Memotion_network_task_B.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Memotion_network_task_B.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZB2HG5cimwDB",
        "colab_type": "text"
      },
      "source": [
        "**Getting Data**\n",
        "\n",
        "In previous notebook we downloaded photos and save them on google drive "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rViDsspAm2ZL",
        "colab_type": "code",
        "outputId": "a993ffa0-bffe-4fdb-b617-2bdaa2e21a0c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 268
        }
      },
      "source": [
        "\"\"\"libraries we use\"\"\"\n",
        "!sudo apt-get install unzip\n",
        "from google.colab import drive\n",
        "from google.colab import files\n",
        "import csv \n",
        "drive.mount('/content/drive')\n",
        "import os.path"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Reading package lists... Done\n",
            "Building dependency tree       \n",
            "Reading state information... Done\n",
            "unzip is already the newest version (6.0-21ubuntu1).\n",
            "The following package was automatically installed and is no longer required:\n",
            "  libnvidia-common-430\n",
            "Use 'sudo apt autoremove' to remove it.\n",
            "0 upgraded, 0 newly installed, 0 to remove and 7 not upgraded.\n",
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&response_type=code&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cvHZwXW_pv9E",
        "colab_type": "text"
      },
      "source": [
        "Upload csv file to read data from it"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uguFnfLOnAiL",
        "colab_type": "code",
        "outputId": "319ebb7d-0cc3-4d6e-a02a-c6edff6aa8fe",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 90
        }
      },
      "source": [
        "!cp drive/My\\ Drive/Memotion/data_train.csv /content/\n",
        "!cp drive/My\\ Drive/Memotion/data_test.csv /content/\n",
        "\n",
        "\n",
        "train_filename = 'data_train.csv'\n",
        "test_filename = 'data_test.csv'\n",
        "\n",
        "# reading the file into a list\n",
        "train_data = [] \n",
        "test_data = []\n",
        "with open(train_filename, 'r') as csvfile: \n",
        "    csvreader = csv.reader(csvfile)  \n",
        "    for row in csvreader: \n",
        "        train_data.append(row) \n",
        "csvfile.close()\n",
        "\n",
        "with open(test_filename, 'r') as csvfile: \n",
        "    csvreader = csv.reader(csvfile)  \n",
        "    for row in csvreader: \n",
        "        test_data.append(row) \n",
        "csvfile.close()\n",
        "\n",
        "print('Example :')\n",
        "print(train_data[0])\n",
        "print(train_data[1])"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Example :\n",
            "['Image_name', 'Image_URL', 'OCR_extracted_text', 'corrected_text', 'Humour', 'Sarcasm', 'offensive', 'Motivational', 'Overall_Sentiment']\n",
            "['10_year_2r94rv.jpg', 'https://i.imgflip.com/2r94rv.jpg', 'LOOK THERE MY FRIEND LIGHTYEAR NOW ALL SOHALIKUT TREND PLAY THE 10 YEARS CHALLENGE AT FACEBOOK imgflip.com ', 'LOOK THERE MY FRIEND LIGHTYEAR NOW ALL SOHALIKUT TREND PLAY THE 10 YEARS CHALLENGE AT FACEBOOK imgflip.com ', 'hilarious', 'general', 'not_offensive', 'not_motivational', 'very_positive']\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wKN-CCeoIG2A",
        "colab_type": "text"
      },
      "source": [
        "Some lines has incorrect data e.g. in some lines corrected ocr is empty , in addition in some lines url contains wrong data so we delet lines which does not have 9 columns"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iRa4gVsSDM3G",
        "colab_type": "code",
        "outputId": "e3f8bdeb-8bd8-4000-fea6-5ab2b6384462",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "del train_data[0] # header part\n",
        "counter = 0  \n",
        "for line in train_data[:] :\n",
        "  if(line[-1] == ''):\n",
        "    train_data.remove(line)\n",
        "print('new train data has' ,str(len(train_data)) , 'lines!')\n",
