train_classifier.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "train_classifier.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "machine_shape": "hm",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Chubbyman2/Emotion_Detector/blob/main/Emotion_Detector.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iTOesErakfpv"
      },
      "source": [
        "import os\n",
        "import keras\n",
        "import numpy as np\n",
        "from keras import optimizers, models, layers\n",
        "from keras.callbacks import ModelCheckpoint, EarlyStopping, ReduceLROnPlateau\n",
        "from keras.layers import Conv2D, Dense, Dropout, Flatten, Input, MaxPool2D, BatchNormalization, Activation\n",
        "from keras.optimizers import Adam\n",
        "from keras.preprocessing.image import ImageDataGenerator"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nYk1TSrdZX8Z"
      },
      "source": [
        "import zipfile\n",
        "from google.colab import drive\n",
        "drive.mount('/content/drive/')\n",
        "\n",
        "!unzip \"/content/drive/My Drive/facial_expressions_train.zip\"\n",
        "!unzip \"/content/drive/My Drive/facial_expressions_validation.zip\""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5CgeZd30mFYT"
      },
      "source": [
        "train_dir = \"train\"\n",
        "val_dir = \"validation\""
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 82
        },
        "id": "arTbWRdTZdbU",
        "outputId": "abdf6863-ca69-4a4d-f6a8-5b1710ca6cac"
      },
      "source": [
        "# Image sample from dataset\n",
        "# Note that they are all grayscale images with size 48x48\n",
        "from PIL import Image\n",
        "\n",
        "img = Image.open(\"train/Angry/0.jpg\")\n",
        "print(img.size)\n",
        "img"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(48, 48)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAADAAAAAwCAAAAAByaaZbAAAGd0lEQVR4nC3RSZOdVRkA4Hc65xvuvX17SqebToh0hChQMpRSaJW4RcotGzf+Nzeu3Ll1YelGrbK0qCACEgJJSHq84zecc973dePzFx78MHJa5MYSVIu3blVW0nbIHpt2f7etKwYv/1lC3l6vYEDpe5Wx0zHuEWxteP/NOlNUNYMQYxUlEFLpfvj3q5qs2WjI3IxynfcbpNHG5qc/wEkBAiFAYUGO4o417d59LOM2AadNVYmMTEM/mJz87IDqsTUBYScmBCIyB2zz/b91Mhr1xXRQyuPT85Lq1z44DlAmgMyERIQS2Q2IPdDOrevZjCRAmDRI3aDTw+NX32iLBTCD4mZEQA5g6m6gWL+0GA40i6R1MfL6aH/2vbOqADtDQGAAEUQCIEZXKEan+gL2wOq5jSCUxunenB1GEqeEjKZu2ck9MBIA0nAgG9+Id9MdGyiRSsuC6GMmELBCBEQcK8GhG4xFPM4WpsoMUU0YpJnU6EncAAXVK19WoWdSDKheBDjZAIQkaKZCcT6fVO5EiGiZtEsT6EjtYKLoiGzZydKADgLIInHv9LAmc1QUQa7R+gGJvU/Y7TKOA5Bp11SWHZ0FZPfkdBbZPXMQNChatQDcwLLrV4J1BldfrmEQcwd3kIOj3cDsuVB30SX1fZlPoZqzb7ebm26/iTCGNEmpAncAd5nvRELklId+07noIlsORweyQ855i+O0IrNmLATuiMBS10SITGXbG7mlslkNzUWczw535+NqG8aqCVBhCQSGgCxMmWp0oWq8/lbgRtZTHcskdZOG1jdD98ImtyI5EFBQikV0rNgpCECZ3gk3l9vmnScrwLaRWTj/680I4dXmKQU1RnNzEy8oyAEm8fDB8xzK9ZvvyPMqH83n8/0/fZZm7+kybxRdmRDcXNaDOsemhtTT4df6UTfT46P5xbQ63N+TkyUePXiYxkAIAMTgKBtDNfcKJYzy0tWz6aLai7FVaSc7P99dhPfzO08/2ZFECCQIJDe9GSIKtiP53Ttr01iqoENLzHe7bbyc3PvSPDvi/x8OKrAUB59o6950bXaTkDYNknt952KNTbXWUo8GoCDgcrcu6nXMI6cqDV40CFuuO4BsTT0/Ro/oCdzIHIhMSuqiDFT1wCEYjZUblISQakhc+6ZqqIiKFyVGQxMBKykmFyWOMbSFwrYrySmnTCNNvXYgHKIZEyKxBFCHGDwamANU7ZC1X5e2rQJVtRXEYoTq7oAASIRumpJ6ysDCZolTv8mt1BwjobbYrQfiOpM7uAEKOICmLL0QiLtqLF3X7hABemnHIefFkqEmMEcHRGFhtDyGgOpu5LTu1mnXwPO4sQm60YuuyhByFkFAECZkspRYzBFi0ZJyd/FkNU7uHXTNfKL5a0N0cC0EACgIjmil5ODohRVVp9988uzGbn8e+K0fl7T8FpXdBNwQEaWIIwEVLUGRCyefljtLerlb+fxHt3ywxytJDGxm7h5cRmB1QtfCDG7OTOHw7UXfb5q3zl4UTF+6Zg5UjLXKKpICqxFjRiJ3NQakYvWObqacp0Pz38e1zTqIXWOYgZOoOTg46AAEAAW6RJrHoWLpVo3kT9ltLbxNukMQcpDc58BUGIwis6ahL0A+aTuKvKbqs+9kxGaUO/e/ez6Jxig6UFVRxgBazErKSQJtVI+zxVHwX65VHk5fm/5y+PR3VwomQ2QGIjHUFLHoRbda9wXwJIyv7+Wr5wz9rVd/9fkf353e/83veyNZTBoYESOzpkwKR98su22fwotm91kbH6aq/fDdvTTgby+a3VvnV61ARwDcUSsUEmIzO3lPYdzidKJa1l+T/+SN/s9fXNtm027DzJ6IjuVARhiAg6Vx2iZwb3dDBRuh7tGjcHbUPTofdAo4LKc7H1yIw8r2zJoCPIDHzpQ5nSea7jNePzxq3tyBL/xEh7142YO9ck+GAH2F+WoWMnHUbRzWK82XN2cPTscvHn7/F7MqXfXz05vl0eTR4P88kbuPwmI43ZpXUlHkqF66Aarbun6Kfzn/+JVD/ba/vly+zFe333jim0dyMF413VdtAq+rMshkb2xPax2FPX/69NcfjeGyuf347dcPun+fH5093VzLcj7Ry2fbLk1nNqlct/uVtdy7dl/9Q44X9Hx5fvby3r3jIVWL6iwPgqmazu9efnedhnAY6mI7klAypf7Li/4P9akoHOv1doXiJ0t9MPsfYjgh+8gDyDkAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<PIL.JpegImagePlugin.JpegImageFile image mode=L size=48x48 at 0x7F602DF3A2E8>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kBbmZyDkvoMa"
