Usecase7_with same_dataset.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Copy of Copy of Usecase6_(1).ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "hl1e0rwtZw3F"
      },
      "source": [
        "import numpy as np # linear algebra\n",
        "import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n",
        "\n",
        "# Input data files are available in the \"../input/\" directory.\n",
        "# For example, running this (by clicking run or pressing Shift+Enter) will list the files in the input directory\n",
        "\n",
        "import os\n",
        "import string\n",
        "from string import digits\n",
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline\n",
        "import re\n",
        "\n",
        "import seaborn as sns\n",
        "from sklearn.utils import shuffle\n",
        "from sklearn.model_selection import train_test_split\n",
        "from tensorflow.keras.layers import Input, LSTM, Embedding, Dense, Add, Bidirectional, TimeDistributed\n",
        "from tensorflow.keras.models import Model,Sequential\n",
        "pd.set_option('display.max_rows', 500)\n",
        "pd.set_option('display.max_columns', 500)\n",
        "pd.set_option('display.width', 1000)\n",
        "import tensorflow as tf\n",
        "from keras.layers import merge\n"
      ],
      "execution_count": 102,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yvHHfeZKhFlj",
        "outputId": "fba3d3d5-7694-4fb4-e29d-1cee6769d7b2"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": 103,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Gre8_g22vOQW"
      },
      "source": [
        "import codecs\n",
        "with codecs.open(\"/content/drive/MyDrive/Colab Notebooks/chatbot/movie_lines_cleaned.txt\", \"rb\", encoding=\"utf-8\", errors=\"ignore\") as f:\n",
        "    lines = f.read().split(\"\\n\")"
      ],
      "execution_count": 104,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EtQjt9--vhck",
        "outputId": "097e10f4-e2b0-48f8-fd2c-a708a6ca7d5a"
      },
      "source": [
        "lines[:4]"
      ],
      "execution_count": 105,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['They do not!', 'They do to!', 'I hope so.', 'She okay?']"
            ]
          },
          "metadata": {},
          "execution_count": 105
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GD6JzA1HbJtY"
      },
      "source": [
        "context_and_target = []\n",
        "for i in range(0, len(lines), 2):\n",
        "  context_and_target.append((lines[i], lines[i+1]))"
      ],
      "execution_count": 106,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oRzDZ3pUxB_5"
      },
      "source": [
        "context_and_target= context_and_target[:152356]"
      ],
      "execution_count": 107,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Kb1T4-u9xCMx"
      },
      "source": [
        "context, target = zip(*context_and_target)\n"
      ],
      "execution_count": 108,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "AJqOqdRBxCZb",
        "outputId": "61743b00-332b-4b58-d2b9-26e855ff36ee"
      },
      "source": [
        "len(target), len(context)"
      ],
      "execution_count": 109,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(152356, 152356)"
            ]
          },
          "metadata": {},
          "execution_count": 109
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mHa-sZYWxCjh"
      },
      "source": [
        "def convert(set):\n",
        "    return list(set)\n",
        "\n",
        "context = convert(context)\n",
        "target = convert(target)"
      ],
      "execution_count": 110,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "du3BmcWV_1eD"
      },
      "source": [
        "##Clean the text data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P2i-PCBoxeZa"
      },
      "source": [
        "context = list (context)\n",
        "target = list (target)"
      ],
      "execution_count": 111,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DetttMxjxeit",
        "outputId": "afe0768d-12ea-4f43-d0bd-1c209c7e8d00"
      },
      "source": [
        "len(context), len(target)"
      ],
      "execution_count": 112,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(152356, 152356)"
            ]
          },
          "metadata": {},
          "execution_count": 112
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VJ_6dR0Uxeqr",
        "outputId": "eeb69262-d082-4019-c43d-015ab786955a"
      },
      "source": [
        "len(target)"
      ],
      "execution_count": 113,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "152356"
            ]
          },
          "metadata": {},
          "execution_count": 113
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TJU-JURcxeu7"
      },
      "source": [
        "import re\n",
        "def clean_text(text):\n",
        "    '''Clean text by removing unnecessary characters and altering the format of words.'''\n",
        "\n",
        "    text = text.lower()\n",
        "    \n",
        "    text = re.sub(r\"i'm\", \"i am\", text)\n",
        "    text = re.sub(r\"he's\", \"he is\", text)\n",
        "    text = re.sub(r\"she's\", \"she is\", text)\n",
        "    text = re.sub(r\"it's\", \"it is\", text)\n",
        "    text = re.sub(r\"that's\", \"that is\", text)\n",
        "    text = re.sub(r\"what's\", \"what is\", text)\n",
        "    text = re.sub(r\"where's\", \"where is\", text)\n",
        "    text = re.sub(r\"how's\", \"how is\", text)\n",
        "    text = re.sub(r\"let's\", \"lets\", text)\n",
        "    text = re.sub(r\"em\", \"them\", text)\n",
        "    text = re.sub(r\"\\'ll\", \" will\", text)\n",
        "    text = re.sub(r\"\\'ve\", \" have\", text)\n",
        "    text = re.sub(r\"\\'re\", \" are\", text)\n",
        "    text = re.sub(r\"\\'d\", \" would\", text)\n",
        "    text = re.sub(r\"gotta\", \"got to\", text)\n",
        "    text = re.sub(r\"outta\", \"out of\", text)\n",
        "    text = re.sub(r\"gotcha\", \"got you\", text)\n",
        "    text = re.sub(r\"here's\", \"here is\", text)\n",
        "    text = re.sub(r\"whaddya\", \"what do you\", text)\n",
        "    text = re.sub(r\"there's\", \"there is\", text)\n",
        "    text = re.sub(r\"wid\", \"with\", text)\n",
        "    text = re.sub(r\"everyone's\", \"everyone is\", text)\n",
        "    text = re.sub(r\"wouldn't\", \"would not\", text)\n",
        "    text = re.sub(r\"couldn't\", \"could not\", text)\n",
        "    text = re.sub(r\"won't\", \"will not\", text)\n",
        "    text = re.sub(r\"can't\", \"cannot\", text)\n",
        "    text = re.sub(r\"that ain't\", \"that is not\", text)\n",
        "    text = re.sub(r\"they ain't\", \"they are not\", text)\n",
        "    text = re.sub(r\"i ain't\", \"i am not\", text)\n",
        "    text = re.sub(r\"he ain't\", \"he is not\", text)\n",
        "    text = re.sub(r\"she ain't\", \"she is not\", text)\n",
        "    text = re.sub(r\"n't\", \" not\", text)\n",
        "    text = re.sub(r\"n'\", \"ng\", text)\n",
        "    text = re.sub(r\"'bout\", \"about\", text)\n",
        "    text = re.sub(r\"'til\", \"until\", text)\n",
        "    re.sub(r\"\\'s\", \"\", text)\n",
        "    text = re.sub(r\"[-()\\\"#/@;:<>{}`+=~|.!?,]\", \" \", text)\n",
        "    \n",
        "    return text"
      ],
      "execution_count": 114,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "id": "rUXW_RHaIRjA",
        "outputId": "3a60808b-f89c-4f10-c568-ab0302fe585c"
      },
      "source": [
        "re.sub(r\"\\'s\", \"\", \"let's\")"
      ],
      "execution_count": 115,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'let'"
            ]
          },
          "metadata": {},
          "execution_count": 115
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EPYO6sQpxe0M"
      },
      "source": [
        "tidy_target = []\n",
        "#every conversation passed of target \n",
        "for conve in target:\n",
        "    text = clean_text(conve)\n",
        "    tidy_target.append(text)"
      ],
      "execution_count": 116,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vir93dX8xe4L",
        "outputId": "8c412d5c-9c9f-4723-a53a-6d518372ad90"
      },
      "source": [
        "len(tidy_target)"
      ],
      "execution_count": 117,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "152356"
            ]
          },
          "metadata": {},
          "execution_count": 117
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GKD3YmEOxe8K"
      },
      "source": [
        "tidy_context = []\n",
        "#every conversation of context passed\n",
        "for conve in context:\n",
        "    text = clean_text(conve)\n",
        "    tidy_context.append(text)"
      ],
      "execution_count": 118,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XRdAQBJ5xfAD",
        "outputId": "a363cc54-ff2f-43db-bca7-4c9fe28f4e65"
      },
      "source": [
        "len(tidy_context), len(tidy_target)"
      ],
      "execution_count": 119,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(152356, 152356)"
            ]
          },
          "metadata": {},
          "execution_count": 119
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aySx2p2SGIQL"
