03_NMT_with_Attention_Tensorflow.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "T4",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Hrithik2212/Neural-Machine-Translation--English-to-Tamil/blob/main/03_NMT_with_Attention_Tensorflow.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#Libraries\n",
        "import pathlib\n",
        "import random\n",
        "import string\n",
        "import re\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras import layers\n",
        "from tensorflow.keras.layers import TextVectorization\n",
        "from tensorflow.keras.callbacks import EarlyStopping\n",
        "from tensorflow.keras import callbacks"
      ],
      "metadata": {
        "id": "P6NCNVZdM-d-"
      },
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "eng_file = '/content/drive/MyDrive/en-ta/train.en'\n",
        "tamil_file = '/content/drive/MyDrive/en-ta/train.ta'"
      ],
      "metadata": {
        "id": "nWlPq_OnNSzQ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Xve9sGmY_y6R",
        "outputId": "13155cbe-3219-466f-9a43-f80a60ab8817"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "with open(eng_file, 'r') as file:\n",
        "    english_sentences = file.readlines()\n",
        "with open(tamil_file, 'r') as file:\n",
        "    tamil_sentences = file.readlines()\n",
        "\n",
        "\n",
        "english_sentences = [sentence.rstrip('\\n').lower() for sentence in english_sentences]\n",
        "tamil_sentences = [sentence.rstrip('\\n') for sentence in tamil_sentences]"
      ],
      "metadata": {
        "id": "s_U2sHClNcoH"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# filter out sentences which contains letters other than whats listed below\n",
        "tamil_vocabulary = [ ' ', '!', '\"', '#', '$', '%', '&', \"'\", '(', ')', '*', '+', ',', '-', '.', '/',\n",
        "                      '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', ':', '<', '=', '>', '?', 'ˌ',\n",
        "                      \"ஃ\", \"அ\", \"ஆ\", \"இ\", \"ஈ\", \"உ\", \"ஊ\", \"எ\", \"ஏ\", \"ஐ\", \"ஒ\", \"ஓ\", \"ஔ\",\n",
        "                      \"க\", \"ங\", \"ச\", \"ஜ\", \"ஞ\", \"ட\", \"ண\",\n",
        "                      \"த\", \"ந\", \"ன\", \"ப\", \"ம\", \"ய\", \"ர\",\n",
        "                      \"ற\", \"ல\", \"ள\", \"ழ\", \"வ\", \"ஶ\", \"ஷ\", \"ஸ\", \"ஹ\",\n",
        "                      \"ா\", \"ி\", \"ீ\", \"ு\", \"ூ\", \"ெ\", \"ே\", \"ை\", \"ொ\", \"ோ\", \"ௌ\", \"்\", \"ௗ\",\n",
        "                      \"௦\", \"௧\", \"௨\", \"௫\", \"௬\", \"௲\", \"௳\"\n",
        "                      ]\n",
        "\n",
        "english_vocabulary = [' ', '!', '\"', '#', '$', '%', '&', \"'\", '(', ')', '*', '+', ',', '-', '.', '/',\n",
        "                        '0', '1', '2', '3', '4', '5', '6', '7', '8', '9',\n",
        "                        ':', '<', '=', '>', '?', '@',\n",
        "                        '[', '\\\\', ']', '^', '_', '`',\n",
        "                        'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l',\n",
        "                        'm', 'n', 'o', 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', 'x',\n",
        "                        'y', 'z',\n",
        "                        '{', '|', '}', '~']"
      ],
      "metadata": {
        "id": "cMD_loL5NfYC"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def is_valid_tokens(sentence , vocab):\n",
        "    for token in list(set(sentence)):\n",
        "        if token not in vocab:\n",
        "            return False\n",
        "    return True\n",
        "\n",
        "valid_sentence_indicies = []\n",
        "for index in range(len(english_sentences[:400000])):\n",
        "    tamil_sentence, english_sentence = tamil_sentences[index], english_sentences[index]\n",
        "    if is_valid_tokens(tamil_sentence, tamil_vocabulary) \\\n",
        "      and is_valid_tokens(english_sentence , english_vocabulary):\n",
        "        valid_sentence_indicies.append(index)"
      ],
      "metadata": {
        "id": "kB-SFjUHNmXa"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "TOTAL_SENTENCES = 200000 # lets take 200000 valid sentences to train the model\n",
        "x = [english_sentences[i] for i in valid_sentence_indicies[:TOTAL_SENTENCES]]\n",
        "y = [tamil_sentences[i] for i in valid_sentence_indicies[:TOTAL_SENTENCES]]"
      ],
      "metadata": {
        "id": "1dEv5HxSN4Zv"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "dataset =  pd.DataFrame({\"English\":x , \"Tamil\":y })\n",
        "dataset.info()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9tW6u08dNyrF",
        "outputId": "7931532f-65ab-46c7-cd93-66ea8b31857f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 200000 entries, 0 to 199999\n",
            "Data columns (total 2 columns):\n",
            " #   Column   Non-Null Count   Dtype \n",
            "---  ------   --------------   ----- \n",
            " 0   English  200000 non-null  object\n",
            " 1   Tamil    200000 non-null  object\n",
            "dtypes: object(2)\n",
            "memory usage: 3.1+ MB\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "dataset.sample(5)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "OkaMHA5uQlif",
        "outputId": "5639917d-5669-46f0-8f9c-47643544412d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                                  English                                              Tamil\n",
              "83292   for that, he thanked people of the area and pa...  இவருக்கு காங்கிரஸ் கட்சியின் நிா்வாகிகளும், அந...\n",
              "160422  she said that these initiatives have up-graded...  வீட்டு வசதி நிதித் துறையை இந்திய ரிசர்வ் வங்கி...\n",
              "159073  hence, action is being taken to set up an agro...  எனவே கிருஷ்ணகிரி மண்டல ஊரக வளர்ச்சி மற்றும் ஊர...\n",
              "86022   recent statistics in the us reveal that in 200...  அமெரிக்காவில் சமீபத்திய புள்ளிவிவரங்கள், 2005ம...\n",
              "97840   even paul had to endure a continuing conflict ...  இது, தேவபக்தியை வெளிக்காட்டுவதில் நல்ல ஆன்மீக ..."