      },
      "source": [
        "**Data Pre-Processing**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "p7ha7jFlvPwV",
        "outputId": "68b65a29-601c-40cb-b4f8-6b425758be61"
      },
      "source": [
        "# Create a dictionary of the classes and a reverse dictionary\n",
        "classes = dict()\n",
        "reverse_classes = dict()\n",
        "count = 0\n",
        "for folder in os.listdir(\"train\"):\n",
        "    classes[folder] = count\n",
        "    reverse_classes[count] = img\n",
        "    count = count + 1\n",
        "\n",
        "classes"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'Angry': 2, 'Happy': 4, 'Neutral': 1, 'Sad': 0, 'Surprise': 3}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lZyc0nM8l68M"
      },
      "source": [
        "BATCH_SIZE = 64\n",
        "\n",
        "# Since the data is separated into training and validation sets this time\n",
        "# we can use the ImageDataGenerator directly\n",
        "\n",
        "# Data augmentation for train only\n",
        "train_datagen = ImageDataGenerator(rescale=1./255,\n",
        "                                   rotation_range=30,\n",
        "                                   shear_range=0.3,\n",
        "                                   width_shift_range=0.3,\n",
        "                                   height_shift_range=0.3,\n",
        "                                   zoom_range=0.3,\n",
        "                                   horizontal_flip=True,\n",
        "                                   fill_mode=\"nearest\")\n",
        "# Resize only\n",
        "val_datagen = ImageDataGenerator(rescale=1./255)"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "e_e0F0g5nAwk",
        "outputId": "52b942c5-bda3-4e49-b04b-b9a40b3b1f05"
      },
      "source": [
        "# Create the generators for model.fit()\n",
        "train_generator = train_datagen.flow_from_directory(train_dir,\n",
        "                                                    color_mode='grayscale',\n",
        "                                                    target_size=(48, 48),\n",
        "                                                    batch_size=BATCH_SIZE,\n",
        "                                                    class_mode='categorical',\n",
        "                                                    shuffle=True)\n",
        "\n",
        "val_generator = val_datagen.flow_from_directory(val_dir,\n",
        "                                                color_mode='grayscale',\n",
        "                                                target_size=(48, 48),\n",
        "                                                batch_size=BATCH_SIZE,\n",
        "                                                class_mode='categorical',\n",
        "                                                shuffle=True)\n",
        "\n",
        "# Note the text shown\n",
        "num_train = 24256\n",
        "num_val = 3006"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 24256 images belonging to 5 classes.\n",
            "Found 3006 images belonging to 5 classes.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8NCTJyAtvlWi"
      },
      "source": [
        "**Create Model**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z2ubgKVGnd3t"
      },
      "source": [
        "# Apply batchnormalization before activation function\n",
