Edureka_Sequence_Learning_Assignment_4_Machine_Translation.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.4"
    },
    "colab": {
      "name": "Edureka Sequence Learning Assignment 4 - Machine Translation.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/nandinib1999/edureka-assignments/blob/main/Edureka_Sequence_Learning_Assignment_4_Machine_Translation.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "miHk-78pJ0q9",
        "outputId": "f1f71f07-2389-47ee-aa67-5571be931f25"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
        "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
        "id": "LDrrjKGwJoeS"
      },
      "source": [
        "import numpy as np \n",
        "import pandas as pd\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 keras.layers import Input, LSTM, Embedding, Dense\n",
        "from keras.models import Model"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hNTWnWz6R0h6"
      },
      "source": [
        "#### Load the dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "_cell_guid": "79c7e3d0-c299-4dcb-8224-4455121ee9b0",
        "_uuid": "d629ff2d2480ee46fbb7e2d37f6b5fab8052498a",
        "id": "ap5Id4x3Joeh"
      },
      "source": [
        "df=pd.read_csv(\"/content/drive/MyDrive/Hindi_English_Truncated_Corpus.csv\",encoding='utf-8')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8uSOM1bOJoei",
        "outputId": "31d74344-5172-4100-c80d-ecefe958c8ba"
      },
      "source": [
        "df['source'].value_counts()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tides        50000\n",
              "ted          39881\n",
              "indic2012    37726\n",
              "Name: source, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "73L_X3_NJoej"
      },
      "source": [
        "df=df[df['source']=='ted']"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 669
        },
        "id": "LfODyHwpJoej",
        "outputId": "53ff7354-1e4a-4a53-ce10-6b7fc3694a55"
      },
      "source": [
        "df.head(20)"
      ],
      "execution_count": null,
      "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>source</th>\n",
              "      <th>english_sentence</th>\n",
              "      <th>hindi_sentence</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>ted</td>\n",
              "      <td>politicians do not have permission to do what ...</td>\n",
              "      <td>राजनीतिज्ञों के पास जो कार्य करना चाहिए, वह कर...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>ted</td>\n",
              "      <td>I'd like to tell you about one such child,</td>\n",
              "      <td>मई आपको ऐसे ही एक बच्चे के बारे में बताना चाहू...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>ted</td>\n",
              "      <td>what we really mean is that they're bad at not...</td>\n",
              "      <td>हम ये नहीं कहना चाहते कि वो ध्यान नहीं दे पाते</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>ted</td>\n",
              "      <td>And who are we to say, even, that they are wrong</td>\n",
              "      <td>और हम होते कौन हैं यह कहने भी वाले कि वे गलत हैं</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>ted</td>\n",
              "      <td>So there is some sort of justice</td>\n",
              "      <td>तो वहाँ न्याय है</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>23</th>\n",
              "      <td>ted</td>\n",
              "      <td>This changed slowly</td>\n",
              "      <td>धीरे धीरे ये सब बदला</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>26</th>\n",
              "      <td>ted</td>\n",
              "      <td>were being produced.</td>\n",
              "      <td>उत्पन्न नहीं कि जाती थी.</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>30</th>\n",
              "      <td>ted</td>\n",
              "      <td>And you can see, this LED is going to glow.</td>\n",
              "      <td>और जैसा आप देख रहे है, ये एल.ई.डी. जल उठेगी।</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>32</th>\n",
              "      <td>ted</td>\n",
              "      <td>to turn on the lights or to bring him a glass ...</td>\n",
              "      <td>लाईट जलाने के लिए या उनके लिए पानी लाने के लिए,</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>35</th>\n",
              "      <td>ted</td>\n",
              "      <td>Can you imagine saying that?</td>\n",
              "      <td>क्या आप ये कल्पना कर सकते है</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>37</th>\n",
              "      <td>ted</td>\n",
              "      <td>Three: this is a good road in - right near whe...</td>\n",
              "      <td>तीसरी: ये हमारी फ़ैक्ट्री के पास की एक अपेक्षा...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>39</th>\n",
              "      <td>ted</td>\n",
              "      <td>What's going on?”</td>\n",
              "      <td>क्या हो रहा है ये?”</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>42</th>\n",
              "      <td>ted</td>\n",
              "      <td>There are also financial reforms in rural China.</td>\n",
              "      <td>ग्रामीण चीन में आर्थिक नवीनीकरण हुये हैं।</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>49</th>\n",
              "      <td>ted</td>\n",
              "      <td>the family planning started in Vietnam and the...</td>\n",
              "      <td>वियतनाम में परिवार योजना शुरू हो गई और उनके पर...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>51</th>\n",
              "      <td>ted</td>\n",
              "      <td>I mean, at that time, trust me,</td>\n",
              "      <td>मेरा मतलब, उस समय, सही मानिए,</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>53</th>\n",
              "      <td>ted</td>\n",
              "      <td>Not only that,</td>\n",
              "      <td>बस वही नहीं,</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>55</th>\n",
              "      <td>ted</td>\n",
              "      <td>humans destroyed the commons that they depende...</td>\n",
              "      <td>मानवों ने उन ही साझे संसाधनों को नष्ट किया जिन...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>56</th>\n",
              "      <td>ted</td>\n",
              "      <td>Almost goes to E, but otherwise the play would...</td>\n",
              "      <td>रचना करीब करीब ई तक जाती है, मगर तब तो नाटक ख़...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>63</th>\n",
              "      <td>ted</td>\n",
              "      <td>So I want to share with you a couple key insights</td>\n",
              "      <td>मैं आपके साथ कुछ मुख्य सूत्र बाँटना चाहता हूँ</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>66</th>\n",
              "      <td>ted</td>\n",
              "      <td>Many countries in the [unclear], they need leg...</td>\n",
              "      <td>[अस्पष्ट] के बहुत सारे राष्ट्रों को मान्यता चा...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   source  ...                                     hindi_sentence\n",
              "0     ted  ...  राजनीतिज्ञों के पास जो कार्य करना चाहिए, वह कर...\n",
              "1     ted  ...  मई आपको ऐसे ही एक बच्चे के बारे में बताना चाहू...\n",
              "3     ted  ...     हम ये नहीं कहना चाहते कि वो ध्यान नहीं दे पाते\n",
              "7     ted  ...   और हम होते कौन हैं यह कहने भी वाले कि वे गलत हैं\n",
              "13    ted  ...                                   तो वहाँ न्याय है\n",
              "23    ted  ...                               धीरे धीरे ये सब बदला\n",