        "\n",
        "del test_data[0] # header part\n",
        "counter = 0  \n",
        "for line in test_data[:] :\n",
        "  if(line[-1] == ''):\n",
        "    test_data.remove(line)\n",
        "print('new test data has' ,str(len(test_data)) , 'lines!')"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "new train data has 6521 lines!\n",
            "new test data has 914 lines!\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "I061KaozL1R9",
        "colab_type": "text"
      },
      "source": [
        "Some of corrected OCR still contains websites url or username ,We try to delete most of them"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DsYujnWPIi_Z",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "for line in train_data :\n",
        "  corrected_ocr = ' '.join(item for item in line[3].split() if not (item.startswith('@') or item.endswith('.com') or item.endswith('.net')))\n",
        "  line[3] = corrected_ocr\n",
        "\n",
        "for line in test_data :\n",
        "  corrected_ocr = ' '.join(item for item in line[3].split() if not (item.startswith('@') or item.endswith('.com') or item.endswith('.net')))\n",
        "  line[3] = corrected_ocr"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "B3mWKPAXpYoN",
        "colab_type": "text"
      },
      "source": [
        "get prepared dataset from google drive"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vpnLRcFzoZub",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!cp drive/My\\ Drive/Memotion/train_dataset.zip /content/\n",
        "!cp drive/My\\ Drive/Memotion/test_dataset.zip /content/\n",
        "\n",
        "!unzip train_dataset.zip\n",
        "!unzip test_dataset.zip"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8lhcRPTE1rWl",
        "colab_type": "text"
      },
      "source": [
        "some photos can not convert to jpeg so we delete them"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2Td-FdUA0LgG",
        "colab_type": "code",
        "outputId": "c9307c2e-2dba-4bba-9dbc-a79c8df8f0e7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 70
        }
      },
      "source": [
        "import os\n",
        "try:\n",
        "  # %tensorflow_version only exists in Colab.\n",
        "  %tensorflow_version 2.x\n",
        "except Exception:\n",
        "  pass\n",
        "import tensorflow as tf\n",
        "\n",
        "TRAIN_DATASET = 'train_dataset/'\n",
        "TEST_DATASET = 'test_dataset/'\n",
        "\n",
        "for image in os.listdir(TRAIN_DATASET) :\n",
        "  try :\n",
        "    img = tf.io.read_file(TRAIN_DATASET + image)\n",
        "    img = tf.image.decode_jpeg(img,channels=3)\n",
        "  except:\n",
        "    os.remove(TRAIN_DATASET + image)\n",
        "print(\"number of photos in train : \" , len(os.listdir(TRAIN_DATASET)))\n",
        "\n",
        "for image in os.listdir(TEST_DATASET) :\n",
        "  try :\n",
        "    img = tf.io.read_file(TEST_DATASET + image)\n",
        "    img = tf.image.decode_jpeg(img,channels=3)\n",
        "  except:\n",
        "    os.remove(TEST_DATASET + image)\n",
        "print(\"number of photos in test : \" , len(os.listdir(TEST_DATASET)))"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "TensorFlow 2.x selected.\n",
            "number of photos in train :  5847\n",
            "number of photos in test :  831\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HdNTSuK9p1BQ",
        "colab_type": "text"
      },
      "source": [
        "\n",
        "Now we want to get images and texts with their labels for Humour ,Sarcasm and Offensive"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EhXGI_vbpW71",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "train_images = []\n",
        "train_texts = []\n",
        "train_labels = []\n",
        "\n",
        "max_len_text = 0\n",
        "\n",
        "\n",
        "for line in train_data : \n",
        "  if os.path.exists(TRAIN_DATASET + line[0]):\n",
        "    train_images.append(TRAIN_DATASET + line[0])\n",
        "    train_texts.append(line[3])\n",
        "    if len(line[3].split()) > max_len_text :\n",