        "# He normal initialization works best on layers w/ ReLU activation\n",
        "model = models.Sequential([\n",
        "    # Block 1\n",
        "    # Leave strides as default since images are small\n",
        "    Conv2D(32, (3, 3), padding=\"same\", kernel_initializer=\"he_normal\", input_shape=(48, 48, 1)),\n",
        "    BatchNormalization(),\n",
        "    Activation(\"relu\"),\n",
        "    Conv2D(32, (3, 3), padding=\"same\", kernel_initializer=\"he_normal\"),\n",
        "    BatchNormalization(),\n",
        "    Activation(\"relu\"),\n",
        "    MaxPool2D((2, 2)),\n",
        "    Dropout(0.25),\n",
        "\n",
        "    # Block 2\n",
        "    Conv2D(64, (3, 3), padding=\"same\", kernel_initializer=\"he_normal\"),\n",
        "    BatchNormalization(),\n",
        "    Activation(\"relu\"),\n",
        "    Conv2D(64, (3, 3), padding=\"same\", kernel_initializer=\"he_normal\"),\n",
        "    BatchNormalization(),\n",
        "    Activation(\"relu\"),\n",
        "    MaxPool2D((2, 2)),\n",
        "    Dropout(0.25),\n",
        "\n",
        "    # Block 3\n",
        "    Conv2D(128, (3, 3), padding=\"same\", kernel_initializer=\"he_normal\"),\n",
        "    BatchNormalization(),\n",
        "    Activation(\"relu\"),\n",
        "    Conv2D(128, (3, 3), padding=\"same\", kernel_initializer=\"he_normal\"),\n",
        "    BatchNormalization(),\n",
        "    Activation(\"relu\"),\n",
        "    MaxPool2D((2, 2)),\n",
        "    Dropout(0.25),\n",
        "\n",
        "    # Block 4\n",
        "    Conv2D(256, (3, 3), padding=\"same\", kernel_initializer=\"he_normal\"),\n",
        "    BatchNormalization(),\n",
        "    Activation(\"relu\"),\n",
        "    Conv2D(256, (3, 3), padding=\"same\", kernel_initializer=\"he_normal\"),\n",
        "    BatchNormalization(),\n",
        "    Activation(\"relu\"),\n",
        "    MaxPool2D((2, 2)),\n",
        "    Dropout(0.25),\n",
        "\n",
        "    # Block 5\n",
        "    Flatten(),\n",
        "    Dense(64, kernel_initializer=\"he_normal\"),\n",
        "    BatchNormalization(),\n",
        "    Activation(\"relu\"),\n",
        "    Dropout(0.5),\n",
        "\n",
        "    # Block 6\n",
        "    Dense(128, kernel_initializer=\"he_normal\"),\n",
        "    BatchNormalization(),\n",
        "    Activation(\"relu\"),\n",
        "    Dropout(0.5),\n",
        "\n",
        "    # Output layer\n",
        "    # Xavier (a.k.a. GlorotNormal) works better on layers w/ softmax activation\n",
        "    Dense(5, kernel_initializer=\"glorot_normal\", activation=\"softmax\")\n",
        "])"
      ],
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GWdGQrRMsaV-"
      },
      "source": [
        "model.compile(optimizer=Adam(lr=0.001), loss=\"categorical_crossentropy\", metrics=[\"accuracy\"])"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eaeoOv3qrbiP"
      },
      "source": [
        "Checkpoint = ModelCheckpoint(\"emotion_classifier.h5\",\n",
        "                             monitor=\"val_loss\",\n",
        "                             mode=\"min\",\n",
        "                             save_best_only=True,\n",
        "                             verbose=1)\n",
        "EarlyStopping = EarlyStopping(monitor=\"val_loss\",\n",
        "                              min_delta=0,\n",
        "                              patience=3,\n",
        "                              verbose=1,\n",
        "                              restore_best_weights=True)\n",
        "Reduce_LR = ReduceLROnPlateau(monitor=\"val_loss\",\n",
        "                              factor=0.2, # Decrease by 20% each time\n",
        "                              patience=3,\n",
        "                              verbose=1,\n",