      },
      "source": [
        "contextdf1= pd.DataFrame(tidy_context, columns=['context'])\n",
        "targetdf1= pd.DataFrame(tidy_target, columns=['target'])"
      ],
      "execution_count": 120,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rqbOfveyzJaJ",
        "outputId": "2918a770-c482-48fe-dd45-37eeff7e2b43"
      },
      "source": [
        "targetdf1.shape, contextdf1.shape"
      ],
      "execution_count": 121,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((152356, 1), (152356, 1))"
            ]
          },
          "metadata": {},
          "execution_count": 121
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "N-n6MqjRGxz4"
      },
      "source": [
        ""
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sWk5HNCGxfDL"
      },
      "source": [
        "lines= pd.concat([contextdf1, targetdf1], axis=1)"
      ],
      "execution_count": 122,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "FswLmtyz0GUu",
        "outputId": "a40c17df-43c7-4a04-a998-6ad81cd66ffb"
      },
      "source": [
        "lines.drop_duplicates(inplace= True)\n",
        "lines= lines.sample(n= 30000, random_state=42)\n",
        "lines.head()"
      ],
      "execution_count": 123,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>context</th>\n",
              "      <th>target</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>4476</th>\n",
              "      <td>gosh  that is a pretty name  buy you a drink</td>\n",
              "      <td>lyla devereux</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>51190</th>\n",
              "      <td>no gas    car dead   we need gas  i do not wan...</td>\n",
              "      <td>this car runs on gas</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>39961</th>\n",
              "      <td>hey  betty  are you okay</td>\n",
              "      <td>oh  appreciate it  i was trying to  ahh  give ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>149180</th>\n",
              "      <td>there was a crash last year  they just got a f...</td>\n",
              "      <td>do not like airlines  lousy unions</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>125343</th>\n",
              "      <td>so anyway  they closed the doors  and she star...</td>\n",
              "      <td>crash paranoia</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                  context                                             target\n",
              "4476        gosh  that is a pretty name  buy you a drink                                      lyla devereux \n",
              "51190   no gas    car dead   we need gas  i do not wan...                              this car runs on gas \n",
              "39961                           hey  betty  are you okay   oh  appreciate it  i was trying to  ahh  give ...\n",
              "149180  there was a crash last year  they just got a f...              do not like airlines  lousy unions   \n",
              "125343  so anyway  they closed the doors  and she star...                                    crash paranoia "
            ]
          },
          "metadata": {},
          "execution_count": 123
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "myym5gGm0Gdl"
      },
      "source": [
        "#Lowercase all characters\n",
        "lines['context']=lines['context'].apply(lambda x: x.lower())\n",
        "lines['target']=lines['target'].apply(lambda x: x.lower())\n",
        "\n",
        "# Remove quotes\n",
        "lines['context']=lines['context'].apply(lambda x: re.sub(\"'\", '', x))\n",
        "lines['target']=lines['target'].apply(lambda x: re.sub(\"'\", '', x))\n",
        "\n",
        "exclude = set(string.punctuation) # Set of all special characters\n",
        "# Remove all the special characters\n",
        "lines['context']=lines['context'].apply(lambda x: ''.join(ch for ch in x if ch not in exclude))\n",
        "lines['target']=lines['target'].apply(lambda x: ''.join(ch for ch in x if ch not in exclude))\n",
        "\n",
        "# Remove all numbers from text\n",
        "remove_digits = str.maketrans('', '', digits)\n",
        "lines['context']=lines['context'].apply(lambda x: x.translate(remove_digits))\n",
        "lines['target']=lines['target'].apply(lambda x: x.translate(remove_digits))\n",
        "\n",
        "\n",
        "# Remove extra spaces\n",
        "lines['context']=lines['context'].apply(lambda x: x.strip())\n",
        "lines['target']=lines['target'].apply(lambda x: x.strip())\n",
        "lines['context']=lines['context'].apply(lambda x: re.sub(\" +\", \" \", x))\n",
        "lines['target']=lines['target'].apply(lambda x: re.sub(\" +\", \" \", x))\n",
        "lines['target']=lines['target'].apply(lambda x: 'START_ ' + x + ' _END')\n",
        "\n",
        "# Add start and end tokens to target sequences"
      ],
      "execution_count": 124,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8bVXfDFv0GhM",
        "outputId": "3648ec08-c1d2-4543-8ed5-27b5e4ea432c"
      },
      "source": [
        "lines.shape"
      ],
      "execution_count": 125,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(30000, 2)"
            ]
          },
          "metadata": {},
          "execution_count": 125
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "T5_iltDGOlcl",
        "outputId": "6407e4bf-23cc-48d0-e6e0-7a9947f49895"
      },
      "source": [
        "lines['context'].values[:20]"
      ],
      "execution_count": 126,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array(['gosh that is a pretty name buy you a drink',\n",
              "       'no gas car dead we need gas i do not want to get shot for running out of gas',\n",
              "       'hey betty are you okay',\n",
              "       'there was a crash last year they just got a favorable ruling on a lawsuit even the plaintiffs do not know',\n",
              "       'so anyway they closed the doors and she starts to scream i have never in my life heard anyone scream quite like that',\n",
              "       'looks like the alternator', 'he is so great u you u hire him',\n",
              "       'come here',\n",
              "       'yeah i am still at bellevue the john doe did not pan out but you will never guess whos lying up in the meat locker',\n",
              "       'if kats not going you are not going', 'what about you',\n",
              "       'i was yeah i was thinking about going there for business',\n",
              "       'hello yes what is on your mind baxter',\n",
              "       'have you ever been to that station', 'what',\n",
              "       'swear on jimmys grave',\n",
              "       'she runs a kind of hair dress ing and you know beauty salon',\n",
              "       'see the jews love to separate things the holy from the profane milk from meat wool from linen the sabbath from the week the jew from the gentile as if one little scrap of this was going to completely contaminate that',\n",
              "       'it is all mthemorabilia but incidental and not integral if you know what i mean',\n",
              "       'who were you talking to'], dtype=object)"
            ]
          },
          "metadata": {},
          "execution_count": 126
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EtU413MW5lbu"
      },
      "source": [
        "# lines['target'] = lines['target'].apply(lambda x : 'START_ '+ x + ' _END')\n",
        "# lines['length_cont_sentence']=lines['context'].apply(lambda x:len(x.split(\" \")))\n",
        "# lines['length_target_sentence']=lines['target'].apply(lambda x:len(x.split(\" \")))\n",
        "\n",
        "# lines=lines[lines['length_cont_sentence']<=20]\n",
        "# lines=lines[lines['length_target_sentence']<=20]\n",
        "# print(\"maximum length of Spanish Sentence \",max(lines['length_cont_sentence']))\n",
        "# print(\"maximum length of English Sentence \",max(lines['length_target_sentence']))\n",
        "\n",
        "# max_length_src=max(lines['length_cont_sentence'])\n",
        "# max_length_tar=max(lines['length_target_sentence'])"
      ],
      "execution_count": 127,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YJ_EzEBQ6Nnq",
        "outputId": "519e13e5-0cc1-4f1b-a9d4-92f20b1dfd6b"
      },
      "source": [
        "lines.shape"
      ],
      "execution_count": 128,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(30000, 2)"
            ]
          },
          "metadata": {},
          "execution_count": 128
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NbNKPFdXR8Gr"
      },
      "source": [
        "# import os\n",
        "# import numpy as np\n",
        "# def load_npy_to_any(path='', name='/content/word-glove-target-idx2word.npy'):\n",
        "#     \"\"\"Load .npy file.\n",
        "\n",
        "#     Examples\n",
        "#     ---------\n",
        "#     - see save_any_to_npy()\n",
        "#     \"\"\"\n",
        "#     file_path = os.path.join(path, name)\n",
        "#     try:\n",
        "#         npy = np.load(file_path, allow_pickle=True).item()\n",
        "#     except:\n",
        "#         npy = np.load(file_path, allow_pickle=True)\n",
        "#     finally:\n",
        "#         try:\n",