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-6242fdf4-1353-4e3a-a48b-bb9fd4aa3a74\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>English</th>\n",
              "      <th>Tamil</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>83292</th>\n",
              "      <td>for that, he thanked people of the area and pa...</td>\n",
              "      <td>இவருக்கு காங்கிரஸ் கட்சியின் நிா்வாகிகளும், அந...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>160422</th>\n",
              "      <td>she said that these initiatives have up-graded...</td>\n",
              "      <td>வீட்டு வசதி நிதித் துறையை இந்திய ரிசர்வ் வங்கி...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>159073</th>\n",
              "      <td>hence, action is being taken to set up an agro...</td>\n",
              "      <td>எனவே கிருஷ்ணகிரி மண்டல ஊரக வளர்ச்சி மற்றும் ஊர...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>86022</th>\n",
              "      <td>recent statistics in the us reveal that in 200...</td>\n",
              "      <td>அமெரிக்காவில் சமீபத்திய புள்ளிவிவரங்கள், 2005ம...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>97840</th>\n",
              "      <td>even paul had to endure a continuing conflict ...</td>\n",
              "      <td>இது, தேவபக்தியை வெளிக்காட்டுவதில் நல்ல ஆன்மீக ...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-6242fdf4-1353-4e3a-a48b-bb9fd4aa3a74')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-6242fdf4-1353-4e3a-a48b-bb9fd4aa3a74 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-6242fdf4-1353-4e3a-a48b-bb9fd4aa3a74');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-39cb6254-f733-4d8d-a077-22446f4444c4\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-39cb6254-f733-4d8d-a077-22446f4444c4')\"\n",
              "            title=\"Suggest charts.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-39cb6254-f733-4d8d-a077-22446f4444c4 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "dataset.to_csv('/content/drive/MyDrive/NMT-Project/data.csv' , index = False )"
      ],
      "metadata": {
        "id": "cG7j82LkBDuJ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "dataset = pd.read_csv('/content/drive/MyDrive/NMT-Project/data.csv')\n",
        "dataset.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "hADUIpxlBaWS",
        "outputId": "b969b503-a1ef-4909-eda2-556db25bc0f3"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                             English  \\\n",
              "0     some 14 months later, the second calf is born.   \n",
              "1                         this photo was taken then.   \n",
              "2  so far two rounds of the jwg meeting have been...   \n",
              "3  the life of the world is nothing but play and ...   \n",
              "4  the maximum temperature will be 35 degrees cel...   \n",
              "\n",
              "                                               Tamil  \n",
              "0  சுமார் 14 மாதங்கள் கழித்து, இரண்டாம் கன்றை ஈனு...  \n",
              "1              அதன்போது எடுக்கப்பட்ட புகைப்படம் இது.  \n",
              "2  இதுவரை இணைப் பணிக் குழு இரண்டுகட்ட பேச்சுவார்த...  \n",
              "3  உலக வாழ்க்கை வீணும் விளையாட்டுமேயன்றி வேறில்லை...  \n",
              "4  அதிகபட்ச வெப்பநிலை 35 டிகிரி செல்சியஸாகவும் கு...  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-d2592115-c0ed-4cec-aa67-798e25e1aab6\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>English</th>\n",
              "      <th>Tamil</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>some 14 months later, the second calf is born.</td>\n",
              "      <td>சுமார் 14 மாதங்கள் கழித்து, இரண்டாம் கன்றை ஈனு...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>this photo was taken then.</td>\n",
              "      <td>அதன்போது எடுக்கப்பட்ட புகைப்படம் இது.</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>so far two rounds of the jwg meeting have been...</td>\n",
              "      <td>இதுவரை இணைப் பணிக் குழு இரண்டுகட்ட பேச்சுவார்த...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>the life of the world is nothing but play and ...</td>\n",
              "      <td>உலக வாழ்க்கை வீணும் விளையாட்டுமேயன்றி வேறில்லை...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>the maximum temperature will be 35 degrees cel...</td>\n",
              "      <td>அதிகபட்ச வெப்பநிலை 35 டிகிரி செல்சியஸாகவும் கு...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-d2592115-c0ed-4cec-aa67-798e25e1aab6')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-d2592115-c0ed-4cec-aa67-798e25e1aab6 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-d2592115-c0ed-4cec-aa67-798e25e1aab6');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-fbc0abbb-7404-4151-bb09-da3a20b2e963\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-fbc0abbb-7404-4151-bb09-da3a20b2e963')\"\n",
              "            title=\"Suggest charts.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-fbc0abbb-7404-4151-bb09-da3a20b2e963 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 2
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "text_pairs = []\n",
        "\n",
        "for i in range(dataset.shape[0]) :\n",
        "    eng = str(dataset[\"English\"][i])\n",
        "    tam = \"[start] \" + str(dataset[\"Tamil\"][i]) + \" [end]\"\n",
        "    text_pairs.append((eng,tam))"
      ],
      "metadata": {
        "id": "zQN83ZIYQ7VB"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "len(text_pairs)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4wJmhIASVF5U",