              "26    ted  ...                           उत्पन्न नहीं कि जाती थी.\n",
              "30    ted  ...       और जैसा आप देख रहे है, ये एल.ई.डी. जल उठेगी।\n",
              "32    ted  ...    लाईट जलाने के लिए या उनके लिए पानी लाने के लिए,\n",
              "35    ted  ...                       क्या आप ये कल्पना कर सकते है\n",
              "37    ted  ...  तीसरी: ये हमारी फ़ैक्ट्री के पास की एक अपेक्षा...\n",
              "39    ted  ...                                क्या हो रहा है ये?”\n",
              "42    ted  ...          ग्रामीण चीन में आर्थिक नवीनीकरण हुये हैं।\n",
              "49    ted  ...  वियतनाम में परिवार योजना शुरू हो गई और उनके पर...\n",
              "51    ted  ...                      मेरा मतलब, उस समय, सही मानिए,\n",
              "53    ted  ...                                       बस वही नहीं,\n",
              "55    ted  ...  मानवों ने उन ही साझे संसाधनों को नष्ट किया जिन...\n",
              "56    ted  ...  रचना करीब करीब ई तक जाती है, मगर तब तो नाटक ख़...\n",
              "63    ted  ...      मैं आपके साथ कुछ मुख्य सूत्र बाँटना चाहता हूँ\n",
              "66    ted  ...  [अस्पष्ट] के बहुत सारे राष्ट्रों को मान्यता चा...\n",
              "\n",
              "[20 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zTzLf2PbSETK"
      },
      "source": [
        "#### Exploring the data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Q6kh3R2qJoel",
        "outputId": "bbce24ef-e6df-4b9c-e009-4d6dae0a6564"
      },
      "source": [
        "df.isna().sum()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "source              0\n",
              "english_sentence    0\n",
              "hindi_sentence      0\n",
              "dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "R--nA-LlJoem",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "63420af6-8d80-4285-c7a5-0b79c940d38e"
      },
      "source": [
        "df = df[~df.isna().any(axis=1)]\n",
        "df.info()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "Int64Index: 39881 entries, 0 to 127606\n",
            "Data columns (total 3 columns):\n",
            " #   Column            Non-Null Count  Dtype \n",
            "---  ------            --------------  ----- \n",
            " 0   source            39881 non-null  object\n",
            " 1   english_sentence  39881 non-null  object\n",
            " 2   hindi_sentence    39881 non-null  object\n",
            "dtypes: object(3)\n",
            "memory usage: 1.2+ MB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ln_vlL4tJoen",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "449b0a29-db32-469f-8a58-007089e97102"
      },
      "source": [
        "df.drop_duplicates(inplace=True)\n",
        "df.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(38803, 3)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9xZLme10Joeo",
        "outputId": "3a2a8ba6-a686-4d77-c2c7-1cb1531539a0"
      },
      "source": [
        "df=df.sample(n=25000,random_state=42)\n",
        "df.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(25000, 3)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "g4xoK9UlSIJi"
      },
      "source": [
        "#### Processing the data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Cs4yV8G_Joep"
      },
      "source": [
        "# Lowercase all characters\n",
        "df['english_sentence']=df['english_sentence'].apply(lambda x: x.lower())\n",
        "df['hindi_sentence']=df['hindi_sentence'].apply(lambda x: x.lower())\n",
        "\n",
        "# Remove quotes\n",
        "df['english_sentence']=df['english_sentence'].apply(lambda x: re.sub(\"'\", '', x))\n",
        "df['hindi_sentence']=df['hindi_sentence'].apply(lambda x: re.sub(\"'\", '', x))\n",
        "\n",
        "exclude = set(string.punctuation) # Set of all special characters\n",
        "# Remove all the special characters\n",
        "df['english_sentence']=df['english_sentence'].apply(lambda x: ''.join(ch for ch in x if ch not in exclude))\n",
        "df['hindi_sentence']=df['hindi_sentence'].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",
        "df['english_sentence']=df['english_sentence'].apply(lambda x: x.translate(remove_digits))\n",
        "df['hindi_sentence']=df['hindi_sentence'].apply(lambda x: x.translate(remove_digits))\n",
        "\n",
        "df['hindi_sentence'] = df['hindi_sentence'].apply(lambda x: re.sub(\"[२३०८१५७९४६]\", \"\", x))\n",
        "\n",
        "# Remove extra spaces\n",
        "df['english_sentence']=df['english_sentence'].apply(lambda x: x.strip())\n",
        "df['hindi_sentence']=df['hindi_sentence'].apply(lambda x: x.strip())\n",
        "df['english_sentence']=df['english_sentence'].apply(lambda x: re.sub(\" +\", \" \", x))\n",
        "df['hindi_sentence']=df['hindi_sentence'].apply(lambda x: re.sub(\" +\", \" \", x))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bk8PRH_SJoer"
      },
      "source": [
        "# Add start and end tokens to target sequences\n",
        "df['hindi_sentence'] = df['hindi_sentence'].apply(lambda x : 'START_ '+ x + ' _END')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dxh2UUHlVcOH",
        "outputId": "d1f40ec8-bd3c-4891-ae2d-0e2ece7b1a29"
      },
      "source": [
        "df.isna().sum()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "source              0\n",
              "english_sentence    0\n",
              "hindi_sentence      0\n",
              "dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "x0dWYAsDVe_b",
        "outputId": "b773e8bb-dceb-4e24-ae4c-bfb7eab4370b"
      },
      "source": [
        "df = df[~df.isna().any(axis=1)]\n",
        "df.info()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "Int64Index: 25000 entries, 82040 to 91889\n",
            "Data columns (total 3 columns):\n",
            " #   Column            Non-Null Count  Dtype \n",
            "---  ------            --------------  ----- \n",
            " 0   source            25000 non-null  object\n",
            " 1   english_sentence  25000 non-null  object\n",
            " 2   hindi_sentence    25000 non-null  object\n",
            "dtypes: object(3)\n",
            "memory usage: 781.2+ KB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "AOkUw1uxJoer",
        "outputId": "ab06fd02-a673-4210-e71b-35329ab75fb7"
      },
      "source": [
        "df.head()"
      ],
      "execution_count": null,
      "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>source</th>\n",
              "      <th>english_sentence</th>\n",
              "      <th>hindi_sentence</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>82040</th>\n",
              "      <td>ted</td>\n",
              "      <td>we still dont know who her parents are who she is</td>\n",
              "      <td>START_ हम अभी तक नहीं जानते हैं कि उसके मातापि...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>85038</th>\n",
              "      <td>ted</td>\n",
              "      <td>no keyboard</td>\n",
              "      <td>START_ कोई कुंजीपटल नहीं _END</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>58018</th>\n",
              "      <td>ted</td>\n",
              "      <td>but as far as being a performer</td>\n",
              "      <td>START_ लेकिन एक कलाकार होने के साथ _END</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>74470</th>\n",
              "      <td>ted</td>\n",
              "      <td>and this particular balloon</td>\n",
              "      <td>START_ और यह खास गुब्बारा _END</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>122330</th>\n",
              "      <td>ted</td>\n",