        "      max_len_text = len(line[3].split())\n",
        "    label=[]\n",
        "    is_other = [1,0]\n",
        "    if line[4] == 'not_funny':\n",
        "      label.append([0,1])\n",
        "    else :\n",
        "      label.append([1,0])\n",
        "      is_other = [0,1]\n",
        "    if line[5] == 'not_sarcastic' or line[5] == 'general':\n",
        "      label.append([0,1])\n",
        "    else :\n",
        "      label.append([1,0])\n",
        "      is_other = [0,1]\n",
        "    if line[6] == 'not_offensive':\n",
        "      label.append([0,1])\n",
        "    else :\n",
        "      label.append([1,0])\n",
        "      is_other = [0,1]\n",
        "     \n",
        "    label.append(is_other)\n",
        "    train_labels.append(label)\n",
        "\n",
        "test_images = []\n",
        "test_texts = []\n",
        "test_labels = []\n",
        "\n",
        "for line in test_data : \n",
        "  if os.path.exists(TEST_DATASET + line[0]):\n",
        "    test_images.append(TEST_DATASET + line[0])\n",
        "    test_texts.append(line[3])\n",
        "    if len(line[3].split()) > max_len_text :\n",
        "      max_len_text = len(line[3].split())\n",
        "    label=[]\n",
        "    is_other = [1,0]\n",
        "    if line[4] == 'not_funny':\n",
        "      label.append([0,1])\n",
        "    else :\n",
        "      label.append([1,0])\n",
        "      is_other = [0,1]\n",
        "    if line[5] == 'not_sarcastic' or line[5] == 'general':\n",
        "      label.append([0,1])\n",
        "    else :\n",
        "      label.append([1,0])\n",
        "      is_other = [0,1]\n",
        "    if line[6] == 'not_offensive':\n",
        "      label.append([0,1])\n",
        "    else :\n",
        "      label.append([1,0])\n",
        "      is_other = [0,1]\n",
        "    label.append(is_other)\n",
        "    test_labels.append(label)\n",
        "\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tLVGWnTfsYju",
        "colab_type": "text"
      },
      "source": [
        "\n",
        "**make dataset with tensorflow**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ai3jw48msfDN",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "\"\"\"libraries we use\"\"\"\n",
        "try:\n",
        "  # %tensorflow_version only exists in Colab.\n",
        "  %tensorflow_version 2.x\n",
        "except Exception:\n",
        "  pass\n",
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras.preprocessing.text import Tokenizer\n",
        "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
        "from tensorflow.keras import utils\n",
        "import tensorflow.keras.layers\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras import applications\n",
        "from tensorflow.keras import regularizers\n",
        "from tensorflow.keras import optimizers\n",
        "from tensorflow.keras import backend as K"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "l9_yHzU3srwz",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# implementing tokenizer\n",
        "tokenizer = Tokenizer(num_words=10000)\n",
        "tokenizer.fit_on_texts(train_texts + test_texts)\n",
        "sequences = tokenizer.texts_to_sequences(train_texts)\n",
        "train_texts = pad_sequences(sequences, maxlen=max_len_text)\n",
        "sequences = tokenizer.texts_to_sequences(test_texts)\n",
        "test_texts = pad_sequences(sequences, maxlen=max_len_text)\n",
        "word_index = tokenizer.word_index"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9Ec23yFzuZcE",
        "colab_type": "text"
      },
      "source": [
        "we need a word embedding , we use glove pre-trained word embedding"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "2ViReIiCt2-H",
        "outputId": "f227de2b-457b-47d7-a5b5-e6f7867b536e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "# !wget http://nlp.stanford.edu/data/glove.6B.zip\n",
        "# !cp /content/drive/My\\ Drive/Memotion/glove.6B.zip /content/\n",
        "# !unzip glove.6B.zip\n",
        "\n",
        "embeddings_index = {}\n",