        "outputId": "aaca871c-b7a9-4d5b-bd58-83e7961ebf99"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "200000"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_sample_size = len(text_pairs) - int(0.1*len(text_pairs))\n",
        "train_pairs = text_pairs[:train_sample_size]\n",
        "test_pairs = text_pairs[train_sample_size:]"
      ],
      "metadata": {
        "id": "fqJLqUwqVOFa"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"Total pairs :\", len(text_pairs))\n",
        "print(\"Train pairs :\", len(train_pairs))\n",
        "print(\"Test pairs :\", len(test_pairs))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uZePVpMOZigz",
        "outputId": "72bc61a8-db43-4dcf-d36c-cdda4fd16621"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Total pairs : 200000\n",
            "Train pairs : 180000\n",
            "Test pairs : 20000\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "strip_chars = string.punctuation + \"|\"\n",
        "strip_chars = strip_chars.replace(\"[\", \"\")\n",
        "strip_chars = strip_chars.replace(\"]\", \"\")\n",
        "strip_chars = strip_chars.replace('\"\"', \"\")\n",
        "\n",
        "vocab_size = 50000\n",
        "sequence_length = 30\n",
        "batch_size = 128"
      ],
      "metadata": {
        "id": "HlFE96z5aGOH"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def custom_standardization(input_string):\n",
        "    lowercase = tf.strings.lower(input_string)\n",
        "    return tf.strings.regex_replace(lowercase, \"[%s]\" % re.escape(strip_chars), \"\")"
      ],
      "metadata": {
        "id": "GVsTtwd-aQX9"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "---"
      ],
      "metadata": {
        "id": "Eka8ur6zAdx4"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "eng_vectorization = TextVectorization(\n",
        "    max_tokens=vocab_size, output_mode=\"int\", output_sequence_length=sequence_length,\n",
        ")\n",
        "\n",
        "tam_vectorization = TextVectorization(\n",
        "    max_tokens=vocab_size,\n",
        "    output_mode=\"int\",\n",
        "    output_sequence_length=sequence_length + 1,\n",
        "    standardize=custom_standardization,\n",
        ")"
      ],
      "metadata": {
        "id": "gWK93ztIaXCU"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_eng_texts = [pair[0] for pair in train_pairs]\n",
        "train_tam_texts = [pair[1] for pair in train_pairs]\n",
        "eng_vectorization.adapt(train_eng_texts)\n",
        "tam_vectorization.adapt(train_tam_texts)"
      ],
      "metadata": {
        "id": "fTsq6xqqacUT"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import json\n",
        "\n",
        "# Get the vocabulary\n",
        "eng_vocab = eng_vectorization.get_vocabulary()\n",
        "tam_vocab = tam_vectorization.get_vocabulary()\n",
        "\n",
        "\n",
        "# Save the vocabulary\n",
        "with open('/content/drive/MyDrive/NMT-Project/vectorizers/eng_vocab.json', 'w') as f:\n",
        "    json.dump(eng_vocab, f)\n",
        "with open('/content/drive/MyDrive/NMT-Project/vectorizers/tam_vocab.json', 'w') as f:\n",
        "    json.dump(tam_vocab, f)"
      ],
      "metadata": {
        "id": "XKwC7ligNnEg"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import json\n",
        "\n",
        "\n",
        "vocab_size = 50000\n",
        "sequence_length = 30\n",
        "batch_size = 128\n",
        "\n",
        "\n",
        "# Load the vocabulary\n",
        "with open('/content/drive/MyDrive/NMT-Project/vectorizers/eng_vocab.json', 'r') as f:\n",
        "    eng_vocab = json.load(f)\n",
        "with open('/content/drive/MyDrive/NMT-Project/vectorizers/tam_vocab.json', 'r') as f:\n",
        "    tam_vocab = json.load(f)\n",
        "\n",
        "# Create a TextVectorization layer with the loaded vocabulary\n",
        "eng_vectorization = TextVectorization(vocabulary=eng_vocab,output_mode=\"int\", output_sequence_length=sequence_length)\n",
        "tam_vectorization = TextVectorization(vocabulary=tam_vocab,output_mode=\"int\",\n",
        "    output_sequence_length=sequence_length + 1,\n",
        "    standardize=custom_standardization)"
      ],
      "metadata": {
        "id": "1j1MTJz_OG3L"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def format_dataset(eng, tam):\n",
        "    eng = eng_vectorization(eng)\n",
        "    tam = tam_vectorization(tam)\n",
        "    return ({\"encoder_inputs\": eng, \"decoder_inputs\": tam[:, :-1],}, tam[:, 1:])"
      ],
      "metadata": {
        "id": "2kofnF7fagxf"
      },
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def make_dataset(pairs):\n",
        "    eng_texts, tam_texts = zip(*pairs)\n",
        "    eng_texts = list(eng_texts)\n",
        "    tam_texts = list(tam_texts)\n",
        "    dataset = tf.data.Dataset.from_tensor_slices((eng_texts, tam_texts))\n",
        "    dataset = dataset.batch(batch_size)\n",
        "    dataset = dataset.map(format_dataset)\n",
        "    return dataset.shuffle(2048).prefetch(16).cache()"
      ],
      "metadata": {
        "id": "mcQhwuqdauKg"
      },
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_ds = make_dataset(train_pairs)\n",
        "test_ds = make_dataset(test_pairs)"
      ],
      "metadata": {
        "id": "3rvJJMvTbPAm"
      },
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "for inputs, targets in train_ds.take(2):\n",
        "    print(f'inputs[\"encoder_inputs\"].shape: {inputs[\"encoder_inputs\"].shape}')\n",
        "    print(f'inputs[\"decoder_inputs\"].shape: {inputs[\"decoder_inputs\"].shape}')\n",
        "    print(f\"targets.shape: {targets.shape}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ihbcYj-SaxLH",
