pl5UhUgvwf4POVb6yrKCqAvycxz1pa+CwZ27WEw13+c3f8jtSlHU1aA/BX9+bvfzyChNaMNQxui9K79l/4WZ2oa0FjtQbcJj4dVGAdqwPN0fA54KGXymoWUtvARF9L7ZeLJzHOBc6vJfDYD/lqto9Jzg54jhcFf/4Hd/xHxAkqQthtlauoXZmbfOycbJLSmbhjavZQOdfMzc2X/izR1DHis9qDbhOtWYZxXPe/XKox9N47m6Xge8NrMXBAPXqp4MWUlwr6oAoNlvc5nHF17GPz17/v9HxEvpdQcPYUSCO0z0uwyCAYJLejhRD1qTd/HajcYdJvwyCqMP2bNKoz3Ub5RvZ/+rMI4CONpno77Rtm+GWUp78lmPF17GOz17+v9HxG/pdSazKPUYlxZba+deXVkZtZeMkjQRDCIsdrAuGgTHg+rMPbdOJqn4zLg3yOi8f9r5AP0UOqn653QxtG1hwFf/wHc/4+tfu9D6YOwLq7dMB5Ui26MFlkPU6qEfw18NjNHm/hD7RvEWO37DbhNeOCrME5x7wYupywXfzblfj8sIh5P6QW/6wDLNhUM/Pr3+f5/XY/O2zaDhNa8o/q5gzIU6C+UKqcXARtT5vfeA/hWRLw2M7u5vrgGMFa7Tr/ahEdT9SwfeZNqZTlZdSgzr42Ip1Fqsg6hfIt8CXAR8PpqTL96ZDxd/37c/42zrY4XBgmt2ZrSw/yVTduPjIhvAJtk5rMi4nTK9MwGCd01j9HHar+CUpMwaVW9nI+mDANbD7gvIn4JfCgzzxto4aaAzPwj8P8GXY6pyuv/YNVIry16PEcHUNo1tW6vA04dZd+prPkHPpMyZE3d9V6qORIy8zOUQGw28GTKqId/Hz3pxBYRb6AER3cBb6OsI/K26vl3q/2SJoGIWBEROzc8nxYRF0dE8+fK0yjLWPecNQmtWY/S4/T8mn2PZU2wtYIuryMgALakDHW6Fh44VjsiXkhpcvjNwErXW++hrDb35qbtX4iIL1ACqC/2v1iTV0R8EzgqM/9YPV6b4cw8qB/lmiqm+PVfjzXTQVM93pvSrD0QBgmt+QbwkaqKZ6RPwuaUXvXHs6aW4ak8cBEQdcenKb2c60Yw7ETpL7JvX0vUP//A6KvgfZsyDFTdtTkwo3q8BeNrOOBU4PUfRwwSWvM2Si3BB4GPN2xfTmkj/8/q+ZWUDjXqrqdSxivX+Tnl7zNZ/QTYi/rJXPYCxsMKlJNKZu7T8HjvumMi4hHVaBN1mdd/fDFIaEFmrgDeFhHHUSa22ZIyVek1mXlrw3HzBlPCSW868JBR9j2Eso7GpBERj2t4+lng5Ij4B+Ac1kzmcgBlMpfX97+EU0dEvAnYODM/Vj1/MnAesFVE/C+wf2YuHmQZJzOv/+AZJIxBFRDMG3Q5pqCrKIsb1VW7/xtl7fnJ5FoeWMU6RJkH4g08cAlbKKvT9XxClSnscEqgNuJzwFLgXZQx/Cdgk08vTcXrf3hE3FQ9HrnX3xYRjcMvt+pXYQwSWhQRT6J0EtsJmAXsVq3G9yHg8sz84UALOLkdC1wYEVcCp1FqcbYCDqaMcHj24IrWE/us+xD1ybZAAkTE5sAzgH0zc15ErGD0xX/UHVPt+v8J2L1p2yLK3DB1x/acQUILqrXMvwv8jDIUr3HlseWUaNcgoUcy89KIeA5liuLPUaLr1ZQ+IM/OzMsGWb5uG+RiLnqQ5axpztqHsjT3yP/brayZ5Eu9MaWuf2bOGXQZmhkktOYjwNzMPKwa4dAYJPwv8MbBFGvqqPp77FYtdLIJ8Ld+LG4ynkTENOqXyZ5S16HP5gNviYjFlPk4ftQwJe/2lKpv9Y7XvwXVe8OFwBu6PQulQUJrHkNpA4MHD8e5A9i0v8WZuqoPxCnzoVhNxXwEcBiw3SiH2Sehd95JGfZ8DWXymkMb9h0E/HQQhZpCvP6t6dl8CgYJrbmZErXWeTx9ahvSlPTvwJHAx4APUYbhrqJMR70+8OHBFW3yy8zfAY+uRpfcmpmNXxLeRekfox7x+g/e0PCw81SsS0R8jNJJ7mWUcfkrKdNi/p1SxfOVzDxucCXUZBUR1wJfoixTuxLYqeowO43qG1ZmHjnIMkoarGqFyvvfH7p5btduaM37KMPsLmFNrcG5lKFqv8Fvc+qd7YD/rdphV1J11MrM1cBJwGsHWDZJk5zNDS2oluXdLyL2pUz/uxmlZ+1FmVk3E57ULbcAD60e/wl4CnBx9XwTYMNBFErS1GCQMAaZeRFOu6z++imwM2Xdiq8Bx0bEppRpwt+C/4+Sip70HTBIGEU11K5lDkNTjxxLWQETSrPWI4BDKDUIF1Dm6JCkoXUf0sZJ7bhYLyJWM4bILDMdhqauqjonbgXcnpl3Dbo8kqYeaxJGdyguUarBmgYsBF5EWaNB0hQSEbsB/wr8E/UTqe3S6zIYJIwiM+cOugya2jLzvohYBIyp6UvSxBcRz6b0RbqIsp7DDynNjM8AFlNG2/WcQcIoImL+GA4fzsyn96wwmso+Crw3Ii7NzL8OujCS+uZ44L8oq12uBN5XzZEyG/gxfVqR2CBhdL/D5gYN3nMo/RIWRcQvgWU0/V9m5oGDKJiknnoccDRlMbth4CEAmbkoIo4FjqMsONhTBgmjyMxDBl0GiTInRzY9b2QgK01O9wLTMnM4Im4CHs2aFTDvAGb1oxAGCdL49hPg5Mx80Gp3EbEVZeEnSZPPr4GgDHW+CDgqIpZQ5kg5nrLoVc85LbM0vh3D6N8YtuaBy5ZLmjw+w5qawvdQ1gr6MeWLwxaUydR6zpoEaXwbYvQmhVnA3/pYFkl9kpk/aHi8JCKeBuxAGeFwXWau6Ec5nExJGmci4rWsWbhpL+BqShtko5nAE4HzM/OlfSyepCnEmgRp/LmbsrATlJqE2ykLijVaQRk3fVIfyyWphyLizcBZmfmX6vHaDGfmf/e6TNYkSONYRJwKHJ+ZNwy6LJJ6q1oOYNfMnF89XpvhfiwHYJAgSZJq2dwgSdI4EBF7juHw4cy8bN2HdcYgQZKk8WEeZTTTyLLPjVX9dSOdet7cYJAgSdL48MSGx1sBp1BWgP0OcDNlfoSXAs+lrFTcc/ZJkCRpnImIc4FrMvPomn0fBHbMzP16XQ5nXJQkafzZl9GXg74E2LsfhTBIkCRp/LkV2H+UfQfw4LlTesI+CZIkjT8nACdGxBzgu6zpk7A/8Hzgrf0ohH0SJEkahyJif8riTk+ljGRYRZmm/cOZeU4/ymCQIEnSOBYR04HNgL9m5qp+5m2QIEmSatlxUZIk1TJIkCRJtQwSJElSLYMESZJU6/8DMfWAZm6EkqkAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input English Sentence: i come home from work about seven every day\n",
      "Actual spanish:  regreso a casa del trabajo a eso de las siete todos los dias\n",
      "Predicted spanish: llegue a la casa del trabajo a las siete de las siete del dia\n"
     ]
    }
   ],
   "source": [
    "k+=1\n",
    "(encoder_input, actual_output), _ = next(train_gen)\n",
    "print(X_train[k:k+1])\n",
    "decoded_sent = visualize(X_train[k:k+1], encoder_input)\n",
    "\n",
    "# Translation of Prediction from Google Translator: arrive at the house of work at seven o'clock seven o'clock\n",
    "print('Input English Sentence:', X_train[k:k+1][0])\n",
    "print('Actual spanish:', y_train[k:k+1][0][6:-4])\n",
    "print('Predicted spanish:', decoded_sent[:-4])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Translation of Predicted Spanish (Source: Google Translator): \"arrive at the house of work at seven o'clock seven o'clock\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "NMT-Training-Inference.ipynb",
   "provenance": [],
   "version": "0.3.2"
  },
  "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": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}