        "                              min_delta=0.0001)"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NtKhSGySsVw5",
        "outputId": "529386e0-f851-4db1-c793-9d0f0e4d382c"
      },
      "source": [
        "history = model.fit(train_generator, \n",
        "                    steps_per_epoch=num_train//BATCH_SIZE,\n",
        "                    epochs=25,\n",
        "                    callbacks=[Checkpoint, EarlyStopping, Reduce_LR],\n",
        "                    validation_data=val_generator,\n",
        "                    validation_steps=num_val//BATCH_SIZE)"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/25\n",
            "379/379 [==============================] - ETA: 0s - loss: 1.7648 - accuracy: 0.2523\n",
            "Epoch 00001: val_loss improved from inf to 1.55815, saving model to emotion_classifier.h5\n",
            "379/379 [==============================] - 22s 58ms/step - loss: 1.7648 - accuracy: 0.2523 - val_loss: 1.5581 - val_accuracy: 0.2945\n",
            "Epoch 2/25\n",
            "379/379 [==============================] - ETA: 0s - loss: 1.6065 - accuracy: 0.2782\n",
            "Epoch 00002: val_loss did not improve from 1.55815\n",
            "379/379 [==============================] - 21s 56ms/step - loss: 1.6065 - accuracy: 0.2782 - val_loss: 1.5845 - val_accuracy: 0.2935\n",
            "Epoch 3/25\n",
            "379/379 [==============================] - ETA: 0s - loss: 1.5629 - accuracy: 0.2939\n",
            "Epoch 00003: val_loss did not improve from 1.55815\n",
            "379/379 [==============================] - 21s 55ms/step - loss: 1.5629 - accuracy: 0.2939 - val_loss: 1.5846 - val_accuracy: 0.2948\n",
            "Epoch 4/25\n",
            "379/379 [==============================] - ETA: 0s - loss: 1.5328 - accuracy: 0.3129\n",
            "Epoch 00004: val_loss improved from 1.55815 to 1.51793, saving model to emotion_classifier.h5\n",
            "379/379 [==============================] - 21s 56ms/step - loss: 1.5328 - accuracy: 0.3129 - val_loss: 1.5179 - val_accuracy: 0.3458\n",
            "Epoch 5/25\n",
            "379/379 [==============================] - ETA: 0s - loss: 1.4638 - accuracy: 0.3557\n",
            "Epoch 00005: val_loss improved from 1.51793 to 1.49259, saving model to emotion_classifier.h5\n",
            "379/379 [==============================] - 21s 57ms/step - loss: 1.4638 - accuracy: 0.3557 - val_loss: 1.4926 - val_accuracy: 0.3916\n",
            "Epoch 6/25\n",
            "378/379 [============================>.] - ETA: 0s - loss: 1.4094 - accuracy: 0.3880\n",
            "Epoch 00006: val_loss improved from 1.49259 to 1.37846, saving model to emotion_classifier.h5\n",
            "379/379 [==============================] - 22s 57ms/step - loss: 1.4096 - accuracy: 0.3878 - val_loss: 1.3785 - val_accuracy: 0.4188\n",
            "Epoch 7/25\n",
            "379/379 [==============================] - ETA: 0s - loss: 1.3411 - accuracy: 0.4308\n",
            "Epoch 00007: val_loss improved from 1.37846 to 1.29957, saving model to emotion_classifier.h5\n",
            "379/379 [==============================] - 21s 55ms/step - loss: 1.3411 - accuracy: 0.4308 - val_loss: 1.2996 - val_accuracy: 0.4511\n",
            "Epoch 8/25\n",
            "379/379 [==============================] - ETA: 0s - loss: 1.2745 - accuracy: 0.4697\n",
            "Epoch 00008: val_loss did not improve from 1.29957\n",
            "379/379 [==============================] - 21s 55ms/step - loss: 1.2745 - accuracy: 0.4697 - val_loss: 1.6050 - val_accuracy: 0.4742\n",
            "Epoch 9/25\n",
            "379/379 [==============================] - ETA: 0s - loss: 1.2181 - accuracy: 0.4976\n",
            "Epoch 00009: val_loss improved from 1.29957 to 1.28719, saving model to emotion_classifier.h5\n",