        "outputId": "9e4f3f96-c877-4e2e-be7d-c9713fb6a96c"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "inputs[\"encoder_inputs\"].shape: (128, 30)\n",
            "inputs[\"decoder_inputs\"].shape: (128, 30)\n",
            "targets.shape: (128, 30)\n",
            "inputs[\"encoder_inputs\"].shape: (128, 30)\n",
            "inputs[\"decoder_inputs\"].shape: (128, 30)\n",
            "targets.shape: (128, 30)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "class TransformerEncoder(layers.Layer):\n",
        "    def __init__(self, embed_dim, dense_dim, num_heads, **kwargs):\n",
        "        super().__init__(**kwargs)\n",
        "        self.embed_dim = embed_dim\n",
        "        self.dense_dim = dense_dim\n",
        "        self.num_heads = num_heads\n",
        "        self.attention = layers.MultiHeadAttention(\n",
        "            num_heads=num_heads, key_dim=embed_dim\n",
        "        )\n",
        "        self.dense_proj = keras.Sequential(\n",
        "            [layers.Dense(dense_dim, activation=\"relu\"), layers.Dense(embed_dim),]\n",
        "        )\n",
        "        self.layernorm_1 = layers.LayerNormalization()\n",
        "        self.layernorm_2 = layers.LayerNormalization()\n",
        "        self.supports_masking = True\n",
        "\n",
        "    def call(self, inputs, mask=None):\n",
        "        if mask is not None:\n",
        "            padding_mask = tf.cast(mask[:, tf.newaxis, :], dtype=\"int32\")\n",
        "        attention_output = self.attention(\n",
        "            query=inputs, value=inputs, key=inputs, attention_mask=padding_mask\n",
        "        )\n",
        "        proj_input = self.layernorm_1(inputs + attention_output)\n",
        "        proj_output = self.dense_proj(proj_input)\n",
        "        return self.layernorm_2(proj_input + proj_output)"
      ],
      "metadata": {
        "id": "J1ELfsDzbJ8Z"
      },
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "class PositionalEmbedding(layers.Layer):\n",
        "    def __init__(self, sequence_length, vocab_size, embed_dim, **kwargs):\n",
        "        super().__init__(**kwargs)\n",
        "        self.token_embeddings = layers.Embedding(\n",
        "            input_dim=vocab_size, output_dim=embed_dim\n",
        "        )\n",
        "        self.position_embeddings = layers.Embedding(\n",
        "            input_dim=sequence_length, output_dim=embed_dim\n",
        "        )\n",
        "        self.sequence_length = sequence_length\n",
        "        self.vocab_size = vocab_size\n",
        "        self.embed_dim = embed_dim\n",
        "\n",
        "    def call(self, inputs):\n",
        "        length = tf.shape(inputs)[-1]\n",
        "        positions = tf.range(start=0, limit=length, delta=1)\n",
        "        embedded_tokens = self.token_embeddings(inputs)\n",
        "        embedded_positions = self.position_embeddings(positions)\n",
        "        return embedded_tokens + embedded_positions\n",
        "\n",
        "    def compute_mask(self, inputs, mask=None):\n",
        "        return tf.math.not_equal(inputs, 0)"
      ],
      "metadata": {
        "id": "zK62TVskbi0I"
      },
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "class TransformerDecoder(layers.Layer):\n",
        "    def __init__(self, embed_dim, latent_dim, num_heads, **kwargs):\n",
        "        super().__init__(**kwargs)\n",
        "        self.embed_dim = embed_dim\n",
        "        self.latent_dim = latent_dim\n",
        "        self.num_heads = num_heads\n",
        "        self.attention_1 = layers.MultiHeadAttention(\n",
        "            num_heads=num_heads, key_dim=embed_dim\n",
        "        )\n",
        "        self.attention_2 = layers.MultiHeadAttention(\n",
        "            num_heads=num_heads, key_dim=embed_dim\n",
        "        )\n",
        "        self.dense_proj = keras.Sequential(\n",
        "            [layers.Dense(latent_dim, activation=\"relu\"), layers.Dense(embed_dim),]\n",
        "        )\n",
        "        self.layernorm_1 = layers.LayerNormalization()\n",
        "        self.layernorm_2 = layers.LayerNormalization()\n",
        "        self.layernorm_3 = layers.LayerNormalization()\n",
        "        self.supports_masking = True\n",
        "\n",
        "    def call(self, inputs, encoder_outputs, mask=None):\n",
        "        causal_mask = self.get_causal_attention_mask(inputs)\n",
        "        if mask is not None:\n",
        "            padding_mask = tf.cast(mask[:, tf.newaxis, :], dtype=\"int32\")\n",
        "            padding_mask = tf.minimum(padding_mask, causal_mask)\n",
        "\n",
        "        attention_output_1 = self.attention_1(\n",
        "            query=inputs, value=inputs, key=inputs, attention_mask=causal_mask\n",
        "        )\n",
        "        out_1 = self.layernorm_1(inputs + attention_output_1)\n",
        "\n",
        "        attention_output_2  = self.attention_2(\n",
        "            query=out_1,\n",
        "            value=encoder_outputs,\n",
        "            key=encoder_outputs,\n",
        "            attention_mask=padding_mask,\n",
        "        )\n",
        "\n",
        "        out_2 = self.layernorm_2(out_1 + attention_output_2)\n",
        "\n",
        "        proj_output = self.dense_proj(out_2)\n",
        "        return self.layernorm_3(out_2 + proj_output)\n",
        "\n",
        "    def get_causal_attention_mask(self, inputs):\n",
        "        input_shape = tf.shape(inputs)\n",
        "        batch_size, sequence_length = input_shape[0], input_shape[1]\n",