              "      <td>and its not as hard as you think integrate cli...</td>\n",
              "      <td>START_ और जितना आपको लगता है यह उतना कठिन नहीं...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "       source  ...                                     hindi_sentence\n",
              "82040     ted  ...  START_ हम अभी तक नहीं जानते हैं कि उसके मातापि...\n",
              "85038     ted  ...                      START_ कोई कुंजीपटल नहीं _END\n",
              "58018     ted  ...            START_ लेकिन एक कलाकार होने के साथ _END\n",
              "74470     ted  ...                     START_ और यह खास गुब्बारा _END\n",
              "122330    ted  ...  START_ और जितना आपको लगता है यह उतना कठिन नहीं...\n",
              "\n",
              "[5 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZtRDkKSFGAb4",
        "outputId": "f05ddbac-88b6-4224-bd2c-bd58ae7e898f"
      },
      "source": [
        "eng_vocab = []\n",
        "eng_sentences = df['english_sentence'].tolist()\n",
        "eng_maxlen = 0\n",
        "for sent in eng_sentences:\n",
        "  words = sent.split(\" \")\n",
        "  words = [x.strip() for x in words if len(x.strip()) > 0]\n",
        "  if eng_maxlen < len(words):\n",
        "    eng_maxlen = len(words)\n",
        "  eng_vocab.extend(words)\n",
        "\n",
        "\n",
        "hindi_vocab = []\n",
        "hindi_sentences = df['hindi_sentence'].tolist()\n",
        "hindi_maxlen = 0\n",
        "for sent in hindi_sentences:\n",
        "  words = sent.split(\" \")\n",
        "  words = [x.strip() for x in words if len(x.strip()) > 0]\n",
        "  if hindi_maxlen < len(words):\n",
        "    hindi_maxlen = len(words)\n",
        "  hindi_vocab.extend(words)\n",
        "\n",
        "\n",
        "all_eng_words = list(set(eng_vocab))\n",
        "all_hindi_words = list(set(hindi_vocab))\n",
        "\n",
        "print(\"English Vocab:\", len(all_eng_words))\n",
        "print(\"Hindi Vocab:\", len(all_hindi_words))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "English Vocab: 14030\n",
            "Hindi Vocab: 17540\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Ar3kiGu5Joes",
        "outputId": "28841dfd-dcf6-4d18-e63a-9b3c40d2e12b"
      },
      "source": [
        "len(all_eng_words)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "14030"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DJ0k0XndJoet",
        "outputId": "5d44fe49-191a-4715-e75c-61f9a1594695"
      },
      "source": [
        "len(all_hindi_words)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "17540"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pqJeK_YPGTIQ"
      },
      "source": [
        "from collections import Counter\n",
        "\n",
        "def word_count(vocabulary):\n",
        "    ct = Counter(vocabulary)\n",
        "    words = []\n",
        "    cnt = []\n",
        "    for i, j in ct.items():\n",
        "      if i != \"START_\" and i != \"_END\":\n",
        "        words.append(i)\n",
        "        cnt.append(j)\n",
        "    data = {\"word\":words , \"count\":cnt}\n",
        "    df = pd.DataFrame(data)\n",
        "    df = df.sort_values(by='count', ascending=False)\n",
        "    df = df.reset_index()[[\"word\",\"count\"]]\n",
        "    return df"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 669
        },
        "id": "Sda_Vta5GVOy",
        "outputId": "aff9a9a3-6e13-46e7-da92-cb9cf98bb54d"
      },
      "source": [
        "word_cnt = word_count(eng_vocab)\n",
        "word_cnt.head(20)"
      ],
      "execution_count": null,
      "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>word</th>\n",
              "      <th>count</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>the</td>\n",
              "      <td>9103</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>and</td>\n",
              "      <td>6268</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>to</td>\n",
              "      <td>5447</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>of</td>\n",
              "      <td>4968</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>a</td>\n",
              "      <td>4609</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>that</td>\n",
              "      <td>3560</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>in</td>\n",
              "      <td>3442</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>i</td>\n",
              "      <td>3059</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>is</td>\n",
              "      <td>2847</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>you</td>\n",
              "      <td>2813</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>we</td>\n",
              "      <td>2343</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>it</td>\n",
              "      <td>2258</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>this</td>\n",
              "      <td>2067</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>so</td>\n",
              "      <td>1601</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>was</td>\n",
              "      <td>1324</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>for</td>\n",
              "      <td>1278</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16</th>\n",
              "      <td>they</td>\n",
              "      <td>1266</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17</th>\n",
              "      <td>have</td>\n",
              "      <td>1259</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>are</td>\n",
              "      <td>1258</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>but</td>\n",
              "      <td>1110</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    word  count\n",
              "0    the   9103\n",
              "1    and   6268\n",
              "2     to   5447\n",
              "3     of   4968\n",
              "4      a   4609\n",
              "5   that   3560\n",
              "6     in   3442\n",
              "7      i   3059\n",
              "8     is   2847\n",
              "9    you   2813\n",
              "10    we   2343\n",
              "11    it   2258\n",
              "12  this   2067\n",
              "13    so   1601\n",
              "14   was   1324\n",
              "15   for   1278\n",
              "16  they   1266\n",
              "17  have   1259\n",
              "18   are   1258\n",
              "19   but   1110"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 20
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 380
        },
        "id": "JEWX6yVXGeTt",
        "outputId": "9dc1d633-3a73-4200-db82-9791641e4c29"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "topn = 50\n",
        "\n",
        "def plot_word_count(df, title):\n",
        "    plt.figure(figsize=(30,3))\n",
        "    plt.bar(df.index, df[\"count\"])\n",
        "    plt.yticks(fontsize=20)\n",
        "    plt.xticks(df.index,df[\"word\"], rotation=90, fontsize=20)\n",
        "    plt.title(title, fontsize=20)\n",
        "    plt.show()\n",
        "\n",
        "plot_word_count(word_cnt[:topn], title=\"The top 20 most frequently appearing words\")\n",
        "print()\n",
        "plot_word_count(word_cnt[-topn:], title=\"The least 20 most frequently appearing words\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 2160x216 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 2160x216 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "t184w1UOJoet",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 890