        "f = open('glove.6B.100d.txt')\n",
        "\n",
        "for line in f:\n",
        "  values = line.split()\n",
        "  word = values[0]\n",
        "  coefs = np.asarray(values[1:], dtype='float32')\n",
        "  embeddings_index[word] = coefs\n",
        "\n",
        "f.close()\n",
        "\n",
        "print('Found %s word vectors.' % len(embeddings_index))\n",
        "\n",
        "#  making embedding matrix \n",
        "embedding_dim = 100\n",
        "embedding_matrix = np.zeros((len(word_index) + 1, embedding_dim))\n",
        "for word, i in word_index.items():\n",
        "    embedding_vector = embeddings_index.get(word)\n",
        "    if embedding_vector is not None:\n",
        "      embedding_matrix[i] = embedding_vector"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 400000 word vectors.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FBk7VNCVus7n",
        "colab_type": "text"
      },
      "source": [
        "now we create dataset using tensorflow"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "e1XSdChLuqK6",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def get_images(inputs,outputs) :\n",
        "  img = tf.io.read_file(inputs[0])\n",
        "  return (img , inputs[1]) , outputs\n",
        "\n",
        "def process_image(inputs,outputs):\n",
        "    img = tf.image.decode_jpeg(inputs[0], channels=3)\n",
        "    img = tf.image.convert_image_dtype(img, tf.float32)\n",
        "    img = tf.image.resize(img, [200, 200])\n",
        "    img = img / 255.0\n",
        "    return (img,inputs[1]),outputs\n",
        "\n",
        "def decode_label(inputs,outputs):\n",
        "    return inputs,(outputs[0],outputs[1],outputs[2],outputs[3])\n",
        "\n",
        "\n",
        "AUTOTUNE = tf.data.experimental.AUTOTUNE\n",
        "X1 = X2 = X3 = X4 =  AUTOTUNE \n",
        "train_dataset = tf.data.Dataset.from_tensor_slices(((train_images,train_texts),train_labels))\n",
        "train_dataset = train_dataset.shuffle(len(train_images))\n",
        "train_dataset = train_dataset.map(get_images, num_parallel_calls=X1)\n",
        "train_dataset = train_dataset.map(process_image,num_parallel_calls=X2)\n",
        "train_dataset = train_dataset.map(decode_label,num_parallel_calls=X3)\n",
        "train_dataset = train_dataset.batch(batch_size=30)\n",
        "train_dataset = train_dataset.prefetch(buffer_size=X4)\n",
        "\n",
        "AUTOTUNE2 = tf.data.experimental.AUTOTUNE\n",
        "Y1 = Y2 = Y3 = Y4 =  AUTOTUNE\n",
        "test_dataset = tf.data.Dataset.from_tensor_slices(((test_images,test_texts),test_labels))\n",
        "test_dataset = test_dataset.map(get_images, num_parallel_calls=Y1)\n",
        "test_dataset = test_dataset.map(process_image,num_parallel_calls=Y2)\n",
        "test_dataset = test_dataset.map(decode_label,num_parallel_calls=Y3)\n",
        "test_dataset = test_dataset.batch(batch_size=100)\n",
        "test_dataset = test_dataset.prefetch(buffer_size=Y4)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "a0OHFQwfvA6c",
        "colab_type": "text"
      },
      "source": [
        "it is time to create network"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cl3qSaey7ylm",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def create_model():\n",
        "  image_input = keras.Input(shape=(200,200,3))\n",
        "  cnn1 = keras.layers.Conv2D(50,(5,5),activation='relu')(image_input)\n",
        "  max_pool_1 = keras.layers.MaxPool2D()(cnn1)\n",
        "  cnn2 = keras.layers.Conv2D(25,(5,5),activation='relu')(max_pool_1)\n",
        "  max_pool_2 = keras.layers.MaxPool2D()(cnn2)\n",
        "  cnn3 = keras.layers.Conv2D(25,(4,4),activation='relu')(max_pool_2)\n",
        "  max_pool_3 = keras.layers.MaxPool2D()(cnn3)\n",
        "  cnn4 = keras.layers.Conv2D(15,(3,3),activation='relu')(max_pool_3)\n",
        "  max_pool_4= keras.layers.MaxPool2D()(cnn4)\n",
        "  cnn5 = keras.layers.Conv2D(5,(3,3),activation='relu',kernel_regularizer=regularizers.l2(0.01))(max_pool_4)\n",
        "  flatten = keras.layers.Flatten()(cnn5)\n",
        "\n",
        "  text_input = keras.Input(shape=(max_len_text))\n",