        "        i = tf.range(sequence_length)[:, tf.newaxis]\n",
        "        j = tf.range(sequence_length)\n",
        "        mask = tf.cast(i >= j, dtype=\"int32\")\n",
        "        mask = tf.reshape(mask, (1, input_shape[1], input_shape[1]))\n",
        "        mult = tf.concat(\n",
        "            [tf.expand_dims(batch_size, -1), tf.constant([1, 1], dtype=tf.int32)],\n",
        "            axis=0,\n",
        "        )\n",
        "        return tf.tile(mask, mult)"
      ],
      "metadata": {
        "id": "Q7Icj1RNbk77"
      },
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "embed_dim = 512\n",
        "latent_dim = 1024\n",
        "num_heads = 8\n",
        "\n",
        "encoder_inputs = keras.Input(shape=(None,), dtype=\"int64\", name=\"encoder_inputs\")\n",
        "x = PositionalEmbedding(sequence_length, vocab_size, embed_dim)(encoder_inputs)\n",
        "encoder_outputs = TransformerEncoder(embed_dim, latent_dim, num_heads)(x)\n",
        "encoder = keras.Model(encoder_inputs, encoder_outputs)"
      ],
      "metadata": {
        "id": "4Ksz9gGNbzS6"
      },
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "decoder_inputs = keras.Input(shape=(None,), dtype=\"int64\", name=\"decoder_inputs\")\n",
        "encoded_seq_inputs = keras.Input(shape=(None, embed_dim), name=\"decoder_state_inputs\")\n",
        "x = PositionalEmbedding(sequence_length, vocab_size, embed_dim)(decoder_inputs)\n",
        "x = TransformerDecoder(embed_dim, latent_dim, num_heads)(x, encoded_seq_inputs)\n",
        "x = layers.Dropout(0.5)(x)\n",
        "decoder_outputs = layers.Dense(vocab_size, activation=\"softmax\")(x)\n",
        "decoder = keras.Model([decoder_inputs, encoded_seq_inputs], decoder_outputs)"
      ],
      "metadata": {
        "id": "Ct_pBX3eb19n"
      },
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "decoder_outputs = decoder([decoder_inputs, encoder_outputs])\n",
        "transformer = keras.Model(\n",
        "    [encoder_inputs, decoder_inputs], decoder_outputs, name=\"transformer\"\n",
        ")\n"
      ],
      "metadata": {
        "id": "JaXNIpDMb4UG"
      },
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "transformer.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Iuslbfhfb6N7",
        "outputId": "0e1a3239-b5d8-43f1-beb6-dfa3fc2837b9"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"transformer\"\n",
            "__________________________________________________________________________________________________\n",
            " Layer (type)                Output Shape                 Param #   Connected to                  \n",
            "==================================================================================================\n",
            " encoder_inputs (InputLayer  [(None, None)]               0         []                            \n",
            " )                                                                                                \n",
            "                                                                                                  \n",
            " positional_embedding (Posi  (None, None, 512)            2561536   ['encoder_inputs[0][0]']      \n",
            " tionalEmbedding)                                         0                                       \n",
            "                                                                                                  \n",
            " decoder_inputs (InputLayer  [(None, None)]               0         []                            \n",
            " )                                                                                                \n",
            "                                                                                                  \n",
            " transformer_encoder (Trans  (None, None, 512)            9453568   ['positional_embedding[0][0]']\n",
            " formerEncoder)                                                                                   \n",
            "                                                                                                  \n",
            " model_1 (Functional)        (None, None, 50000)          6912136   ['decoder_inputs[0][0]',      \n",
            "                                                          0          'transformer_encoder[0][0]'] \n",
            "                                                                                                  \n",
            "==================================================================================================\n",
            "Total params: 104190288 (397.45 MB)\n",
            "Trainable params: 104190288 (397.45 MB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "class CustomSchedule(tf.keras.optimizers.schedules.LearningRateSchedule):\n",
        "  def __init__(self, d_model, warmup_steps=4000):\n",
        "    super().__init__()\n",
        "\n",
        "    self.d_model = d_model\n",
        "    self.d_model = tf.cast(self.d_model, tf.float32)\n",
        "\n",
        "    self.warmup_steps = warmup_steps\n",
        "\n",
        "  def __call__(self, step):\n",
        "    step = tf.cast(step, dtype=tf.float32)\n",
        "    arg1 = tf.math.rsqrt(step)\n",
        "    arg2 = step * (self.warmup_steps ** -1.5)\n",
        "\n",
        "    return tf.math.rsqrt(self.d_model) * tf.math.minimum(arg1, arg2)\n",
        "\n",
        "  def get_config(self):\n",
        "    return {\"d_model\": self.d_model.numpy(), \"warmup_steps\": self.warmup_steps}"
      ],
      "metadata": {
        "id": "rDRJZ6Od8k7O"
      },