        },
        "outputId": "c0f3788a-b0cb-4da1-d34a-0b7774013532"
      },
      "source": [
        "df['eng_sent_len'] = df['english_sentence'].apply(lambda x:len(x.split(\" \")))\n",
        "df['hindi_sent_len'] = df['hindi_sentence'].apply(lambda x:len(x.split(\" \")))\n",
        "df.head(20)"
      ],
      "execution_count": null,
      "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>source</th>\n",
              "      <th>english_sentence</th>\n",
              "      <th>hindi_sentence</th>\n",
              "      <th>eng_sent_len</th>\n",
              "      <th>hindi_sent_len</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>82040</th>\n",
              "      <td>ted</td>\n",
              "      <td>we still dont know who her parents are who she is</td>\n",
              "      <td>START_ हम अभी तक नहीं जानते हैं कि उसके मातापि...</td>\n",
              "      <td>11</td>\n",
              "      <td>16</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>85038</th>\n",
              "      <td>ted</td>\n",
              "      <td>no keyboard</td>\n",
              "      <td>START_ कोई कुंजीपटल नहीं _END</td>\n",
              "      <td>2</td>\n",
              "      <td>5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>58018</th>\n",
              "      <td>ted</td>\n",
              "      <td>but as far as being a performer</td>\n",
              "      <td>START_ लेकिन एक कलाकार होने के साथ _END</td>\n",
              "      <td>7</td>\n",
              "      <td>8</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>74470</th>\n",
              "      <td>ted</td>\n",
              "      <td>and this particular balloon</td>\n",
              "      <td>START_ और यह खास गुब्बारा _END</td>\n",
              "      <td>4</td>\n",
              "      <td>6</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>122330</th>\n",
              "      <td>ted</td>\n",
              "      <td>and its not as hard as you think integrate cli...</td>\n",
              "      <td>START_ और जितना आपको लगता है यह उतना कठिन नहीं...</td>\n",
              "      <td>16</td>\n",
              "      <td>20</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>79517</th>\n",
              "      <td>ted</td>\n",
              "      <td>and saw the demo by jeff han</td>\n",
              "      <td>START_ और जेफ़ हान द्वारा प्रदर्शन देखा होगा _END</td>\n",
              "      <td>7</td>\n",
              "      <td>9</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>23089</th>\n",
              "      <td>ted</td>\n",
              "      <td>this baby is fully electric</td>\n",
              "      <td>START_ यह बच्चा पूरी तरह से बिजली से चलता है _END</td>\n",
              "      <td>5</td>\n",
              "      <td>11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>116699</th>\n",
              "      <td>ted</td>\n",
              "      <td>kids have no or very little say in making the ...</td>\n",
              "      <td>START_ फिर भी बच्चों को नियम बनाने का बिलकुल न...</td>\n",
              "      <td>11</td>\n",
              "      <td>17</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>64132</th>\n",
              "      <td>ted</td>\n",
              "      <td>im going to add a little bit to my description...</td>\n",
              "      <td>START_ मै अपने द्वारा दिए गए उम्र के बढ़्ने के...</td>\n",
              "      <td>12</td>\n",
              "      <td>16</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>74513</th>\n",
              "      <td>ted</td>\n",
              "      <td>expands and cools until it gets to the point w...</td>\n",
              "      <td>START_ फैलने लगता है फिर ये ठंडा होकर उस अवस्थ...</td>\n",
              "      <td>13</td>\n",
              "      <td>21</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6311</th>\n",
              "      <td>ted</td>\n",
              "      <td>by making it seamless</td>\n",
              "      <td>START_ इसे निरंतर बना कर _END</td>\n",
              "      <td>4</td>\n",
              "      <td>6</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>41215</th>\n",
              "      <td>ted</td>\n",
              "      <td>its perceptions thoughts feelings dreams</td>\n",
              "      <td>START_ इसके अपने दृष्टिकोण विचार हैं अपननी भाव...</td>\n",
              "      <td>5</td>\n",
              "      <td>14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25876</th>\n",
              "      <td>ted</td>\n",
              "      <td>so imagine you have carbon nanotubes growing</td>\n",
              "      <td>START_ तो आप कल्पना करे की कार्बन नैनोट्यूब बड...</td>\n",
              "      <td>7</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>27454</th>\n",
              "      <td>ted</td>\n",
              "      <td>falling down and disintegrating</td>\n",
              "      <td>START_ ऊपर से गिरकर बिखर गया _END</td>\n",
              "      <td>4</td>\n",
              "      <td>7</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>98063</th>\n",
              "      <td>ted</td>\n",
              "      <td>so when this came up in my local shop naturall...</td>\n",
              "      <td>START_ तो स्वाभाविक रूप से जब यह अपने स्थानीय ...</td>\n",
              "      <td>12</td>\n",
              "      <td>14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>91972</th>\n",
              "      <td>ted</td>\n",
              "      <td>but then how to act compassionately if you don...</td>\n",
              "      <td>START_ पर आप सहानुभूति पूर्वक कार्य कैसे करोगे...</td>\n",
              "      <td>11</td>\n",
              "      <td>15</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8657</th>\n",
              "      <td>ted</td>\n",
              "      <td>its limited by our cognitive biases</td>\n",
              "      <td>START_ ये बंधा हुआ है हमारे दिमागी पक्षपात से ...</td>\n",
              "      <td>6</td>\n",
              "      <td>10</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>82742</th>\n",
              "      <td>ted</td>\n",
              "      <td>were going to have about eight billion perhaps...</td>\n",
              "      <td>START_ संसार में अरब या उस से भी ज्यादा लोग हो...</td>\n",
              "      <td>10</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>74821</th>\n",
              "      <td>ted</td>\n",
              "      <td>and everyone knew the truth</td>\n",
              "      <td>START_ और सब को सच्चाई का पता चल गया _END</td>\n",
              "      <td>5</td>\n",
              "      <td>10</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>60006</th>\n",
              "      <td>ted</td>\n",
              "      <td>if i give you the letter q</td>\n",
              "      <td>START_ यदि मैं आपको एक वर्ण q दूँ _END</td>\n",
              "      <td>7</td>\n",
              "      <td>9</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "       source  ... hindi_sent_len\n",
              "82040     ted  ...             16\n",
              "85038     ted  ...              5\n",
              "58018     ted  ...              8\n",
              "74470     ted  ...              6\n",
              "122330    ted  ...             20\n",
              "79517     ted  ...              9\n",
              "23089     ted  ...             11\n",
              "116699    ted  ...             17\n",
              "64132     ted  ...             16\n",