        "#             return npy\n",
        "#         except:\n",
        "#             print(\"[!] Fail to load %s\" % file_path)\n",
        "#             exit()"
      ],
      "execution_count": 129,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E-S2ZmEy582y"
      },
      "source": [
        "### Get English and Spanish Vocabulary\n",
        "all_cont_words=set()\n",
        "for i in lines['context']:\n",
        "    for word in i.split():\n",
        "        if word not in all_cont_words:\n",
        "            all_cont_words.add(word)\n",
        "\n",
        "all_target_words=set()\n",
        "for i in lines['target']:\n",
        "    for word in i.split():\n",
        "        if word not in all_target_words:\n",
        "            all_target_words.add(word)\n"
      ],
      "execution_count": 130,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sqzkcGrS5vOf",
        "outputId": "948349fd-582a-4a83-f2e4-c8689d03cc3e"
      },
      "source": [
        "lines.shape"
      ],
      "execution_count": 131,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(30000, 2)"
            ]
          },
          "metadata": {},
          "execution_count": 131
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 419
        },
        "id": "LQhOWaA74Ky3",
        "outputId": "17fb183b-9627-46ce-a871-3140318fbf0d"
      },
      "source": [
        "input_words = sorted(list(all_cont_words))\n",
        "target_words = sorted(list(all_target_words))\n",
        "\n",
        "totalword= input_words + target_words\n",
        "totalword= list(set(totalword))\n",
        "\n",
        "num_encoder_tokens = len(totalword)\n",
        "\n",
        "num_encoder_tokens = num_encoder_tokens+1 #for zero padding\n",
        "\n",
        "input_token_index = dict([(word, i+1) for i, word in enumerate(totalword)])#(avoiding index 0 for lstm layer)\n",
        "reverse_input_char_index = dict((i, word) for word, i in input_token_index.items())\n",
        "\n",
        "lines = shuffle(lines)\n",
        "lines.head(-20)"
      ],
      "execution_count": 132,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>context</th>\n",
              "      <th>target</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>68081</th>\n",
              "      <td>forget it bunghole</td>\n",
              "      <td>START_ me first _END</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>53013</th>\n",
              "      <td>why do you want me to go back to my parents th...</td>\n",
              "      <td>START_ i am not ever been there but it do not ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>112742</th>\n",
              "      <td>yes i am very sorry</td>\n",
              "      <td>START_ did you kill all those guards _END</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6833</th>\n",
              "      <td>maybe but that do not make me a killer</td>\n",
              "      <td>START_ you hated west more than any of us _END</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>56432</th>\n",
              "      <td>you are an animal</td>\n",
              "      <td>START_ it dies we live _END</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>109724</th>\n",
              "      <td>expect it did you get anything from the room</td>\n",
              "      <td>START_ he will be contacting us again _END</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>87921</th>\n",
              "      <td>alert all commands calculate every possible de...</td>\n",
              "      <td>START_ lord vader our ships have completed the...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14947</th>\n",
              "      <td>what</td>\n",
              "      <td>START_ we got to be close _END</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>70270</th>\n",
              "      <td>well i did not see him anymore i will tell you...</td>\n",
              "      <td>START_ i do not know if you can fire a psychia...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>98870</th>\n",
              "      <td>next channel darling</td>\n",
              "      <td>START_ mute _END</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>29980 rows × 2 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                  context                                             target\n",
              "68081                                  forget it bunghole                               START_ me first _END\n",
              "53013   why do you want me to go back to my parents th...  START_ i am not ever been there but it do not ...\n",
              "112742                                yes i am very sorry          START_ did you kill all those guards _END\n",
              "6833               maybe but that do not make me a killer     START_ you hated west more than any of us _END\n",
              "56432                                   you are an animal                        START_ it dies we live _END\n",
              "...                                                   ...                                                ...\n",
              "109724       expect it did you get anything from the room         START_ he will be contacting us again _END\n",
              "87921   alert all commands calculate every possible de...  START_ lord vader our ships have completed the...\n",
              "14947                                                what                     START_ we got to be close _END\n",
              "70270   well i did not see him anymore i will tell you...  START_ i do not know if you can fire a psychia...\n",
              "98870                                next channel darling                                   START_ mute _END\n",
              "\n",
              "[29980 rows x 2 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 132
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tfPhZuaOWWJk",
        "outputId": "db5be38d-2e43-4774-87ab-ac2213bc4761"
      },
      "source": [
        "input_token_index['START_']"
      ],
      "execution_count": 133,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2672"
            ]
          },
          "metadata": {},
          "execution_count": 133
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FnKzPWIF6ey-",
        "outputId": "29014507-868d-4781-efe4-0db0cfa86c08"
      },
      "source": [
        "num_encoder_tokens"
      ],
      "execution_count": 134,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "24783"
            ]
          },
          "metadata": {},
          "execution_count": 134
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FdXJHgYEV7HM",
        "outputId": "1971ba43-80c8-435d-e46c-d1655ef00e71"
      },
      "source": [
        "\n",
        "lines['length_cont_sentence']=lines['context'].apply(lambda x:len(x.split(\" \")))\n",
        "lines['length_target_sentence']=lines['target'].apply(lambda x:len(x.split(\" \")))\n",
        "\n",
        "lines=lines[lines['length_cont_sentence']<=20]\n",
        "lines=lines[lines['length_target_sentence']<=20]\n",
        "print(\"maximum length of Spanish Sentence \",max(lines['length_cont_sentence']))\n",
        "print(\"maximum length of English Sentence \",max(lines['length_target_sentence']))\n",
        "\n",
        "max_length_src=max(lines['length_cont_sentence'])\n",
        "max_length_tar=max(lines['length_target_sentence'])"
      ],
      "execution_count": 135,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "maximum length of Spanish Sentence  20\n",
            "maximum length of English Sentence  20\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Q_AxUQlVWL1Z",
        "outputId": "2067b9e5-5c9f-4766-dd12-1fa5e39fd64c"
      },
      "source": [
        "lines.iloc[:7]['context']"
      ],
      "execution_count": 136,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "68081                         forget it bunghole\n",
              "112742                       yes i am very sorry\n",
              "6833      maybe but that do not make me a killer\n",
              "56432                          you are an animal\n",
              "138865                             lucien baines\n",
              "92512                             i got no music\n",
              "75087                                   what did\n",
              "Name: context, dtype: object"
            ]
          },
          "metadata": {},
          "execution_count": 136
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8JvtE5o7vu_K",
        "outputId": "8d60bb63-1240-4244-972f-8c977710caa5"
      },
      "source": [
        "lines.iloc[:7]['target']"
      ],
      "execution_count": 137,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "68081                               START_ me first _END\n",
              "112742         START_ did you kill all those guards _END\n",
              "6833      START_ you hated west more than any of us _END\n",
              "56432                        START_ it dies we live _END\n",
              "138865                           START_ he hates me _END\n",
              "92512                       START_ you are for real _END\n",