        "  embedding_layer = keras.layers.Embedding(len(word_index) + 1,embedding_dim,weights=[embedding_matrix],name = 'embedding_layer',trainable=False)\n",
        "  embedded = embedding_layer(text_input)\n",
        "  lstm1 = keras.layers.LSTM(300,return_sequences=True)(embedded)\n",
        "  lstm2 = keras.layers.LSTM(300,return_sequences=True)(lstm1)\n",
        "  lstm3 = keras.layers.LSTM(300,kernel_regularizer=regularizers.l2(0.01))(lstm2)\n",
        "\n",
        "  concatinate = tf.concat([flatten,lstm3] , axis = 1)\n",
        "  dense1 = keras.layers.Dense(100)(concatinate)\n",
        "  dropout = keras.layers.Dropout(0.2)(dense1)\n",
        "  dense2 = keras.layers.Dense(25)(dropout)\n",
        "  dense3 = keras.layers.Dense(2,activation = 'softmax',name = 'Humour')(dense2)\n",
        "  dense4 = keras.layers.Dense(2,activation = 'softmax',name = 'Sarcasm')(dense2)\n",
        "  dense5 = keras.layers.Dense(2,activation = 'softmax',name = 'Offensive')(dense2)\n",
        "  dense6 = keras.layers.Dense(2,activation = 'softmax',name = 'Other')(dense2)\n",
        "\n",
        "  model = keras.Model([image_input,text_input],[dense3,dense4,dense5,dense6])\n",
        "\n",
        "  \n",
        "  tf.keras.utils.plot_model(model, 'model.png', show_shapes=True)\n",
        "  model.compile(optimizer='adam' , loss='binary_crossentropy',metrics=['acc']) \n",
        "\n",
        "  return model"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SqxePZ3DvKvD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model = create_model()\n",
        "\n",
        "checkpoint_callback_2 = tf.keras.callbacks.ModelCheckpoint('model_B/', monitor='val_Offensive_acc', verbose=1, save_best_only=True,save_weights_only=True, mode='max')\n",
        "\n",
        "history = model.fit(train_dataset,epochs = 40,validation_data=test_dataset,callbacks=[checkpoint_callback_1,checkpoint_callback_2])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gh56UILceoTJ",
        "colab_type": "text"
      },
      "source": [
        "plot the training loop"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2CQPd77hena2",
        "colab_type": "code",
        "outputId": "dca55252-c0d3-4cc7-92fa-98a2af0c02d5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# Plot training & validation loss values\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.title('Model loss')\n",
        "plt.ylabel('Loss')\n",
        "plt.xlabel('Epoch')\n",
        "plt.legend(['Train', 'Test'], loc='upper left')\n",
        "plt.show()\n",
        "\n",
        "\n",
        "# Plot training & validation acc values\n",
        "plt.plot(history.history['Humour_acc'])\n",
        "plt.plot(history.history['val_Humour_acc'])\n",
        "plt.title('Model Acc Humour')\n",
        "plt.ylabel('Acc')\n",
        "plt.xlabel('Epoch')\n",
        "plt.legend(['Train', 'Test'], loc='upper left')\n",
        "plt.show()\n",
        "\n",
        "# Plot training & validation acc values\n",
        "plt.plot(history.history['Sarcasm_acc'])\n",
        "plt.plot(history.history['val_Sarcasm_acc'])\n",
        "plt.title('Model Acc Sarcasm')\n",
        "plt.ylabel('Acc')\n",
        "plt.xlabel('Epoch')\n",
        "plt.legend(['Train', 'Test'], loc='upper left')\n",
        "plt.show()\n",
        "\n",
        "# Plot training & validation acc values\n",
        "plt.plot(history.history['Offensive_acc'])\n",
        "plt.plot(history.history['val_Offensive_acc'])\n",
        "plt.title('Model Acc Offensive')\n",
        "plt.ylabel('Acc')\n",
        "plt.xlabel('Epoch')\n",
        "plt.legend(['Train', 'Test'], loc='upper left')\n",
        "plt.show()\n",
        "\n",
        "# Plot training & validation acc values\n",
        "plt.plot(history.history['Other_acc'])\n",
        "plt.plot(history.history['val_Other_acc'])\n",
        "plt.title('Model Acc Other')\n",
        "plt.ylabel('Acc')\n",
        "plt.xlabel('Epoch')\n",