              "74513     ted  ...             21\n",
              "6311      ted  ...              6\n",
              "41215     ted  ...             14\n",
              "25876     ted  ...             12\n",
              "27454     ted  ...              7\n",
              "98063     ted  ...             14\n",
              "91972     ted  ...             15\n",
              "8657      ted  ...             10\n",
              "82742     ted  ...             12\n",
              "74821     ted  ...             10\n",
              "60006     ted  ...              9\n",
              "\n",
              "[20 rows x 5 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 22
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "cQHOIJgwULdb",
        "outputId": "09a342a5-4f4a-4af1-e7db-d9c9dbe5e993"
      },
      "source": [
        "plt.boxplot(df['eng_sent_len'])\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "4IYLm13fUdR6",
        "outputId": "ade151de-ac81-4350-b4aa-70f24be198fe"
      },
      "source": [
        "plt.boxplot(df['hindi_sent_len'])\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "U557IXS7Joeu"
      },
      "source": [
        "df=df[df['eng_sent_len']<=20]\n",
        "df=df[df['hindi_sent_len']<=20]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IxLYaJabJoeu",
        "outputId": "848c29b3-edc2-4359-fcf3-34f2e8c2827b"
      },
      "source": [
        "df.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(24774, 5)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "CySIKcLfUrOg",
        "outputId": "d4867959-6db5-4bf7-e468-bc6b0dcea628"
      },
      "source": [
        "plt.boxplot(df['eng_sent_len'])\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "N_2z2ht3Un62",
        "outputId": "727fd592-0e43-4eca-c31a-01548bb19fe7"
      },
      "source": [
        "plt.boxplot(df['hindi_sent_len'])\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "plCt-NLtUzVv",
        "outputId": "1a315991-e73c-4dd2-ece7-d586ba878edd"
      },
      "source": [
        "plt.hist(df['eng_sent_len'])\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "OntMrXR3Uumi",
        "outputId": "80c9dfea-6508-443f-fd7e-f48b25913c32"
      },
      "source": [
        "plt.hist(df['hindi_sent_len'])\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gLB-FgP4VDbR",
        "outputId": "99f5e052-82c6-479d-9895-7dc4ff9dd4b9"
      },
      "source": [
        "max_length_src = max(df['hindi_sent_len'])\n",
        "max_length_tar = max(df['eng_sent_len'])\n",
        "\n",
        "print(\"Maximum length of Hindi Sentence:\", max_length_src)\n",
        "print(\"Maximum length of English Sentence:\", max_length_tar)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Maximum length of Hindi Sentence: 20\n",
            "Maximum length of English Sentence: 20\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zCgDpMWZJoev",
        "outputId": "e568eb81-9fd2-45ad-8c1f-ee41b6e1e811"
      },
      "source": [
        "input_words = sorted(list(all_eng_words))\n",
        "target_words = sorted(list(all_hindi_words))\n",
        "num_encoder_tokens = len(all_eng_words)\n",
        "num_decoder_tokens = len(all_hindi_words)\n",
        "num_encoder_tokens, num_decoder_tokens"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(14030, 17540)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 32
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TcOktsIJJoew"
      },
      "source": [
        "num_decoder_tokens += 1\n",
        "num_encoder_tokens += 1"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uPdZLmfWJoew"
      },
      "source": [
        "input_token_index = dict([(word, i+1) for i, word in enumerate(input_words)])\n",
        "target_token_index = dict([(word, i+1) for i, word in enumerate(target_words)])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h7W-L3MPJoex"
      },
      "source": [
        "reverse_input_char_index = dict((i, word) for word, i in input_token_index.items())\n",
        "reverse_target_char_index = dict((i, word) for word, i in target_token_index.items())"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 427
        },
        "id": "R4slRrg0Joex",
        "outputId": "0927c0c7-e445-4a23-f314-c0a2904145f9"
      },
      "source": [
        "df = shuffle(df)\n",
        "df.head(10)"
      ],
      "execution_count": null,
      "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>source</th>\n",
              "      <th>english_sentence</th>\n",
              "      <th>hindi_sentence</th>\n",
              "      <th>eng_sent_len</th>\n",
              "      <th>hindi_sent_len</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>71318</th>\n",
              "      <td>ted</td>\n",
              "      <td>thats been viewed four million times</td>\n",
              "      <td>START_ और उसको भी चालीस लाख बार देखा गया है। _END</td>\n",
              "      <td>6</td>\n",
              "      <td>11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>46829</th>\n",
              "      <td>ted</td>\n",
              "      <td>this herbavate cream for eczema possible</td>\n",
              "      <td>START_ हरबावेट क्रीम का निर्माण सँभव हुआ । _END</td>\n",
              "      <td>6</td>\n",
              "      <td>9</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>96392</th>\n",
              "      <td>ted</td>\n",
              "      <td>and thats a good thing because in order</td>\n",
              "      <td>START_ और यह अच्छी बात है क्योंकि _END</td>\n",
              "      <td>8</td>\n",
              "      <td>8</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>89048</th>\n",
              "      <td>ted</td>\n",
              "      <td>somalia had disintegrated into anarchy</td>\n",
              "      <td>START_ समालीया में आराजकता छा गयी थी _END</td>\n",
              "      <td>5</td>\n",
              "      <td>8</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>38443</th>\n",
              "      <td>ted</td>\n",
              "      <td>and asked me to define happiness</td>\n",
              "      <td>START_ और मुझसे ख़ुशी की परिभाषा पूछते _END</td>\n",
              "      <td>6</td>\n",
              "      <td>8</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>126860</th>\n",
              "      <td>ted</td>\n",
              "      <td>the insurgency thats been unleashed on your pe...</td>\n",
              "      <td>START_ ये आपके खिलाफ किया हुआ विद्रोह _END</td>\n",
              "      <td>8</td>\n",
              "      <td>8</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>69870</th>\n",
              "      <td>ted</td>\n",
              "      <td>making microdecisions dozens and dozens</td>\n",
              "      <td>START_ छोटे छोटे दर्ज़नों फ़ैसले लेने होते हैं...</td>\n",
              "      <td>5</td>\n",
              "      <td>9</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>110879</th>\n",
              "      <td>ted</td>\n",
              "      <td>on the banks of the rivers</td>\n",
              "      <td>START_ वैश्यावृत्ति में बदलकर रह गया है। _END</td>\n",
              "      <td>6</td>\n",
              "      <td>8</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20352</th>\n",
              "      <td>ted</td>\n",
              "      <td>our heros are human heros</td>\n",
              "      <td>START_ मानव हीरो हमारा हीरो है _END</td>\n",
              "      <td>5</td>\n",
              "      <td>7</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>88219</th>\n",