              "75087                        START_ hey captain god _END\n",
              "Name: target, dtype: object"
            ]
          },
          "metadata": {},
          "execution_count": 137
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hqxNr92Gg043"
      },
      "source": [
        "X, y = lines['context'], lines['target']\n",
        "testX, testy = lines['context'], lines['target']\n",
        "# X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.2,random_state=42)\n",
        "# X_train.shape, X_test.shape"
      ],
      "execution_count": 138,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Qp5VdAnIAMWb"
      },
      "source": [
        "##Generator to generate the data for training"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LeH2vARvg8U-"
      },
      "source": [
        "def generate_batch(x1, y1, batch_size):\n",
        "    ''' Generate a batch of data '''\n",
        "    while True:\n",
        "        for j in range(0, len(x1), batch_size):\n",
        "            encoder_input_data = np.zeros((batch_size, max_length_src),dtype='float32')\n",
        "            decoder_input_data = np.zeros((batch_size, max_length_tar),dtype='float32')\n",
        "            decoder_target_data = np.zeros((batch_size, max_length_tar, num_encoder_tokens),dtype='float32')\n",
        "            for i, (input_text, target_text) in enumerate(zip(x1[j:j+batch_size], y1[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] = input_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, input_token_index[word]] = 1.\n",
        "            yield([encoder_input_data, decoder_input_data], decoder_target_data)"
      ],
      "execution_count": 139,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_9-qbW_OAW23"
      },
      "source": [
        "##Create the model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3XizX5Qcg_s6"
      },
      "source": [
        "#Model\n",
        "#curent dimesion of data is 128 batch size, 20, num_encoder_tokens\n",
        "latent_dim=300\n",
        "# Encoder input in any shape, here None is batch size but data will be 128 batch size, 20 timesteps\n",
        "encoder_inputs = Input(shape=(None, ))\n",
        "\n",
        "#embedding layer:- input layer is vocabulary , output dimension of embedding layer with None batch size, maskzero only if input is of variable length \n",
        "enc_emb =  Embedding(num_encoder_tokens, latent_dim, mask_zero= True)(encoder_inputs)\n",
        "\n",
        "\n",
        "\n"
      ],
      "execution_count": 140,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NU4zWm5B7QwM",
        "outputId": "e68d0197-506b-44bf-b8bd-22529d24a1a6"
      },
      "source": [
        "enc_emb.shape"
      ],
      "execution_count": 141,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "TensorShape([None, None, 300])"
            ]
          },
          "metadata": {},
          "execution_count": 141
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "l-3yHMS17Oh0"
      },
      "source": [
        "encoder_lstm = LSTM(latent_dim, return_state=True, activation=\"tanh\", recurrent_activation=\"sigmoid\", return_sequences= True)\n",
        "#the output of 1st lstm is (None, 20, latent dims)\n",
        "\n",
        "backward_lstm2= LSTM(latent_dim, return_state=True, activation=\"tanh\", recurrent_activation=\"sigmoid\", return_sequences= True, go_backwards=True)\n",
        "\n",
        "encoderbilstm= Bidirectional(layer= encoder_lstm, backward_layer= backward_lstm2, merge_mode= 'sum')(enc_emb)\n",
        "\n",
        "\n"
      ],
      "execution_count": 142,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NwAknmknmfZJ",
        "outputId": "40dcaf8f-bd57-4e4b-ba8e-b03769d45e34"
      },
      "source": [
        "encoderbilstm[0]"
      ],
      "execution_count": 143,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<KerasTensor: shape=(None, None, 300) dtype=float32 (created by layer 'bidirectional_2')>"
            ]
          },
          "metadata": {},
          "execution_count": 143
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SoqEaEcgjr8D"
      },
      "source": [
        "# below used hidden and cell state of last model\n",
        "encoder_lstm2 = LSTM(latent_dim, return_state=True, activation=\"tanh\", recurrent_activation=\"sigmoid\")\n",
        "#the output of 1st lstm is (None, 20, latent dims)\n",
        "\n",
        "backward_lstm3= LSTM(latent_dim, return_state=True, activation=\"tanh\", recurrent_activation=\"sigmoid\", go_backwards=True)\n",
        "# input layer for lstm must be 3D https://machinelearningmastery.com/reshape-input-data-long-short-term-memory-networks-keras/\n",
        "bidir2= Bidirectional(layer= encoder_lstm2, backward_layer= backward_lstm3, merge_mode= 'sum')(encoderbilstm[0], initial_state= encoderbilstm[1:])\n"
      ],
      "execution_count": 144,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PZ1kaGM37Z7n",
        "outputId": "75744117-8ba7-4a63-b2df-7fcb712a1552"
      },
      "source": [
        "bidir2[1].dtype"
      ],
      "execution_count": 145,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tf.float32"
            ]
          },
          "metadata": {},
          "execution_count": 145
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Emr2MIqtE5UG"
      },
      "source": [
        "#first normal lstm hidden states are generated than lstm with backward layer are generated\n",
        "forwardh, forwardc, backwardh, backwardc= bidir2[1], bidir2[2], bidir2[3], bidir2[4]"
      ],
      "execution_count": 146,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yh0sZByKFJ4H"
      },
      "source": [
        "stateh= tf.add(forwardh,backwardh)\n",
        "statec= tf.add(forwardc,backwardc)\n",
        "state= [stateh, statec]"
      ],
      "execution_count": 147,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 533
        },
        "id": "vBht-SeC4ymt",
        "outputId": "1c45ec7d-3bc5-47c3-eb89-875ff6e86554"
      },
      "source": [
        "encodermodel=Model(encoder_inputs, [state])\n",
        "from tensorflow.keras.utils import plot_model\n",
        "plot_model(encodermodel, show_shapes=True)\n"
      ],
      "execution_count": 148,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 148
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KqwFm5447I0C"
      },
      "source": [
        "\n",
        "decoder_inputs = Input(shape=(None, ))\n",
        "#in teacher forcing alwayas correct output is passed to training hence attention model can e applied directly to decoder input\n",
        "#embedding layer:- input layer is vocabulary , output dimension of embedding layer with None batch size, maskzero only if input is of variable length \n",
        "decoder_emb =  Embedding(num_encoder_tokens, latent_dim, mask_zero= True)(decoder_inputs)\n",
        "\n",
        "#reverse of target is not required and eturn sequence is true as we need 3 dimensional target output\n",
        "decoderlstm1= LSTM(latent_dim, return_state=True, activation=\"tanh\", recurrent_activation=\"sigmoid\", return_sequences= True)\n",
        "\n",
        "decodedlstmlayer, _, _=decoderlstm1(decoder_emb, initial_state= [stateh, statec])\n",
        "timedistributedlayer= TimeDistributed(Dense(num_encoder_tokens, activation= 'softmax'))\n",
        "decodedlstmlayer= timedistributedlayer(decodedlstmlayer)\n",
        "#there should no lambda used for attenteion embedding layer in decoder should be replaced with attention layer, keep return sequenc as false for decoder\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ],
      "execution_count": 149,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1oFNUe8KGJ-G"
      },
      "source": [
        "model= Model([encoder_inputs, decoder_inputs], [decodedlstmlayer])\n",
        "\n",
        "model.compile(optimizer= 'adam', loss='categorical_crossentropy', metrics= ['accuracy'])"
      ],
      "execution_count": 150,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 584
        },
        "id": "3_c5AFSkIxUF",
        "outputId": "7dbc7930-77b5-4f76-fcdb-0592c8f0db78"
      },
      "source": [
        "from tensorflow.keras.utils import plot_model\n",
        "plot_model(model, show_shapes=True)\n"
      ],
      "execution_count": 151,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 151
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oNyD3fHj7Xp3"
      },
      "source": [
        "a= Input(shape= (300, ))\n",
        "b= Input(shape= (300, ))\n",
        "combinestates= [a,b]\n",
        "decodedlstmlayer1, h, c=decoderlstm1(decoder_emb, initial_state= combinestates)\n",
        "decodedlstmlayer1= timedistributedlayer(decodedlstmlayer1)\n",
        "combinedoutputstate= [h,c]\n",
        "decodermodel= Model([decoder_inputs] + combinestates, [decodedlstmlayer1] + combinedoutputstate)\n"
      ],
      "execution_count": 152,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 405
        },
        "id": "gcYQ50u5hI3W",
        "outputId": "27ae5426-8a0d-4e66-b88e-bc035ba6487d"
      },
      "source": [
        "\n",
        "plot_model(decodermodel, show_shapes=True)"
      ],
      "execution_count": 153,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 153