            "379/379 [==============================] - 21s 56ms/step - loss: 1.2181 - accuracy: 0.4976 - val_loss: 1.2872 - val_accuracy: 0.5248\n",
            "Epoch 10/25\n",
            "379/379 [==============================] - ETA: 0s - loss: 1.1831 - accuracy: 0.5185\n",
            "Epoch 00010: val_loss did not improve from 1.28719\n",
            "379/379 [==============================] - 21s 56ms/step - loss: 1.1831 - accuracy: 0.5185 - val_loss: 1.2952 - val_accuracy: 0.5577\n",
            "Epoch 11/25\n",
            "379/379 [==============================] - ETA: 0s - loss: 1.1511 - accuracy: 0.5354\n",
            "Epoch 00011: val_loss improved from 1.28719 to 1.22681, saving model to emotion_classifier.h5\n",
            "379/379 [==============================] - 21s 56ms/step - loss: 1.1511 - accuracy: 0.5354 - val_loss: 1.2268 - val_accuracy: 0.5537\n",
            "Epoch 12/25\n",
            "379/379 [==============================] - ETA: 0s - loss: 1.1246 - accuracy: 0.5479\n",
            "Epoch 00012: val_loss did not improve from 1.22681\n",
            "379/379 [==============================] - 21s 55ms/step - loss: 1.1246 - accuracy: 0.5479 - val_loss: 1.2918 - val_accuracy: 0.5577\n",
            "Epoch 13/25\n",
            "379/379 [==============================] - ETA: 0s - loss: 1.1062 - accuracy: 0.5571\n",
            "Epoch 00013: val_loss improved from 1.22681 to 1.20617, saving model to emotion_classifier.h5\n",
            "379/379 [==============================] - 21s 56ms/step - loss: 1.1062 - accuracy: 0.5571 - val_loss: 1.2062 - val_accuracy: 0.5642\n",
            "Epoch 14/25\n",
            "379/379 [==============================] - ETA: 0s - loss: 1.0925 - accuracy: 0.5665\n",
            "Epoch 00014: val_loss did not improve from 1.20617\n",
            "379/379 [==============================] - 21s 55ms/step - loss: 1.0925 - accuracy: 0.5665 - val_loss: 1.2751 - val_accuracy: 0.5275\n",
            "Epoch 15/25\n",
            "379/379 [==============================] - ETA: 0s - loss: 1.0700 - accuracy: 0.5766\n",
            "Epoch 00015: val_loss did not improve from 1.20617\n",
            "379/379 [==============================] - 21s 55ms/step - loss: 1.0700 - accuracy: 0.5766 - val_loss: 1.3305 - val_accuracy: 0.5516\n",
            "Epoch 16/25\n",
            "379/379 [==============================] - ETA: 0s - loss: 1.0528 - accuracy: 0.5833\n",
            "Epoch 00016: val_loss did not improve from 1.20617\n",
            "Restoring model weights from the end of the best epoch.\n",
            "\n",
            "Epoch 00016: ReduceLROnPlateau reducing learning rate to 0.00020000000949949026.\n",
            "379/379 [==============================] - 21s 55ms/step - loss: 1.0528 - accuracy: 0.5833 - val_loss: 1.2901 - val_accuracy: 0.5499\n",
            "Epoch 00016: early stopping\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vbRkSvKVuz-K"
      },
      "source": [
        "model.save(\"emotion_classifier.h5\")"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TphXya10vBn5",
        "outputId": "f0cddf5e-aa0a-4bce-9bdd-73e61ea91b40"
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d (Conv2D)              (None, 48, 48, 32)        320       \n",
            "_________________________________________________________________\n",
            "batch_normalization (BatchNo (None, 48, 48, 32)        128       \n",
            "_________________________________________________________________\n",
            "activation (Activation)      (None, 48, 48, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_1 (Conv2D)            (None, 48, 48, 32)        9248      \n",
            "_________________________________________________________________\n",
            "batch_normalization_1 (Batch (None, 48, 48, 32)        128       \n",