              "      <td>ted</td>\n",
              "      <td>he had a side parting and a neat mustache</td>\n",
              "      <td>START_ उनके पास एक साइड पार्टिंग और एक साफ मूं...</td>\n",
              "      <td>9</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "       source  ... hindi_sent_len\n",
              "71318     ted  ...             11\n",
              "46829     ted  ...              9\n",
              "96392     ted  ...              8\n",
              "89048     ted  ...              8\n",
              "38443     ted  ...              8\n",
              "126860    ted  ...              8\n",
              "69870     ted  ...              9\n",
              "110879    ted  ...              8\n",
              "20352     ted  ...              7\n",
              "88219     ted  ...             12\n",
              "\n",
              "[10 rows x 5 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 36
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_fUPUKqDSgSs"
      },
      "source": [
        "#### Splitting the data into train and test"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iOyHRFvvJoey",
        "outputId": "5a072107-a20e-423e-e66b-b1de76f54e6a"
      },
      "source": [
        "X, y = df['english_sentence'], df['hindi_sentence']\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": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((19819,), (4955,))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 37
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BJp03orMJoez"
      },
      "source": [
        "def generate_batch(X = X_train, y = y_train, batch_size = 128):\n",
        "    ''' Generate a batch of data '''\n",
        "    while True:\n",
        "        for j in range(0, len(X), batch_size):\n",
        "            encoder_input_data = np.zeros((batch_size, max_length_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_decoder_tokens),dtype='float32')\n",
        "            for i, (input_text, target_text) in enumerate(zip(X[j:j+batch_size], y[j:j+batch_size])):\n",
        "                for t, word in enumerate(input_text.split()):\n",
        "                    encoder_input_data[i, t] = input_token_index[word] # encoder input seq\n",
        "                for t, word in enumerate(target_text.split()):\n",
        "                    if t<len(target_text.split())-1:\n",
        "                        decoder_input_data[i, t] = target_token_index[word] # decoder input seq\n",
        "                    if t>0:\n",
        "                        # decoder target sequence (one hot encoded)\n",
        "                        # does not include the START_ token\n",
        "                        # Offset by one timestep\n",
        "                        decoder_target_data[i, t - 1, target_token_index[word]] = 1.\n",
        "            yield([encoder_input_data, decoder_input_data], decoder_target_data)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yyl5NF52Joez"
      },
      "source": [
        "### Encoder-Decoder Architecture\n",
        "\n",
        "(Reference from the problem statement PDF provided)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "F_DalbqVJoe1"
      },
      "source": [
        "latent_dim=300"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iEWdsNSAJoe1"
      },
      "source": [
        "# Encoder\n",
        "encoder_inputs = Input(shape=(None,))\n",
        "enc_emb =  Embedding(num_encoder_tokens, latent_dim, mask_zero = True)(encoder_inputs)\n",
        "encoder_lstm = LSTM(latent_dim, return_state=True)\n",
        "encoder_outputs, state_h, state_c = encoder_lstm(enc_emb)\n",
        "# We discard `encoder_outputs` and only keep the states.\n",
        "encoder_states = [state_h, state_c]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EGlE81F5Joe2"
      },
      "source": [
        "# Set up the decoder, using `encoder_states` as initial state.\n",
        "decoder_inputs = Input(shape=(None,))\n",
        "dec_emb_layer = Embedding(num_decoder_tokens, latent_dim, mask_zero = True)\n",
        "dec_emb = dec_emb_layer(decoder_inputs)\n",
        "# We set up our decoder to return full output sequences,\n",
        "# and to return internal states as well. We don't use the\n",
        "# return states in the training model, but we will use them in inference.\n",
        "decoder_lstm = LSTM(latent_dim, return_sequences=True, return_state=True)\n",
        "decoder_outputs, _, _ = decoder_lstm(dec_emb, initial_state=encoder_states)\n",
        "decoder_dense = Dense(num_decoder_tokens, activation='softmax')\n",
        "decoder_outputs = decoder_dense(decoder_outputs)\n",
        "\n",
        "# Define the model that will turn\n",
        "# `encoder_input_data` & `decoder_input_data` into `decoder_target_data`\n",
        "model = Model([encoder_inputs, decoder_inputs], decoder_outputs)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CtYP27OaJoe2"
      },
      "source": [
        "model.compile(optimizer='rmsprop', loss='categorical_crossentropy')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "H6LTpJF8Joe2",
        "outputId": "bebf17cd-7f69-4503-ab5b-ed45736e95b7"
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_1 (InputLayer)            [(None, None)]       0                                            \n",
            "__________________________________________________________________________________________________\n",
            "input_2 (InputLayer)            [(None, None)]       0                                            \n",
            "__________________________________________________________________________________________________\n",
            "embedding (Embedding)           (None, None, 300)    4209300     input_1[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "embedding_1 (Embedding)         (None, None, 300)    5262300     input_2[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "lstm (LSTM)                     [(None, 300), (None, 721200      embedding[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "lstm_1 (LSTM)                   [(None, None, 300),  721200      embedding_1[0][0]                \n",
            "                                                                 lstm[0][1]                       \n",
            "                                                                 lstm[0][2]                       \n",
            "__________________________________________________________________________________________________\n",
            "dense (Dense)                   (None, None, 17541)  5279841     lstm_1[0][0]                     \n",
            "==================================================================================================\n",
            "Total params: 16,193,841\n",
            "Trainable params: 16,193,841\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "J-y0E236Joe3"
      },
      "source": [
        "train_samples = len(X_train)\n",
        "val_samples = len(X_test)\n",
        "batch_size = 128\n",
        "epochs = 50"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uosatohrJoe3",
        "outputId": "e87c5b5d-1a58-4c97-faa3-6520e54a24f8"
      },
      "source": [
        "history = model.fit_generator(generator = generate_batch(X_train, y_train, batch_size = batch_size),\n",
        "                    steps_per_epoch = train_samples//batch_size,\n",
        "                    epochs=epochs,\n",