        "plt.legend(['Train', 'Test'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTKoAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n9P8B1q3yUpEpmfgAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEMYYY/yyBGGMMcYvSxDGGGP8+n99MbwIMhwprAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrloAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Wf6sD7TKDeB3",
        "colab_type": "text"
      },
      "source": [
        "evaluate the best model "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SfrY6re7Rjzb",
        "colab_type": "code",
        "outputId": "a646781d-09bd-49cd-853e-aa299e6ccd9b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 90
        }
      },
      "source": [
        "model = create_model()\n",
        "model.evaluate(test_dataset)\n",
        "model.load_weights('model_B/')\n",
        "score = model.evaluate(test_dataset)\n",
        "avg_acc = (score[-1] + score[-2] + score[-3] + score[-4])/4\n",
        "print('Total accuracy : ' , avg_acc)"
      ],
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "9/9 [==============================] - 6s 659ms/step - loss: 7.6258 - Humour_loss: 0.6951 - Sarcasm_loss: 0.6910 - Offensive_loss: 0.6935 - Other_loss: 0.6846 - Humour_acc: 0.4958 - Sarcasm_acc: 0.6847 - Offensive_acc: 0.4308 - Other_acc: 0.3369\n",
            "9/9 [==============================] - 5s 568ms/step - loss: 2.2827 - Humour_loss: 0.5443 - Sarcasm_loss: 0.5727 - Offensive_loss: 0.6405 - Other_loss: 0.4879 - Humour_acc: 0.8375 - Sarcasm_acc: 0.7918 - Offensive_acc: 0.6980 - Other_acc: 0.9146\n",
            "Total accuracy :  0.8104693293571472\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QE_pZlv4DiC-",
        "colab_type": "text"
      },
      "source": [
        "we have to get f1-score and recall for this task"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3FcAIytDS5GN",
        "colab_type": "code",
        "outputId": "c1c65895-9269-4cd5-8dae-39b7dff19d75",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "from sklearn.metrics import f1_score\n",
        "from sklearn.metrics import recall_score\n",
        "\n",
        "\n",
        "Humour_labels = []\n",
        "Sarcasm_labels = []\n",
        "Offensive_labels = []\n",
        "Other_labels = []\n",
        "\n",
        "score = model.predict(test_dataset)\n",
        "\n",
        "for _, label in test_dataset: \n",
        "    humour_label = label[0]\n",
        "    sarcasm_label = label[1]\n",
        "    offensive_label = label[2]\n",
        "    other_label = label[3]\n",
        "    for i in humour_label :\n",
        "      Humour_labels.append(i)\n",
        "    for i in sarcasm_label :\n",
        "      Sarcasm_labels.append(i)\n",
        "    for i in offensive_label :\n",
        "      Offensive_labels.append(i)\n",
        "    for i in other_label :\n",
        "      Other_labels.append(i)\n",
        "\n",
        "f1_humour = f1_score(Humour_labels,score[0] > 0.5,average='micro')\n",
        "f1_sarcasm = f1_score(Sarcasm_labels,score[1] > 0.5,average='micro')\n",
        "f1_offensive = f1_score(Offensive_labels,score[2] > 0.5,average='micro')\n",
        "f1_other = f1_score(Other_labels,score[3] > 0.5,average='micro')\n",
        "\n",
        "f1_score = (f1_humour + f1_sarcasm + f1_offensive + f1_other)/4\n",
        "print('f1_score : ' , f1_score)\n",
        "\n",
        "recall_humour = recall_score(Humour_labels,score[0] > 0.5,average='micro')\n",
        "recall_sarcasm = recall_score(Sarcasm_labels,score[1] > 0.5,average='micro')\n",
        "recall_offensive = recall_score(Offensive_labels,score[2] > 0.5,average='micro')\n",
        "recall_other = recall_score(Other_labels,score[3] > 0.5,average='micro')\n",
        "\n",
        "recall = (recall_humour + recall_sarcasm + recall_offensive + recall_other)/4\n",
        "print('recall : ' , recall)\n"
      ],
      "execution_count": 44,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "f1_score :  0.8104693140794224\n",
            "recall :  0.8104693140794224\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "n9bgQqDob8wd",
        "colab_type": "code",
        "outputId": "d5d611fc-a2e9-4375-977b-2a89811f1ecd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 728
        }
      },
      "source": [