            "_________________________________________________________________\n",
            "activation_1 (Activation)    (None, 48, 48, 32)        0         \n",
            "_________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D) (None, 24, 24, 32)        0         \n",
            "_________________________________________________________________\n",
            "dropout (Dropout)            (None, 24, 24, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_2 (Conv2D)            (None, 24, 24, 64)        18496     \n",
            "_________________________________________________________________\n",
            "batch_normalization_2 (Batch (None, 24, 24, 64)        256       \n",
            "_________________________________________________________________\n",
            "activation_2 (Activation)    (None, 24, 24, 64)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_3 (Conv2D)            (None, 24, 24, 64)        36928     \n",
            "_________________________________________________________________\n",
            "batch_normalization_3 (Batch (None, 24, 24, 64)        256       \n",
            "_________________________________________________________________\n",
            "activation_3 (Activation)    (None, 24, 24, 64)        0         \n",
            "_________________________________________________________________\n",
            "max_pooling2d_1 (MaxPooling2 (None, 12, 12, 64)        0         \n",
            "_________________________________________________________________\n",
            "dropout_1 (Dropout)          (None, 12, 12, 64)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_4 (Conv2D)            (None, 12, 12, 128)       73856     \n",
            "_________________________________________________________________\n",
            "batch_normalization_4 (Batch (None, 12, 12, 128)       512       \n",
            "_________________________________________________________________\n",
            "activation_4 (Activation)    (None, 12, 12, 128)       0         \n",
            "_________________________________________________________________\n",
            "conv2d_5 (Conv2D)            (None, 12, 12, 128)       147584    \n",
            "_________________________________________________________________\n",
            "batch_normalization_5 (Batch (None, 12, 12, 128)       512       \n",
            "_________________________________________________________________\n",
            "activation_5 (Activation)    (None, 12, 12, 128)       0         \n",
            "_________________________________________________________________\n",
            "max_pooling2d_2 (MaxPooling2 (None, 6, 6, 128)         0         \n",
            "_________________________________________________________________\n",
            "dropout_2 (Dropout)          (None, 6, 6, 128)         0         \n",
            "_________________________________________________________________\n",
            "conv2d_6 (Conv2D)            (None, 6, 6, 256)         295168    \n",
            "_________________________________________________________________\n",
            "batch_normalization_6 (Batch (None, 6, 6, 256)         1024      \n",
            "_________________________________________________________________\n",
            "activation_6 (Activation)    (None, 6, 6, 256)         0         \n",
            "_________________________________________________________________\n",
            "conv2d_7 (Conv2D)            (None, 6, 6, 256)         590080    \n",
            "_________________________________________________________________\n",
            "batch_normalization_7 (Batch (None, 6, 6, 256)         1024      \n",
            "_________________________________________________________________\n",
            "activation_7 (Activation)    (None, 6, 6, 256)         0         \n",