        "                    validation_data = generate_batch(X_test, y_test, batch_size = batch_size),\n",
        "                    validation_steps = val_samples//batch_size)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/keras/engine/training.py:1915: 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"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "154/154 [==============================] - 57s 209ms/step - loss: 3.2088 - val_loss: 2.8643\n",
            "Epoch 2/50\n",
            "154/154 [==============================] - 30s 192ms/step - loss: 2.7583 - val_loss: 2.6874\n",
            "Epoch 3/50\n",
            "154/154 [==============================] - 29s 189ms/step - loss: 2.5803 - val_loss: 2.6119\n",
            "Epoch 4/50\n",
            "154/154 [==============================] - 30s 194ms/step - loss: 2.4631 - val_loss: 2.5657\n",
            "Epoch 5/50\n",
            "154/154 [==============================] - 30s 195ms/step - loss: 2.3769 - val_loss: 2.5434\n",
            "Epoch 6/50\n",
            "154/154 [==============================] - 30s 193ms/step - loss: 2.2923 - val_loss: 2.4941\n",
            "Epoch 7/50\n",
            "154/154 [==============================] - 30s 194ms/step - loss: 2.2164 - val_loss: 2.4787\n",
            "Epoch 8/50\n",
            "154/154 [==============================] - 30s 193ms/step - loss: 2.1484 - val_loss: 2.4698\n",
            "Epoch 9/50\n",
            "154/154 [==============================] - 30s 193ms/step - loss: 2.0791 - val_loss: 2.4502\n",
            "Epoch 10/50\n",
            "154/154 [==============================] - 29s 191ms/step - loss: 2.0120 - val_loss: 2.4431\n",
            "Epoch 11/50\n",
            "154/154 [==============================] - 29s 191ms/step - loss: 1.9419 - val_loss: 2.4409\n",
            "Epoch 12/50\n",
            "154/154 [==============================] - 30s 193ms/step - loss: 1.8751 - val_loss: 2.4380\n",
            "Epoch 13/50\n",
            "154/154 [==============================] - 29s 190ms/step - loss: 1.8138 - val_loss: 2.4449\n",
            "Epoch 14/50\n",
            "154/154 [==============================] - 29s 189ms/step - loss: 1.7512 - val_loss: 2.4539\n",
            "Epoch 15/50\n",
            "154/154 [==============================] - 29s 191ms/step - loss: 1.6924 - val_loss: 2.4574\n",
            "Epoch 16/50\n",
            "154/154 [==============================] - 30s 194ms/step - loss: 1.6319 - val_loss: 2.4772\n",
            "Epoch 17/50\n",
            "154/154 [==============================] - 29s 190ms/step - loss: 1.5776 - val_loss: 2.4808\n",
            "Epoch 18/50\n",
            "154/154 [==============================] - 29s 191ms/step - loss: 1.5236 - val_loss: 2.4918\n",
            "Epoch 19/50\n",
            "154/154 [==============================] - 29s 191ms/step - loss: 1.4690 - val_loss: 2.5128\n",
            "Epoch 20/50\n",
            "154/154 [==============================] - 29s 191ms/step - loss: 1.4151 - val_loss: 2.5352\n",
            "Epoch 21/50\n",
            "154/154 [==============================] - 29s 191ms/step - loss: 1.3589 - val_loss: 2.5437\n",
            "Epoch 22/50\n",
            "154/154 [==============================] - 30s 192ms/step - loss: 1.3061 - val_loss: 2.5632\n",
            "Epoch 23/50\n",
            "154/154 [==============================] - 30s 192ms/step - loss: 1.2545 - val_loss: 2.5908\n",
            "Epoch 24/50\n",
            "154/154 [==============================] - 29s 191ms/step - loss: 1.2058 - val_loss: 2.6118\n",
            "Epoch 25/50\n",
            "154/154 [==============================] - 30s 193ms/step - loss: 1.1594 - val_loss: 2.6292\n",
            "Epoch 26/50\n",
            "154/154 [==============================] - 29s 191ms/step - loss: 1.1113 - val_loss: 2.6581\n",
            "Epoch 27/50\n",
            "154/154 [==============================] - 29s 191ms/step - loss: 1.0626 - val_loss: 2.6871\n",
            "Epoch 28/50\n",
            "154/154 [==============================] - 30s 193ms/step - loss: 1.0186 - val_loss: 2.6999\n",
            "Epoch 29/50\n",
            "154/154 [==============================] - 29s 190ms/step - loss: 0.9761 - val_loss: 2.7171\n",
            "Epoch 30/50\n",
            "154/154 [==============================] - 29s 189ms/step - loss: 0.9351 - val_loss: 2.7438\n",
            "Epoch 31/50\n",
            "154/154 [==============================] - 29s 189ms/step - loss: 0.8926 - val_loss: 2.7664\n",
            "Epoch 32/50\n",
            "154/154 [==============================] - 30s 195ms/step - loss: 0.8501 - val_loss: 2.7842\n",
            "Epoch 33/50\n",
            "154/154 [==============================] - 29s 187ms/step - loss: 0.8091 - val_loss: 2.8143\n",
            "Epoch 34/50\n",
            "154/154 [==============================] - 29s 188ms/step - loss: 0.7693 - val_loss: 2.8298\n",
            "Epoch 35/50\n",
            "154/154 [==============================] - 29s 187ms/step - loss: 0.7341 - val_loss: 2.8635\n",
            "Epoch 36/50\n",
            "154/154 [==============================] - 29s 189ms/step - loss: 0.6983 - val_loss: 2.8829\n",
            "Epoch 37/50\n",
            "154/154 [==============================] - 29s 190ms/step - loss: 0.6650 - val_loss: 2.9004\n",
            "Epoch 38/50\n",
            "154/154 [==============================] - 29s 190ms/step - loss: 0.6351 - val_loss: 2.9267\n",
            "Epoch 39/50\n",
            "154/154 [==============================] - 29s 188ms/step - loss: 0.6051 - val_loss: 2.9490\n",
            "Epoch 40/50\n",
            "154/154 [==============================] - 29s 190ms/step - loss: 0.5795 - val_loss: 2.9758\n",
            "Epoch 41/50\n",
            "154/154 [==============================] - 29s 190ms/step - loss: 0.5530 - val_loss: 2.9967\n",
            "Epoch 42/50\n",
            "154/154 [==============================] - 29s 187ms/step - loss: 0.5256 - val_loss: 3.0234\n",
            "Epoch 43/50\n",
            "154/154 [==============================] - 29s 189ms/step - loss: 0.5020 - val_loss: 3.0369\n",
            "Epoch 44/50\n",
            "154/154 [==============================] - 29s 188ms/step - loss: 0.4796 - val_loss: 3.0554\n",
            "Epoch 45/50\n",
            "154/154 [==============================] - 29s 190ms/step - loss: 0.4565 - val_loss: 3.0805\n",
            "Epoch 46/50\n",
            "154/154 [==============================] - 29s 191ms/step - loss: 0.4366 - val_loss: 3.0999\n",
            "Epoch 47/50\n",
            "154/154 [==============================] - 30s 192ms/step - loss: 0.4154 - val_loss: 3.1174\n",
            "Epoch 48/50\n",
            "154/154 [==============================] - 29s 187ms/step - loss: 0.3972 - val_loss: 3.1423\n",
            "Epoch 49/50\n",
            "154/154 [==============================] - 29s 189ms/step - loss: 0.3778 - val_loss: 3.1599\n",
            "Epoch 50/50\n",
            "154/154 [==============================] - 29s 189ms/step - loss: 0.3602 - val_loss: 3.1785\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "edeEArQ-Joe3"
      },
      "source": [
        "model.save_weights('model_weights.h5')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pxOjc2UtTH_l"
      },
      "source": [
        "#### Evaluating model performance"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yXp0EzZoJoe4"
      },
      "source": [