        "from sklearn.metrics import classification_report\n",
        "\n",
        "target_names_1 = ['Humour','Not Humour']\n",
        "target_names_2 = ['Sarcasm','Not Sarcasm']\n",
        "target_names_3 = ['Offensive','Not Offensive']\n",
        "target_names_4 = ['Other','Not Other']\n",
        "\n",
        "Humour_labels = []\n",
        "Sarcasm_labels = []\n",
        "Offensive_labels = []\n",
        "Other_labels = []\n",
        "\n",
        "score = model.predict(test_dataset)\n",
        "\n",
        "for _, label in test_dataset: \n",
        "    humour_label = label[0]\n",
        "    sarcasm_label = label[1]\n",
        "    offensive_label = label[2]\n",
        "    other_label = label[3]\n",
        "    for i in humour_label :\n",
        "      Humour_labels.append(i)\n",
        "    for i in sarcasm_label :\n",
        "      Sarcasm_labels.append(i)\n",
        "    for i in offensive_label :\n",
        "      Offensive_labels.append(i)\n",
        "    for i in other_label :\n",
        "      Other_labels.append(i)\n",
        "\n",
        "print(classification_report(Humour_labels, score[0] > 0.5, target_names=target_names_1))\n",
        "print(classification_report(Sarcasm_labels, score[1] > 0.5, target_names=target_names_2))\n",
        "print(classification_report(Offensive_labels, score[2] > 0.5, target_names=target_names_3))\n",
        "print(classification_report(Other_labels, score[3] > 0.5, target_names=target_names_4))"
      ],
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "      Humour       0.83      0.99      0.90       645\n",
            "  Not Humour       0.90      0.31      0.46       186\n",
            "\n",
            "   micro avg       0.84      0.84      0.84       831\n",
            "   macro avg       0.87      0.65      0.68       831\n",
            "weighted avg       0.85      0.84      0.80       831\n",
            " samples avg       0.84      0.84      0.84       831\n",
            "\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "     Sarcasm       0.87      0.26      0.41       223\n",
            " Not Sarcasm       0.79      0.99      0.87       608\n",
            "\n",
            "   micro avg       0.79      0.79      0.79       831\n",
            "   macro avg       0.83      0.62      0.64       831\n",
            "weighted avg       0.81      0.79      0.75       831\n",
            " samples avg       0.79      0.79      0.79       831\n",
            "\n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "    Offensive       0.67      0.97      0.80       501\n",
            "Not Offensive       0.88      0.28      0.42       330\n",
            "\n",
            "    micro avg       0.70      0.70      0.70       831\n",
            "    macro avg       0.77      0.63      0.61       831\n",
            " weighted avg       0.75      0.70      0.65       831\n",
            "  samples avg       0.70      0.70      0.70       831\n",
            "\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "       Other       0.94      0.32      0.48       102\n",
            "   Not Other       0.91      1.00      0.95       729\n",
            "\n",
            "   micro avg       0.91      0.91      0.91       831\n",
            "   macro avg       0.93      0.66      0.72       831\n",
            "weighted avg       0.92      0.91      0.90       831\n",
            " samples avg       0.91      0.91      0.91       831\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qY7E8YtFDK0Y",
        "colab_type": "code",
        "outputId": "956dca06-eaba-4e17-a594-eb6f2be65823",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 106
        }
      },
      "source": [
        "!zip -r model_B.zip 'model_B'"
      ],
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "  adding: model_B/ (stored 0%)\n",
            "  adding: model_B/.data-00001-of-00002 (deflated 7%)\n",
            "  adding: model_B/checkpoint (deflated 34%)\n",
            "  adding: model_B/.index (deflated 74%)\n",
            "  adding: model_B/.data-00000-of-00002 (deflated 27%)\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}