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "learning_rate = CustomSchedule(latent_dim)\n",
        "\n",
        "optimizer = tf.keras.optimizers.Adam(learning_rate, beta_1=0.9, beta_2=0.98,\n",
        "                                     epsilon=1e-9)"
      ],
      "metadata": {
        "id": "yxYNhrB18s1V"
      },
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(learning_rate(tf.range(40000, dtype=tf.float32)))\n",
        "plt.ylabel('Learning Rate')\n",
        "plt.xlabel('Train Step')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 466
        },
        "id": "QbU5_Miv8zzQ",
        "outputId": "72d52ccd-d99a-4ed6-ec38-af0d38aee3f8"
      },
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 0, 'Train Step')"
            ]
          },
          "metadata": {},
          "execution_count": 23
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "## call backs\n",
        "early_stopping = callbacks.EarlyStopping(\n",
        "    monitor='val_loss',\n",
        "    patience=10,\n",
        "    restore_best_weights=True\n",
        ")\n",
        "\n",
        "checkpoint_filepath = '/content/drive/MyDrive/NMT-Project/checkpoint'\n",
        "model_checkpoint_callback = callbacks.ModelCheckpoint(\n",
        "    filepath=checkpoint_filepath,\n",
        "    save_weights_only=False,\n",
        "    monitor='val_accuracy',\n",
        "    mode='max',\n",
        "    save_best_only=True)\n",
        "\n",
        "csv_logger = callbacks.CSVLogger('/content/drive/MyDrive/NMT-Project/training_log.csv' , append=True)\n",
        "\n",
        "\n",
        "backup_callback = callbacks.BackupAndRestore(backup_dir=\"/content/drive/MyDrive/NMT-Project/backup\" ,\n",
        "                                             delete_checkpoint=False)"
      ],
      "metadata": {
        "id": "Ew4UXJG7gULL"
      },
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "epochs = 30\n",
        "transformer.compile(\n",
        "    optimizer=optimizer,\n",
        "    loss=\"sparse_categorical_crossentropy\",\n",
        "    metrics=[\"accuracy\"]\n",
        ")"
      ],
      "metadata": {
        "id": "njReIjF2cIUk"
      },
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# tf.config.run_functions_eagerly(True)"
      ],
      "metadata": {
        "id": "RrqJcDAGPUZm"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "history = transformer.fit(train_ds,\n",
        "                epochs=epochs,\n",
        "                validation_data=test_ds ,\n",
        "                callbacks=[\n",
        "                      early_stopping ,\n",
        "                      model_checkpoint_callback ,\n",
        "                      csv_logger ,\n",
        "                      backup_callback ] )"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CDF5Iadd9TqO",
        "outputId": "57354526-52b5-4b5e-e091-168e3ab9c5de"
      },
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/30\n",
            "1407/1407 [==============================] - 750s 532ms/step - loss: 6.8777 - accuracy: 0.2729 - val_loss: 6.2177 - val_accuracy: 0.3025\n",
            "Epoch 2/30\n",
            "1407/1407 [==============================] - 744s 529ms/step - loss: 6.0646 - accuracy: 0.3047 - val_loss: 5.7225 - val_accuracy: 0.3244\n",
            "Epoch 3/30\n",
            "1407/1407 [==============================] - 739s 526ms/step - loss: 5.7221 - accuracy: 0.3232 - val_loss: 5.4949 - val_accuracy: 0.3390\n",
            "Epoch 4/30\n",
            "1407/1407 [==============================] - 747s 531ms/step - loss: 5.4075 - accuracy: 0.3429 - val_loss: 5.2972 - val_accuracy: 0.3506\n",
            "Epoch 5/30\n",
            "1407/1407 [==============================] - 740s 526ms/step - loss: 5.1014 - accuracy: 0.3628 - val_loss: 5.1886 - val_accuracy: 0.3573\n",
            "Epoch 6/30\n",
            "1407/1407 [==============================] - 731s 520ms/step - loss: 4.8687 - accuracy: 0.3798 - val_loss: 5.1709 - val_accuracy: 0.3562\n",
            "Epoch 7/30\n",
            "1407/1407 [==============================] - 728s 517ms/step - loss: 4.6874 - accuracy: 0.3944 - val_loss: 5.1379 - val_accuracy: 0.3569\n",
            "Epoch 8/30\n",
            "1407/1407 [==============================] - 728s 517ms/step - loss: 4.5287 - accuracy: 0.4080 - val_loss: 5.1896 - val_accuracy: 0.3557\n",
            "Epoch 9/30\n",
            "1407/1407 [==============================] - 731s 520ms/step - loss: 4.3815 - accuracy: 0.4211 - val_loss: 5.2997 - val_accuracy: 0.3555\n",
            "Epoch 10/30\n",
            "1407/1407 [==============================] - 729s 518ms/step - loss: 4.2478 - accuracy: 0.4335 - val_loss: 5.2768 - val_accuracy: 0.3470\n",
            "Epoch 11/30\n",
            "1407/1407 [==============================] - 736s 523ms/step - loss: 4.1222 - accuracy: 0.4450 - val_loss: 5.3278 - val_accuracy: 0.3416\n",
            "Epoch 12/30\n",
            "1407/1407 [==============================] - 731s 520ms/step - loss: 3.9960 - accuracy: 0.4570 - val_loss: 5.4279 - val_accuracy: 0.3375\n",
            "Epoch 13/30\n",
            "1407/1407 [==============================] - 731s 520ms/step - loss: 3.8741 - accuracy: 0.4691 - val_loss: 5.4465 - val_accuracy: 0.3427\n",
            "Epoch 14/30\n",
            "1407/1407 [==============================] - 728s 518ms/step - loss: 3.7576 - accuracy: 0.4805 - val_loss: 5.4750 - val_accuracy: 0.3396\n",
            "Epoch 15/30\n",
            "1407/1407 [==============================] - 733s 521ms/step - loss: 3.6530 - accuracy: 0.4922 - val_loss: 5.5337 - val_accuracy: 0.3288\n",
            "Epoch 16/30\n",