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nQeU1QfYhNPX",
        "outputId": "c70918e9-5f85-43f3-f5ec-ab54e256759c"
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": 154,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_4\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_5 (InputLayer)            [(None, None)]       0                                            \n",
            "__________________________________________________________________________________________________\n",
            "embedding_2 (Embedding)         (None, None, 300)    7434900     input_5[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "bidirectional_2 (Bidirectional) [(None, None, 300),  1442400     embedding_2[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "input_6 (InputLayer)            [(None, None)]       0                                            \n",
            "__________________________________________________________________________________________________\n",
            "bidirectional_3 (Bidirectional) [(None, 300), (None, 1442400     bidirectional_2[0][0]            \n",
            "                                                                 bidirectional_2[0][1]            \n",
            "                                                                 bidirectional_2[0][2]            \n",
            "                                                                 bidirectional_2[0][3]            \n",
            "                                                                 bidirectional_2[0][4]            \n",
            "__________________________________________________________________________________________________\n",
            "embedding_3 (Embedding)         (None, None, 300)    7434900     input_6[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "tf.math.add_2 (TFOpLambda)      (None, 300)          0           bidirectional_3[0][1]            \n",
            "                                                                 bidirectional_3[0][3]            \n",
            "__________________________________________________________________________________________________\n",
            "tf.math.add_3 (TFOpLambda)      (None, 300)          0           bidirectional_3[0][2]            \n",
            "                                                                 bidirectional_3[0][4]            \n",
            "__________________________________________________________________________________________________\n",
            "lstm_9 (LSTM)                   [(None, None, 300),  721200      embedding_3[0][0]                \n",
            "                                                                 tf.math.add_2[0][0]              \n",
            "                                                                 tf.math.add_3[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "time_distributed_1 (TimeDistrib (None, None, 24783)  7459683     lstm_9[0][0]                     \n",
            "==================================================================================================\n",
            "Total params: 25,935,483\n",
            "Trainable params: 25,935,483\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PKw6xFHDhQYN"
      },
      "source": [
        "train_samples = len(X)\n",
        "batch_size = 64\n",
        "epochs = 105\n",
        "\n"
      ],
      "execution_count": 155,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "aWMwvLk06Gqt",
        "outputId": "2c632d72-732a-4aee-f2f9-34407b2520e5"
      },
      "source": [
        "train_samples, num_encoder_tokens"
      ],
      "execution_count": 156,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(21776, 24783)"
            ]
          },
          "metadata": {},
          "execution_count": 156
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "z9L574WXAzGw"
      },
      "source": [
        "##Train the model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZGdrrkXNhShq",
        "outputId": "6bd13990-6583-404e-98af-23e619bea9db"
      },
      "source": [
        "from tensorflow.keras.callbacks import EarlyStopping, ModelCheckpoint\n",
        "es= EarlyStopping(monitor= 'loss', mode= 'min', verbose= 1, patience= 10)\n",
        "mc= ModelCheckpoint('/content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5', monitor= 'accuracy', mode= 'max', verbose= 1, save_best_only= True)\n",
        "model.fit_generator(generator = generate_batch(X, y, batch_size = batch_size),steps_per_epoch = train_samples//batch_size,\n",
        "                    epochs=epochs,callbacks= [mc, es])"
      ],
      "execution_count": 157,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/keras/engine/training.py:1972: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n",
            "  warnings.warn('`Model.fit_generator` is deprecated and '\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/105\n",
            "340/340 [==============================] - 87s 212ms/step - loss: 2.4686 - accuracy: 0.1333\n",
            "\n",
            "Epoch 00001: accuracy improved from -inf to 0.13331, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 2/105\n",
            "340/340 [==============================] - 72s 212ms/step - loss: 2.1377 - accuracy: 0.1817\n",
            "\n",
            "Epoch 00002: accuracy improved from 0.13331 to 0.18174, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 3/105\n",
            "340/340 [==============================] - 72s 213ms/step - loss: 2.0032 - accuracy: 0.2098\n",
            "\n",
            "Epoch 00003: accuracy improved from 0.18174 to 0.20981, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 4/105\n",
            "340/340 [==============================] - 72s 213ms/step - loss: 1.9295 - accuracy: 0.2201\n",
            "\n",
            "Epoch 00004: accuracy improved from 0.20981 to 0.22009, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 5/105\n",
            "340/340 [==============================] - 72s 213ms/step - loss: 1.8704 - accuracy: 0.2279\n",
            "\n",
            "Epoch 00005: accuracy improved from 0.22009 to 0.22791, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 6/105\n",
            "340/340 [==============================] - 72s 213ms/step - loss: 1.8183 - accuracy: 0.2339\n",
            "\n",
            "Epoch 00006: accuracy improved from 0.22791 to 0.23392, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 7/105\n",
            "340/340 [==============================] - 72s 212ms/step - loss: 1.7713 - accuracy: 0.2399\n",
            "\n",
            "Epoch 00007: accuracy improved from 0.23392 to 0.23990, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 8/105\n",
            "340/340 [==============================] - 72s 213ms/step - loss: 1.7268 - accuracy: 0.2456\n",
            "\n",
            "Epoch 00008: accuracy improved from 0.23990 to 0.24555, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 9/105\n",
            "340/340 [==============================] - 72s 213ms/step - loss: 1.6835 - accuracy: 0.2516\n",
            "\n",
            "Epoch 00009: accuracy improved from 0.24555 to 0.25155, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 10/105\n",
            "340/340 [==============================] - 73s 214ms/step - loss: 1.6394 - accuracy: 0.2568\n",
            "\n",
            "Epoch 00010: accuracy improved from 0.25155 to 0.25678, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 11/105\n",
            "340/340 [==============================] - 73s 215ms/step - loss: 1.5959 - accuracy: 0.2632\n",
            "\n",
            "Epoch 00011: accuracy improved from 0.25678 to 0.26324, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 12/105\n",
            "340/340 [==============================] - 73s 215ms/step - loss: 1.5515 - accuracy: 0.2719\n",
            "\n",
            "Epoch 00012: accuracy improved from 0.26324 to 0.27190, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 13/105\n",
            "340/340 [==============================] - 73s 216ms/step - loss: 1.5081 - accuracy: 0.2818\n",
            "\n",
            "Epoch 00013: accuracy improved from 0.27190 to 0.28181, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 14/105\n",
            "340/340 [==============================] - 73s 215ms/step - loss: 1.4648 - accuracy: 0.2935\n",
            "\n",
            "Epoch 00014: accuracy improved from 0.28181 to 0.29348, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 15/105\n",
            "340/340 [==============================] - 73s 215ms/step - loss: 1.4231 - accuracy: 0.3057\n",
            "\n",
            "Epoch 00015: accuracy improved from 0.29348 to 0.30571, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 16/105\n",
            "340/340 [==============================] - 74s 216ms/step - loss: 1.3805 - accuracy: 0.3193\n",
            "\n",
            "Epoch 00016: accuracy improved from 0.30571 to 0.31929, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 17/105\n",
            "340/340 [==============================] - 74s 219ms/step - loss: 1.3401 - accuracy: 0.3339\n",
            "\n",