            "_________________________________________________________________\n",
            "max_pooling2d_3 (MaxPooling2 (None, 3, 3, 256)         0         \n",
            "_________________________________________________________________\n",
            "dropout_3 (Dropout)          (None, 3, 3, 256)         0         \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 2304)              0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 64)                147520    \n",
            "_________________________________________________________________\n",
            "batch_normalization_8 (Batch (None, 64)                256       \n",
            "_________________________________________________________________\n",
            "activation_8 (Activation)    (None, 64)                0         \n",
            "_________________________________________________________________\n",
            "dropout_4 (Dropout)          (None, 64)                0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 128)               8320      \n",
            "_________________________________________________________________\n",
            "batch_normalization_9 (Batch (None, 128)               512       \n",
            "_________________________________________________________________\n",
            "activation_9 (Activation)    (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dropout_5 (Dropout)          (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 5)                 645       \n",
            "=================================================================\n",
            "Total params: 1,332,773\n",
            "Trainable params: 1,330,469\n",
            "Non-trainable params: 2,304\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 65
        },
        "id": "zJFENU0jvCsp",
        "outputId": "022f8a69-7e22-45ff-df06-2a39d2f53ceb"
      },
      "source": [
        "# The model predicted the sample image accurately\n",
        "image = Image.open(\"train/Angry/0.jpg\")\n",
        "image = image.resize((48, 48))\n",
        "arr = np.array(image)\n",
        "x_data = [arr]\n",
        "x_data = np.array(x_data, dtype = \"float32\")\n",
        "x_data = x_data.reshape((len(x_data), 48, 48, 1))\n",
        "x_data /= 255\n",
        "\n",
        "pred_array = model.predict(x_data)\n",
        "result = reverse_classes[np.argmax(pred_array)]\n",
        "result"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<PIL.JpegImagePlugin.JpegImageFile image mode=L size=48x48 at 0x7F602DF3A2E8>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "moldzwawvhyb"
      },
      "source": [
        "**Plot Training and Validation Graphs**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 499
        },
        "id": "QjsuQwjzveBO",
        "outputId": "afdbce9e-3338-4656-8372-136f34926faa"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "acc = history.history['accuracy']\n",
        "val_acc = history.history['val_accuracy']\n",
        "\n",
        "loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']\n",
        "\n",
        "epochs_range = range(16)\n",
        "\n",
        "plt.figure(figsize=(8, 8))\n",
        "plt.subplot(1, 2, 1)\n",
        "plt.plot(epochs_range, acc, label='Training Accuracy')\n",
        "plt.plot(epochs_range, val_acc, label='Validation Accuracy')\n",
        "plt.legend(loc='lower right')\n",
        "plt.title('Training and Validation Accuracy')\n",
        "\n",
        "plt.subplot(1, 2, 2)\n",
        "plt.plot(epochs_range, loss, label='Training Loss')\n",
        "plt.plot(epochs_range, val_loss, label='Validation Loss')\n",
        "plt.legend(loc='upper right')\n",
        "plt.title('Training and Validation Loss')\n",
        "plt.show()"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x576 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}