            "1407/1407 [==============================] - 733s 521ms/step - loss: 3.5695 - accuracy: 0.5030 - val_loss: 5.4905 - val_accuracy: 0.3332\n",
            "Epoch 17/30\n",
            "1407/1407 [==============================] - 732s 520ms/step - loss: 3.5056 - accuracy: 0.5128 - val_loss: 5.5362 - val_accuracy: 0.3349\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "transformer.save('/content/drive/MyDrive/NMT-Project/V1-50k-106M.h5')\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "V3aBvgB-yFYx",
        "outputId": "3a8d6f95-6e44-4302-f85b-54e52afccb74"
      },
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py:3079: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.\n",
            "  saving_api.save_model(\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import os\n",
        "def plot_loss_and_accuracy(history, save_dir=None, filename=None):\n",
        "    # Extract the loss and accuracy values from the history object\n",
        "    loss = history.history['loss']\n",
        "    val_loss = history.history['val_loss']\n",
        "    accuracy = history.history['accuracy']\n",
        "    val_accuracy = history.history['val_accuracy']\n",
        "\n",
        "    # Get the number of epochs\n",
        "    epochs = range(1, len(loss) + 1)\n",
        "\n",
        "    # Plot training and validation loss\n",
        "    plt.figure(figsize=(12, 5))\n",
        "    plt.subplot(1, 2, 1)\n",
        "    plt.plot(epochs, loss, 'bo-', label='Training Loss')\n",
        "    plt.plot(epochs, val_loss, 'ro-', label='Validation Loss')\n",
        "    plt.title('Training and Validation Loss')\n",
        "    plt.xlabel('Epochs')\n",
        "    plt.ylabel('Loss')\n",
        "    plt.legend()\n",
        "\n",
        "    # Plot training and validation accuracy\n",
        "    plt.subplot(1, 2, 2)\n",
        "    plt.plot(epochs, accuracy, 'bo-', label='Training Accuracy')\n",
        "    plt.plot(epochs, val_accuracy, 'ro-', label='Validation Accuracy')\n",
        "    plt.title('Training and Validation Accuracy')\n",
        "    plt.xlabel('Epochs')\n",
        "    plt.ylabel('Accuracy')\n",
        "    plt.legend()\n",
        "\n",
        "    plt.tight_layout()\n",
        "\n",
        "    # Save the figure if save_dir and filename are provided\n",
        "    if save_dir and filename:\n",
        "        os.makedirs(save_dir, exist_ok=True)\n",
        "        save_path = os.path.join(save_dir, filename)\n",
        "        plt.savefig(save_path)\n",
        "        print(f\"Plots saved as '{save_path}'\")\n",
        "\n",
        "    plt.show()\n",
        "plot_loss_and_accuracy(history ,\"/content/drive/MyDrive/NMT-Project\", \"V1-50k-106M_trainloss.png\" )"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 524
        },
        "id": "92V01oYBxxDx",
        "outputId": "3a0fa319-7310-4e8f-afd3-45555cab7dc7"
      },
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Plots saved as '/content/drive/MyDrive/NMT-Project/V1-50k-106M_trainloss.png'\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1200x500 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "tam_vocab = tam_vectorization.get_vocabulary()\n",
        "tam_index_lookup = dict(zip(range(len(tam_vocab)), tam_vocab))\n",
        "max_decoded_sentence_length = 20"
      ],
      "metadata": {
        "id": "x4qe0ttVc2mI"
      },
      "execution_count": 32,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def decode_sequence(input_sentence):\n",
        "    tokenized_input_sentence = eng_vectorization([input_sentence])\n",
        "    decoded_sentence = \"[start]\"\n",
        "    for i in range(max_decoded_sentence_length):\n",
        "        tokenized_target_sentence = tam_vectorization([decoded_sentence])[:, :-1]\n",
        "        predictions = transformer([tokenized_input_sentence, tokenized_target_sentence])\n",
        "\n",
        "        sampled_token_index = np.argmax(predictions[0, i, :])\n",
        "        sampled_token = tam_index_lookup[sampled_token_index]\n",
        "        decoded_sentence += \" \" + sampled_token\n",
        "\n",
        "        if sampled_token == \"[end]\":\n",
        "            break\n",
        "    return decoded_sentence"
      ],
      "metadata": {
        "id": "_0jL02Brex6k"
      },
      "execution_count": 33,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Testing on Random Train Samples"
      ],
      "metadata": {
        "id": "48uwpDWYw0LA"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 1\n",
        "sample  = random.choice(train_pairs)\n",
        "input_sentence = sample[0]\n",
        "translated = decode_sequence(input_sentence)\n",
        "print('English - ' , input_sentence )\n",
        "print('Tamil' )\n",
        "print('Ground Truth : ', sample[1])\n",
        "print('Machine Translation : ' , translated )"
      ],
      "metadata": {
        "id": "tf0EipyJfGZX",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "9b661003-223a-45f3-e197-c738daeb3d1d"
      },
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "English -  green chillies, chopped - 1 tbsp\n",
            "Tamil\n",
            "Ground Truth :  [start] பொடியாக நறுக்கிய பச்சை மிளகாய், கறிவேப்பிலை - தலா 1 டீஸ்பூன், [end]\n",
            "Machine Translation :  [start] பச்சை மிளகாய் 1 டீஸ்பூன் [end]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 2\n",
        "sample  = random.choice(train_pairs)\n",
        "input_sentence = sample[0]\n",
        "translated = decode_sequence(input_sentence)\n",
        "print('English - ' , input_sentence )\n",
        "print('Tamil' )\n",