            "Epoch 00017: accuracy improved from 0.31929 to 0.33390, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 18/105\n",
            "340/340 [==============================] - 75s 220ms/step - loss: 1.3012 - accuracy: 0.3469\n",
            "\n",
            "Epoch 00018: accuracy improved from 0.33390 to 0.34687, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 19/105\n",
            "340/340 [==============================] - 75s 220ms/step - loss: 1.2639 - accuracy: 0.3606\n",
            "\n",
            "Epoch 00019: accuracy improved from 0.34687 to 0.36063, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 20/105\n",
            "340/340 [==============================] - 75s 219ms/step - loss: 1.2254 - accuracy: 0.3748\n",
            "\n",
            "Epoch 00020: accuracy improved from 0.36063 to 0.37483, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 21/105\n",
            "340/340 [==============================] - 75s 219ms/step - loss: 1.1891 - accuracy: 0.3887\n",
            "\n",
            "Epoch 00021: accuracy improved from 0.37483 to 0.38871, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 22/105\n",
            "340/340 [==============================] - 75s 219ms/step - loss: 1.1522 - accuracy: 0.4045\n",
            "\n",
            "Epoch 00022: accuracy improved from 0.38871 to 0.40446, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 23/105\n",
            "340/340 [==============================] - 74s 219ms/step - loss: 1.1182 - accuracy: 0.4175\n",
            "\n",
            "Epoch 00023: accuracy improved from 0.40446 to 0.41747, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 24/105\n",
            "340/340 [==============================] - 74s 219ms/step - loss: 1.0841 - accuracy: 0.4313\n",
            "\n",
            "Epoch 00024: accuracy improved from 0.41747 to 0.43132, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 25/105\n",
            "340/340 [==============================] - 74s 219ms/step - loss: 1.0514 - accuracy: 0.4462\n",
            "\n",
            "Epoch 00025: accuracy improved from 0.43132 to 0.44616, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 26/105\n",
            "340/340 [==============================] - 75s 220ms/step - loss: 1.0186 - accuracy: 0.4600\n",
            "\n",
            "Epoch 00026: accuracy improved from 0.44616 to 0.46000, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 27/105\n",
            "340/340 [==============================] - 76s 222ms/step - loss: 0.9871 - accuracy: 0.4736\n",
            "\n",
            "Epoch 00027: accuracy improved from 0.46000 to 0.47363, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 28/105\n",
            "340/340 [==============================] - 75s 222ms/step - loss: 0.9568 - accuracy: 0.4883\n",
            "\n",
            "Epoch 00028: accuracy improved from 0.47363 to 0.48832, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 29/105\n",
            "340/340 [==============================] - 75s 222ms/step - loss: 0.9261 - accuracy: 0.5018\n",
            "\n",
            "Epoch 00029: accuracy improved from 0.48832 to 0.50181, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 30/105\n",
            "340/340 [==============================] - 75s 222ms/step - loss: 0.8934 - accuracy: 0.5182\n",
            "\n",
            "Epoch 00030: accuracy improved from 0.50181 to 0.51823, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 31/105\n",
            "340/340 [==============================] - 75s 220ms/step - loss: 0.8618 - accuracy: 0.5334\n",
            "\n",
            "Epoch 00031: accuracy improved from 0.51823 to 0.53342, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 32/105\n",
            "340/340 [==============================] - 74s 219ms/step - loss: 0.8324 - accuracy: 0.5484\n",
            "\n",
            "Epoch 00032: accuracy improved from 0.53342 to 0.54843, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 33/105\n",
            "340/340 [==============================] - 75s 219ms/step - loss: 0.8034 - accuracy: 0.5626\n",
            "\n",
            "Epoch 00033: accuracy improved from 0.54843 to 0.56259, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 34/105\n",
            "340/340 [==============================] - 75s 220ms/step - loss: 0.7747 - accuracy: 0.5767\n",
            "\n",
            "Epoch 00034: accuracy improved from 0.56259 to 0.57675, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 35/105\n",
            "340/340 [==============================] - 75s 219ms/step - loss: 0.7471 - accuracy: 0.5911\n",
            "\n",
            "Epoch 00035: accuracy improved from 0.57675 to 0.59115, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 36/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.7210 - accuracy: 0.6042\n",
            "\n",
            "Epoch 00036: accuracy improved from 0.59115 to 0.60419, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 37/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.6932 - accuracy: 0.6189\n",
            "\n",
            "Epoch 00037: accuracy improved from 0.60419 to 0.61886, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 38/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.6670 - accuracy: 0.6324\n",
            "\n",
            "Epoch 00038: accuracy improved from 0.61886 to 0.63236, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 39/105\n",
            "340/340 [==============================] - 75s 220ms/step - loss: 0.6437 - accuracy: 0.6444\n",
            "\n",
            "Epoch 00039: accuracy improved from 0.63236 to 0.64439, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 40/105\n",
            "340/340 [==============================] - 75s 219ms/step - loss: 0.6203 - accuracy: 0.6570\n",
            "\n",
            "Epoch 00040: accuracy improved from 0.64439 to 0.65696, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 41/105\n",
            "340/340 [==============================] - 75s 220ms/step - loss: 0.5975 - accuracy: 0.6694\n",
            "\n",
            "Epoch 00041: accuracy improved from 0.65696 to 0.66938, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 42/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.5766 - accuracy: 0.6786\n",
            "\n",
            "Epoch 00042: accuracy improved from 0.66938 to 0.67859, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 43/105\n",
            "340/340 [==============================] - 75s 220ms/step - loss: 0.5550 - accuracy: 0.6918\n",
            "\n",
            "Epoch 00043: accuracy improved from 0.67859 to 0.69181, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 44/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.5353 - accuracy: 0.7023\n",
            "\n",
            "Epoch 00044: accuracy improved from 0.69181 to 0.70227, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 45/105\n",
            "340/340 [==============================] - 75s 220ms/step - loss: 0.5124 - accuracy: 0.7167\n",
            "\n",
            "Epoch 00045: accuracy improved from 0.70227 to 0.71671, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 46/105\n",
            "340/340 [==============================] - 75s 220ms/step - loss: 0.4904 - accuracy: 0.7283\n",
            "\n",
            "Epoch 00046: accuracy improved from 0.71671 to 0.72828, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 47/105\n",
            "340/340 [==============================] - 75s 220ms/step - loss: 0.4714 - accuracy: 0.7390\n",
            "\n",
            "Epoch 00047: accuracy improved from 0.72828 to 0.73903, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 48/105\n",
            "340/340 [==============================] - 75s 220ms/step - loss: 0.4535 - accuracy: 0.7500\n",
            "\n",
            "Epoch 00048: accuracy improved from 0.73903 to 0.75001, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 49/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.4351 - accuracy: 0.7598\n",
            "\n",
            "Epoch 00049: accuracy improved from 0.75001 to 0.75978, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 50/105\n",
            "340/340 [==============================] - 75s 222ms/step - loss: 0.4185 - accuracy: 0.7700\n",
            "\n",
            "Epoch 00050: accuracy improved from 0.75978 to 0.77004, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 51/105\n",
            "340/340 [==============================] - 75s 222ms/step - loss: 0.4021 - accuracy: 0.7789\n",
            "\n",
            "Epoch 00051: accuracy improved from 0.77004 to 0.77892, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 52/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.3837 - accuracy: 0.7901\n",
            "\n",
            "Epoch 00052: accuracy improved from 0.77892 to 0.79015, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 53/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.3672 - accuracy: 0.7992\n",
            "\n",
            "Epoch 00053: accuracy improved from 0.79015 to 0.79920, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 54/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.3532 - accuracy: 0.8067\n",
            "\n",