        "# Encode the input sequence to get the \"thought vectors\"\n",
        "encoder_model = Model(encoder_inputs, encoder_states)\n",
        "\n",
        "# Decoder setup\n",
        "# Below tensors will hold the states of the previous time step\n",
        "decoder_state_input_h = Input(shape=(latent_dim,))\n",
        "decoder_state_input_c = Input(shape=(latent_dim,))\n",
        "decoder_states_inputs = [decoder_state_input_h, decoder_state_input_c]\n",
        "\n",
        "dec_emb2= dec_emb_layer(decoder_inputs) # Get the embeddings of the decoder sequence\n",
        "\n",
        "# To predict the next word in the sequence, set the initial states to the states from the previous time step\n",
        "decoder_outputs2, state_h2, state_c2 = decoder_lstm(dec_emb2, initial_state=decoder_states_inputs)\n",
        "decoder_states2 = [state_h2, state_c2]\n",
        "decoder_outputs2 = decoder_dense(decoder_outputs2) # A dense softmax layer to generate prob dist. over the target vocabulary\n",
        "\n",
        "# Final decoder model\n",
        "decoder_model = Model(\n",
        "    [decoder_inputs] + decoder_states_inputs,\n",
        "    [decoder_outputs2] + decoder_states2)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WI5Ctxv7Joe4"
      },
      "source": [
        "def decode_sequence(input_seq):\n",
        "    # Encode the input as state vectors.\n",
        "    states_value = encoder_model.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] = target_token_index['START_']\n",
        "\n",
        "    # Sampling loop for a batch of sequences\n",
        "    # (to simplify, here we assume a batch of size 1).\n",
        "    stop_condition = False\n",
        "    decoded_sentence = ''\n",
        "    while not stop_condition:\n",
        "        output_tokens, h, c = decoder_model.predict([target_seq] + states_value)\n",
        "\n",
        "        # Sample a token\n",
        "        sampled_token_index = np.argmax(output_tokens[0, -1, :])\n",
        "        sampled_char = reverse_target_char_index[sampled_token_index]\n",
        "        decoded_sentence += ' '+sampled_char\n",
        "\n",
        "        # Exit condition: either hit max length\n",
        "        # or find stop character.\n",
        "        if (sampled_char == '_END' or\n",
        "           len(decoded_sentence) > 50):\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": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BEIiNi4mJoe5"
      },
      "source": [
        "train_gen = generate_batch(X_train, y_train, batch_size = 1)\n",
        "k=-1"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QVk2DJ7CJoe5",
        "outputId": "28da8439-f962-401a-fdfd-8ca032c4c6e0"
      },
      "source": [
        "k+=1\n",
        "(input_seq, actual_output), _ = next(train_gen)\n",
        "decoded_sentence = decode_sequence(input_seq)\n",
        "print('Input English sentence:', X_train[k:k+1].values[0])\n",
        "print('Actual Hindi Translation:', y_train[k:k+1].values[0][6:-4])\n",
        "print('Predicted Hindi Translation:', decoded_sentence[:-4])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Input English sentence: and of intelligence\n",
            "Actual Hindi Translation:  और बुद्धिमत्ता को \n",
            "Predicted Hindi Translation:  और बुद्धिमत्ता को \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SyPGde6rJoe6",
        "outputId": "c7fc8dad-38bf-4a81-a2ec-be2789dd21a4"
      },
      "source": [
        "k+=1\n",
        "(input_seq, actual_output), _ = next(train_gen)\n",
        "decoded_sentence = decode_sequence(input_seq)\n",
        "print('Input English sentence:', X_train[k:k+1].values[0])\n",
        "print('Actual Hindi Translation:', y_train[k:k+1].values[0][6:-4])\n",
        "print('Predicted Hindi Translation:', decoded_sentence[:-4])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Input English sentence: i would have died in a day\n",
            "Actual Hindi Translation:  तो मैं वहां एक दिन में मर गया होता \n",
            "Predicted Hindi Translation:  तो मैं एक दिन में काम करना चाहता था \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DueLyKiEJoe6",
        "outputId": "ef6622ca-dafe-4a7f-91d6-b7b71bda040f"
      },
      "source": [
        "k+=1\n",
        "(input_seq, actual_output), _ = next(train_gen)\n",
        "decoded_sentence = decode_sequence(input_seq)\n",
        "print('Input English sentence:', X_train[k:k+1].values[0])\n",
        "print('Actual Hindi Translation:', y_train[k:k+1].values[0][6:-4])\n",
        "print('Predicted Hindi Translation:', decoded_sentence[:-4])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Input English sentence: and people started renovating their own structures\n",
            "Actual Hindi Translation:  और लोगों ने अपने ही ढ़ाँचे को पुनः बनाना शुरु किया। \n",
            "Predicted Hindi Translation:  और लोग अपने घर का जन्म क्षेत्र में था \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "k2h7LQSMJoe7",
        "outputId": "0af17f8c-c8cb-42a1-cd25-20a24eb3235d"
      },
      "source": [
        "k+=1\n",
        "(input_seq, actual_output), _ = next(train_gen)\n",
        "decoded_sentence = decode_sequence(input_seq)\n",
        "print('Input English sentence:', X_train[k:k+1].values[0])\n",
        "print('Actual Hindi Translation:', y_train[k:k+1].values[0][6:-4])\n",
        "print('Predicted Hindi Translation:', decoded_sentence[:-4])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Input English sentence: bullets from bursting\n",
            "Actual Hindi Translation:  गोलियां फोड़ना बंद नहीं करता है \n",
            "Predicted Hindi Translation:  नहीं जाएगा \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "p1rL7rE5Joe7",
        "outputId": "bce84594-e236-4574-abc0-1262ce6a8131"
      },
      "source": [
        "k+=1\n",
        "(input_seq, actual_output), _ = next(train_gen)\n",
        "decoded_sentence = decode_sequence(input_seq)\n",
        "print('Input English sentence:', X_train[k:k+1].values[0])\n",
        "print('Actual Hindi Translation:', y_train[k:k+1].values[0][6:-4])\n",
        "print('Predicted Hindi Translation:', decoded_sentence[:-4])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Input English sentence: and so its not just an offshore problem\n",
            "Actual Hindi Translation:  तो यह केवल देश के बाहर की समस्या नहीं है \n",
            "Predicted Hindi Translation:  और यह तो किसी के लिए ये मुमकिन नहीं है । \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1ZAk-2e5TP3D"
      },
      "source": [
        "### **Conclusion**\n",
        "\n",
        "As you may observe, the performance of model is not satisfactory as the translations are not accurate. However, there are a number of possible reasons of why the performance is bad. \n",
        "\n",
        "First of all, we used a small amount data for training the model. Secondly, we improve the model performance by using pretraining embeddings like Glove or contextual embedding models like transformers (BERT, RoBERTA and so on). Adding attention layer can also improve the model performance significantly.\n",
        "\n",
        "We can also make sure that the training data of english sentences is correctly mapped to its hindi translation.\n",
        "\n",
        "Neural Language Translation is one of the toughest real-world NLP problems and translation from english -> hindi is tougher due to the complex nature of hindi language.\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ]
    }
  ]
}