        "print('Ground Truth : ', sample[1])\n",
        "print('Machine Translation : ' , translated )"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rJ8Hv9Qvy7mu",
        "outputId": "d73e8a1a-7581-4faa-d6d4-62a82e7c423d"
      },
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "English -  potassium carbonate is mainly used in the production of soap and glass.\n",
            "Tamil\n",
            "Ground Truth :  [start] பொட்டாசியம் கார்பனேட் ,சோப்பு மற்றும் கண்ணாடி உற்பத்தி செய்யப் பயன்படுத்தப்படுகிறது. [end]\n",
            "Machine Translation :  [start] பொட்டாசியம் கால்சியம் பாஸ்பரஸ் இரும்பு மற்றும் பொட்டாசியம் [UNK] போன்ற [UNK] உள்ளன [end]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 3\n",
        "sample  = random.choice(train_pairs)\n",
        "input_sentence = sample[0]\n",
        "translated = decode_sequence(input_sentence)\n",
        "print('English - ' , input_sentence )\n",
        "print('Tamil' )\n",
        "print('Ground Truth : ', sample[1])\n",
        "print('Machine Translation : ' , translated )"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KBIfrd-Yt13o",
        "outputId": "370c1f57-adc0-4cab-af3b-30115655fbb3"
      },
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "English -  these 11...\n",
            "Tamil\n",
            "Ground Truth :  [start] இந்த 11 கெத்து குணங்கள்தாம் . [end]\n",
            "Machine Translation :  [start] இந்த 11 [end]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Testing on Random Test samples"
      ],
      "metadata": {
        "id": "W4KV5PbIzZgs"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 1\n",
        "sample  = random.choice(test_pairs)\n",
        "input_sentence = sample[0]\n",
        "translated = decode_sequence(input_sentence)\n",
        "print('English - ' , input_sentence )\n",
        "print('Tamil' )\n",
        "print('Ground Truth : ', sample[1])\n",
        "print('Machine Translation : ' , translated )"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "u0V7qFjmzZH-",
        "outputId": "44932fe2-4922-46ac-9143-4854ee1eff23"
      },
      "execution_count": 44,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "English -  quarters for additional mahila judge and special judges at cuddalore in cuddalore district\n",
            "Tamil\n",
            "Ground Truth :  [start] கடலூர் மாவட்டம் கடலூரில் ஒருங்கிணைந்த நீதிமன்ற வளாகத்தில் இருசக்கர வாகனம் நிறுத்துமிடம் அமைத்தல் [end]\n",
            "Machine Translation :  [start] மாவட்ட ஆட்சியர் அலுவலக வளாகம் மற்றும் சென்னை [UNK] [end]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 2\n",
        "sample  = random.choice(test_pairs)\n",
        "input_sentence = sample[0]\n",
        "translated = decode_sequence(input_sentence)\n",
        "print('English - ' , input_sentence )\n",
        "print('Tamil' )\n",
        "print('Ground Truth : ', sample[1])\n",
        "print('Machine Translation : ' , translated )"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cB9NtIaxzjC9",
        "outputId": "3f5eb19c-1070-4b3b-fb7b-19ba90628253"
      },
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "English -  his car window was broken .\n",
            "Tamil\n",
            "Ground Truth :  [start] இதில் அவரது கார் கண்ணாடிகள் உடைந்தன. [end]\n",
            "Machine Translation :  [start] அந்த கார் [UNK] [UNK] [end]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 3\n",
        "sample  = random.choice(test_pairs)\n",
        "input_sentence = sample[0]\n",
        "translated = decode_sequence(input_sentence)\n",
        "print('English - ' , input_sentence )\n",
        "print('Tamil' )\n",
        "print('Ground Truth : ', sample[1])\n",
        "print('Machine Translation : ' , translated )"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1PgwLW6jzp46",
        "outputId": "3d923548-c9c9-46bb-c32c-52fa4b08987a"
      },
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "English -  the film was loved by fans.\n",
            "Tamil\n",
            "Ground Truth :  [start] ரசிகர்களுக்கு படம் பிடித்துவிட்டது. [end]\n",
            "Machine Translation :  [start] இந்த படம் ரசிகர்கள் மத்தியில் பெரும் வரவேற்பை பெற்றுள்ளது [end]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Testing on custom input"
      ],
      "metadata": {
        "id": "9v2hWFgm07NJ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "input_sentence = \"My name is Hrithik\"\n",
        "translated = decode_sequence(input_sentence)\n",
        "print('English - ' , input_sentence )\n",
        "print('Tamil' )\n",
        "print('Machine Translation : ' , translated )"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "b_DVf4Kh061Q",
        "outputId": "5c15be91-c17c-405e-eaef-0506c28c16f6"
      },
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "English -  My name is Hrithik\n",
            "Tamil\n",
            "Machine Translation :  [start] என் பெயர் [UNK] [end]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "input_sentence = \"I will buy milk later\"\n",
        "translated = decode_sequence(input_sentence)\n",
        "print('English - ' , input_sentence )\n",
        "print('Tamil' )\n",
        "print('Machine Translation : ' , translated )"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NI6nn8WSzvVP",
        "outputId": "e65e2a5d-cfa4-4bab-a7f6-f4c514b409eb"
      },
      "execution_count": 55,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "English -  I will buy milk later\n",
            "Tamil\n",
            "Machine Translation :  [start] பின்னர் பால் [UNK] [end]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "---"
      ],
      "metadata": {
        "id": "toB5M_9u1rHx"
      }
    }
  ]
}