            "Epoch 00054: accuracy improved from 0.79920 to 0.80674, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 55/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.3377 - accuracy: 0.8168\n",
            "\n",
            "Epoch 00055: accuracy improved from 0.80674 to 0.81681, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 56/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.3240 - accuracy: 0.8239\n",
            "\n",
            "Epoch 00056: accuracy improved from 0.81681 to 0.82393, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 57/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.3105 - accuracy: 0.8320\n",
            "\n",
            "Epoch 00057: accuracy improved from 0.82393 to 0.83201, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 58/105\n",
            "340/340 [==============================] - 75s 222ms/step - loss: 0.2986 - accuracy: 0.8386\n",
            "\n",
            "Epoch 00058: accuracy improved from 0.83201 to 0.83857, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 59/105\n",
            "340/340 [==============================] - 75s 222ms/step - loss: 0.2890 - accuracy: 0.8441\n",
            "\n",
            "Epoch 00059: accuracy improved from 0.83857 to 0.84410, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 60/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.2763 - accuracy: 0.8516\n",
            "\n",
            "Epoch 00060: accuracy improved from 0.84410 to 0.85164, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 61/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.2604 - accuracy: 0.8622\n",
            "\n",
            "Epoch 00061: accuracy improved from 0.85164 to 0.86215, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 62/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.2489 - accuracy: 0.8699\n",
            "\n",
            "Epoch 00062: accuracy improved from 0.86215 to 0.86990, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 63/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.2364 - accuracy: 0.8773\n",
            "\n",
            "Epoch 00063: accuracy improved from 0.86990 to 0.87732, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 64/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.2256 - accuracy: 0.8838\n",
            "\n",
            "Epoch 00064: accuracy improved from 0.87732 to 0.88377, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 65/105\n",
            "340/340 [==============================] - 75s 222ms/step - loss: 0.2163 - accuracy: 0.8890\n",
            "\n",
            "Epoch 00065: accuracy improved from 0.88377 to 0.88902, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 66/105\n",
            "340/340 [==============================] - 76s 222ms/step - loss: 0.2084 - accuracy: 0.8933\n",
            "\n",
            "Epoch 00066: accuracy improved from 0.88902 to 0.89332, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 67/105\n",
            "340/340 [==============================] - 76s 222ms/step - loss: 0.2007 - accuracy: 0.8977\n",
            "\n",
            "Epoch 00067: accuracy improved from 0.89332 to 0.89767, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 68/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.1937 - accuracy: 0.9014\n",
            "\n",
            "Epoch 00068: accuracy improved from 0.89767 to 0.90141, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 69/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.1845 - accuracy: 0.9075\n",
            "\n",
            "Epoch 00069: accuracy improved from 0.90141 to 0.90750, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 70/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.1776 - accuracy: 0.9111\n",
            "\n",
            "Epoch 00070: accuracy improved from 0.90750 to 0.91110, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 71/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.1709 - accuracy: 0.9153\n",
            "\n",
            "Epoch 00071: accuracy improved from 0.91110 to 0.91528, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 72/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.1655 - accuracy: 0.9178\n",
            "\n",
            "Epoch 00072: accuracy improved from 0.91528 to 0.91777, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 73/105\n",
            "340/340 [==============================] - 75s 221ms/step - loss: 0.1585 - accuracy: 0.9221\n",
            "\n",
            "Epoch 00073: accuracy improved from 0.91777 to 0.92209, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 74/105\n",
            "340/340 [==============================] - 75s 222ms/step - loss: 0.1523 - accuracy: 0.9258\n",
            "\n",
            "Epoch 00074: accuracy improved from 0.92209 to 0.92577, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 75/105\n",
            "340/340 [==============================] - 75s 222ms/step - loss: 0.1467 - accuracy: 0.9285\n",
            "\n",
            "Epoch 00075: accuracy improved from 0.92577 to 0.92852, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 76/105\n",
            "340/340 [==============================] - 76s 222ms/step - loss: 0.1412 - accuracy: 0.9318\n",
            "\n",
            "Epoch 00076: accuracy improved from 0.92852 to 0.93184, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 77/105\n",
            "340/340 [==============================] - 75s 222ms/step - loss: 0.1354 - accuracy: 0.9356\n",
            "\n",
            "Epoch 00077: accuracy improved from 0.93184 to 0.93557, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 78/105\n",
            "340/340 [==============================] - 75s 220ms/step - loss: 0.1302 - accuracy: 0.9384\n",
            "\n",
            "Epoch 00078: accuracy improved from 0.93557 to 0.93841, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 79/105\n",
            "340/340 [==============================] - 74s 218ms/step - loss: 0.1259 - accuracy: 0.9409\n",
            "\n",
            "Epoch 00079: accuracy improved from 0.93841 to 0.94091, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 80/105\n",
            "340/340 [==============================] - 75s 219ms/step - loss: 0.1216 - accuracy: 0.9424\n",
            "\n",
            "Epoch 00080: accuracy improved from 0.94091 to 0.94241, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 81/105\n",
            "340/340 [==============================] - 74s 219ms/step - loss: 0.1175 - accuracy: 0.9453\n",
            "\n",
            "Epoch 00081: accuracy improved from 0.94241 to 0.94529, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 82/105\n",
            "340/340 [==============================] - 74s 218ms/step - loss: 0.1141 - accuracy: 0.9470\n",
            "\n",
            "Epoch 00082: accuracy improved from 0.94529 to 0.94696, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 83/105\n",
            "340/340 [==============================] - 74s 218ms/step - loss: 0.1093 - accuracy: 0.9496\n",
            "\n",
            "Epoch 00083: accuracy improved from 0.94696 to 0.94965, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 84/105\n",
            "340/340 [==============================] - 74s 218ms/step - loss: 0.1050 - accuracy: 0.9520\n",
            "\n",
            "Epoch 00084: accuracy improved from 0.94965 to 0.95200, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 85/105\n",
            "340/340 [==============================] - 74s 218ms/step - loss: 0.1013 - accuracy: 0.9543\n",
            "\n",
            "Epoch 00085: accuracy improved from 0.95200 to 0.95426, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 86/105\n",
            "340/340 [==============================] - 74s 218ms/step - loss: 0.0987 - accuracy: 0.9549\n",
            "\n",
            "Epoch 00086: accuracy improved from 0.95426 to 0.95494, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 87/105\n",
            "340/340 [==============================] - 75s 219ms/step - loss: 0.0965 - accuracy: 0.9563\n",
            "\n",
            "Epoch 00087: accuracy improved from 0.95494 to 0.95634, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 88/105\n",
            "340/340 [==============================] - 74s 219ms/step - loss: 0.0930 - accuracy: 0.9578\n",
            "\n",
            "Epoch 00088: accuracy improved from 0.95634 to 0.95781, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 89/105\n",
            "340/340 [==============================] - 75s 220ms/step - loss: 0.0899 - accuracy: 0.9590\n",
            "\n",
            "Epoch 00089: accuracy improved from 0.95781 to 0.95904, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 90/105\n",
            "340/340 [==============================] - 74s 218ms/step - loss: 0.0885 - accuracy: 0.9598\n",
            "\n",
            "Epoch 00090: accuracy improved from 0.95904 to 0.95976, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 91/105\n",
            "340/340 [==============================] - 74s 219ms/step - loss: 0.0853 - accuracy: 0.9618\n",
            "\n",
            "Epoch 00091: accuracy improved from 0.95976 to 0.96183, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 92/105\n",
            "340/340 [==============================] - 74s 218ms/step - loss: 0.0826 - accuracy: 0.9633\n",
            "\n",
            "Epoch 00092: accuracy improved from 0.96183 to 0.96326, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 93/105\n",
            "340/340 [==============================] - 74s 218ms/step - loss: 0.0806 - accuracy: 0.9641\n",
            "\n",
            "Epoch 00093: accuracy improved from 0.96326 to 0.96410, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 94/105\n",
            "340/340 [==============================] - 74s 218ms/step - loss: 0.0792 - accuracy: 0.9650\n",
            "\n",
            "Epoch 00094: accuracy improved from 0.96410 to 0.96501, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 95/105\n",
            "340/340 [==============================] - 74s 219ms/step - loss: 0.0790 - accuracy: 0.9642\n",
            "\n",
            "Epoch 00095: accuracy did not improve from 0.96501\n",
            "Epoch 96/105\n",
            "340/340 [==============================] - 74s 218ms/step - loss: 0.0769 - accuracy: 0.9649\n",
            "\n",
            "Epoch 00096: accuracy did not improve from 0.96501\n",
            "Epoch 97/105\n",
            "340/340 [==============================] - 74s 218ms/step - loss: 0.0750 - accuracy: 0.9662\n",
            "\n",
            "Epoch 00097: accuracy improved from 0.96501 to 0.96622, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 98/105\n",
            "340/340 [==============================] - 74s 219ms/step - loss: 0.0728 - accuracy: 0.9673\n",
            "\n",
            "Epoch 00098: accuracy improved from 0.96622 to 0.96731, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 99/105\n",
            "340/340 [==============================] - 74s 218ms/step - loss: 0.0701 - accuracy: 0.9688\n",
            "\n",
            "Epoch 00099: accuracy improved from 0.96731 to 0.96877, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 100/105\n",
            "340/340 [==============================] - 74s 218ms/step - loss: 0.0693 - accuracy: 0.9693\n",
            "\n",
            "Epoch 00100: accuracy improved from 0.96877 to 0.96933, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 101/105\n",
            "340/340 [==============================] - 74s 217ms/step - loss: 0.0687 - accuracy: 0.9686\n",
            "\n",
            "Epoch 00101: accuracy did not improve from 0.96933\n",
            "Epoch 102/105\n",
            "340/340 [==============================] - 74s 217ms/step - loss: 0.0669 - accuracy: 0.9697\n",
            "\n",
            "Epoch 00102: accuracy improved from 0.96933 to 0.96968, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 103/105\n",
            "340/340 [==============================] - 74s 218ms/step - loss: 0.0665 - accuracy: 0.9699\n",
            "\n",
            "Epoch 00103: accuracy improved from 0.96968 to 0.96986, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 104/105\n",
            "340/340 [==============================] - 74s 218ms/step - loss: 0.0645 - accuracy: 0.9708\n",
            "\n",
            "Epoch 00104: accuracy improved from 0.96986 to 0.97083, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n",
            "Epoch 105/105\n",
            "340/340 [==============================] - 74s 219ms/step - loss: 0.0622 - accuracy: 0.9719\n",
            "\n",
            "Epoch 00105: accuracy improved from 0.97083 to 0.97192, saving model to /content/drive/MyDrive/Colab Notebooks/chatbot/chatbot.h5\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7f121c4f2450>"
            ]
          },
          "metadata": {},
          "execution_count": 157
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E41zfzKjhYij"
      },
      "source": [
        "#iuhbjhv"
      ],
      "execution_count": 158,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ksnf2P_RNP2E"
      },
      "source": [
        ""
      ],
      "execution_count": 158,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FwjFtBeRhX1T"
      },
      "source": [
        "model.save_weights('/content/drive/MyDrive/Colab Notebooks/chatbot/nmt_weights.h5')"
      ],
      "execution_count": 159,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZMQszPVcJvGy",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b02c9cc5-4eca-4055-fff5-6be80cdf55d2"
      },
      "source": [
        "model_json = encodermodel.to_json()\n",
        "with open(\"/content/drive/MyDrive/Colab Notebooks/chatbot/encodermodel1.json\", \"w\") as json_file:\n",
        "    json_file.write(model_json)\n",
        "\n",
        "encodermodel.save_weights(\"/content/drive/MyDrive/Colab Notebooks/chatbot/encoder_model1.h5\")\n",
        "print(\"Saved Model!\")"
      ],
      "execution_count": 160,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Saved Model!\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "F-qKYUBaJvKh",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2bbc9b95-c3a6-4209-c601-b5fa8c28c6e3"
      },
      "source": [
        "model_json = decodermodel.to_json()\n",
        "with open(\"/content/drive/MyDrive/Colab Notebooks/chatbot/decodermodel1.json\", \"w\") as json_file:\n",
        "    json_file.write(model_json)\n",
        "\n",
        "decodermodel.save_weights(\"/content/drive/MyDrive/Colab Notebooks/chatbot/decoder_model.h5\")\n",
        "print(\"Saved Model!\")"
      ],
      "execution_count": 168,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Saved Model!\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SbNtkWJAJvcC"
      },
      "source": [
        "def decode_sequence(input_seq):\n",
        "    # Encode the input as state vectors.\n",
        "    states_value = encodermodel.predict(input_seq)\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] = input_token_index['START_']\n",
        "\n",
        "    # Sampling loop for a batch of sequences\n",
        "    # (to simplify, here we assume a batch of size 1).\n",
        "    stop_condition = False\n",
        "    decoded_sentence = ''\n",
        "    while not stop_condition:\n",
        "        output_tokens, h, c = decodermodel.predict([target_seq] + states_value)\n",
        "\n",
        "        # Sample a token/ it used last time stamp which is only one so doenst matter\n",
        "        sampled_token_index = np.argmax(output_tokens[0, -1, :])\n",
        "        sampled_char = reverse_input_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) > 20):\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"
      ],
      "execution_count": 162,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WhEDyxTcJzpA"
      },
      "source": [
        "train_gen = generate_batch(X, y, batch_size = 1)\n",
        "k=-1\n"
      ],
      "execution_count": 163,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TMX5JndkpOLD",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d07a354d-0952-41a0-d8c2-89c545354315"
      },
      "source": [
        "k+=1\n",
        "(input_seq, actual_output), _ = next(train_gen)\n",
        "decoded_sentence = decode_sequence(input_seq)\n",
        "print('Input  sentence:', X[k:k+1].values[0])\n",
        "print('Actual  conversation:', y[k:k+1].values[0][6:-4])\n",
        "print('Predicted word:', decoded_sentence[:-4])"
      ],
      "execution_count": 164,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Input  sentence: forget it bunghole\n",
            "Actual  conversation:  me first \n",
            "Predicted word:  me first \n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EiSt1qIC5t8v"
      },
      "source": [
        "\n"
      ],
      "execution_count": 169,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hWqVBboK51dY"
      },
      "source": [
        "def chatbot(cases):\n",
        "\n",
        "  for i in range(cases):\n",
        "    stringss= input('START THE CONVERSATION::::')\n",
        "    stringss= clean_text(stringss)\n",
        "    stringss= ' '.join(stringss.split())\n",
        "    stringss= stringss.strip()\n",
        "    newdata= np.zeros((1, max_length_src),dtype='float32')\n",
        "    for t, word in enumerate(stringss.split()):\n",
        "      newdata[0, t] = input_token_index[word]\n",
        "\n",
        "    decoded_sentence = decode_sequence(newdata)\n",
        "    print('CHATBOT:::::', decoded_sentence[:-4])\n"
      ],
      "execution_count": 179,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_sP4Rv3E7P1A",
        "outputId": "dcc73f28-7d75-439c-d261-d706247ccb4e"
      },
      "source": [
        "chatbot(4)\n",
        "\n",
        "\n",
        "\n"
      ],
      "execution_count": 181,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "START THE CONVERSATION::::good night\n",
            "CHATBOT:::::  driving is not the probl\n",
            "START THE CONVERSATION::::why\n",
            "CHATBOT:::::  i am really afraid \n",
            "START THE CONVERSATION::::where are you\n",
            "CHATBOT:::::  go to the back door\n",
            "START THE CONVERSATION::::why\n",
            "CHATBOT:::::  i am really afraid \n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "N7B3EpIz7Z9w"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}