CNN_model_updated.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "JphD-Vck7fPd"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "from keras import optimizers, losses, activations, models\n",
        "from keras.layers import Activation\n",
        "from keras.callbacks import ModelCheckpoint, EarlyStopping, LearningRateScheduler\n",
        "from keras.layers import Dense, Input, Dropout, Convolution1D, MaxPool1D, GlobalMaxPool1D, GlobalAveragePooling1D,concatenate,BatchNormalization\n",
        "from keras import regularizers\n",
        "from keras.regularizers import l2\n",
        "from keras.regularizers import l1_l2\n",
        "from sklearn.preprocessing import StandardScaler,MinMaxScaler,RobustScaler\n",
        "import librosa\n",
        "#from playsound import playsound\n",
        "from tensorflow import keras\n",
        "import tensorflow as tf\n",
        "import glob\n",
        "import os\n",
        "#import noisereduce as nr\n",
        "from keras.callbacks import ReduceLROnPlateau\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "from sklearn.metrics import f1_score,recall_score,accuracy_score\n",
        "from sklearn.model_selection import train_test_split\n",
        "from tqdm import tqdm"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IaFSHkBC7fPl",
        "outputId": "3c51401d-f085-4ee3-84f3-d77cbd821b10"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/gdrive\n"
          ]
        }
      ],
      "source": [
        "#df=pd.read_csv(\"recordings-diagnosis.csv\")\n",
        "from google.colab import drive\n",
        "drive.mount('/content/gdrive')"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df = pd.read_csv('/content/gdrive/MyDrive/CS/third year/semester e/Voice Processing/mPower data/recordings-diagnosis.csv')\n",
        "Features = pd.read_csv('/content/gdrive/MyDrive/CS/third year/semester e/Voice Processing/Data/15k-features.csv')"
      ],
      "metadata": {
        "id": "vY9VkOof73TG"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "OjjLYDaW7fPm",
        "outputId": "41acb321-df61-4eb4-d3f0-870a68307c96",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 80
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   professional-diagnosis                                          file_name\n",
              "0                   False  audio_audio.m4a-b68fe4c0-db31-4b39-8f8f-7fb874..."
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-a7382c8c-291f-4250-b301-855278b17959\">\n",
              "    <div 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>professional-diagnosis</th>\n",
              "      <th>file_name</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>False</td>\n",
              "      <td>audio_audio.m4a-b68fe4c0-db31-4b39-8f8f-7fb874...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-a7382c8c-291f-4250-b301-855278b17959')\"\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 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\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",
              "    [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-a7382c8c-291f-4250-b301-855278b17959 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-a7382c8c-291f-4250-b301-855278b17959');\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",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ],
      "source": [
        "df.head(1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "YjtQHPH17fPn",
        "outputId": "c97419ec-5b09-43e4-c010-11e370d70ae8",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True     39271\n",
              "False    22785\n",
              "Name: professional-diagnosis, dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ],
      "source": [
        "df['professional-diagnosis'].value_counts()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "LH3rfBuB7fPo"
      },
      "outputs": [],
      "source": [
        "files_dir= \"wav_files\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "1DCx2e5V7fPo"
      },
      "outputs": [],
      "source": [
        "class CFG:\n",
        "    # batch_size and epochs\n",
        "    batch_size = 64\n",
        "    epochs = 50\n",
        "    patience = 35 \n",
        "    patience_for_plato = 10\n",
        "    model = 1 # 1 for Regular model. 2 for Resnet model. 3 for Attention model\n",
        "    feature_extract_train = False # To create new data\n",
        "    feature_extract_test = False\n",
        "    load_weights = True\n",
        "    save_weights = False"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ltg3m_Sx7fPp"
      },
      "outputs": [],
      "source": [
        "# data =[]\n",
        "# for path in os.listdir(files_dir):\n",
        "#     isPD = df[\"professional-diagnosis\"].loc[df[\"file_name\"]==path].values[0]\n",
        "#     data.append(Audio_File(os.path.join(files_dir,path),isPD))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1fMQLLxd7fPq"
      },
      "source": [
        "# Data Augmentation\n",
        "\n",
        "* Data augmentation is the process by which we create new synthetic data samples by adding small perturbations on our initial training set.\n",
        "\n",
        "* To generate syntactic data for audio, we can apply noise injection, shifting time, changing pitch and speed.\n",
        "\n",
        "* The objective is to make our model invariant to those perturbations and enhace its ability to generalize.\n",
        "\n",
        "* In order to this to work adding the perturbations must conserve the same label as the original training sample.\n",
        "* In images data augmention can be performed by shifting the image, zooming, rotating ...\n",
        "\n",
        "First, let's check which augmentation techniques works better for our dataset."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Pqcqh8tw7fPs"
      },
      "outputs": [],
      "source": [
        "def noise(data):\n",
        "    noise_amp = 0.035*np.random.uniform()*np.amax(data)\n",
        "    data = data + noise_amp*np.random.normal(size=data.shape[0])\n",
        "    return data\n",
        "\n",
        "def stretch(data, rate=0.8):\n",
        "    return librosa.effects.time_stretch(data, rate)\n",
        "\n",
        "def shift(data):\n",
        "    shift_range = int(np.random.uniform(low=-5, high = 5)*1000)\n",
        "    return np.roll(data, shift_range)\n",
        "\n",
        "def pitch(data, sampling_rate, pitch_factor=0.7):\n",
        "    return librosa.effects.pitch_shift(data, sampling_rate, pitch_factor)\n",
        "\n",
        "# taking any example and checking for techniques.\n",
        "path = '/content/gdrive/MyDrive/CS/third year/semester e/Voice Processing/Data/sample.wav'\n",
        "data, sample_rate = librosa.load(path)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VqURGkbr-1QF",
        "outputId": "becb9285-978b-4398-ffcd-2e028c70e427"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.00057509, 0.00098759, 0.00088838, ..., 0.00280021, 0.0022276 ,\n",
              "       0.00196877], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "sample_rate"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TLSQKZhp-28T",
        "outputId": "2a198547-c2e8-45f2-b570-900ef5112c74"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "22050"
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "60Zhr_1Z7fPt"
      },
      "source": [
        "## Feature Extraction\n",
        "- Extraction of features is a very important part in analyzing and finding relations between different things. As we already know that the data provided of audio cannot be understood by the models directly so we need to convert them into an understandable format for which feature extraction is used.\n",
        "\n",
        "\n",
        "The audio signal is a three-dimensional signal in which three axes represent time, amplitude and frequency.\n",
        "\n",
        "![image.png](https://miro.medium.com/max/633/1*7sKM9aECRmuoqTadCYVw9A.jpeg)\n",
        "\n",
        "I am no expert on audio signals and feature extraction on audio files so i need to search and found a very good blog written by [Askash Mallik](https://medium.com/heuristics/audio-signal-feature-extraction-and-clustering-935319d2225) on feature extraction.\n",
        "\n",
        "As stated there with the help of the sample rate and the sample data, one can perform several transformations on it to extract valuable features out of it.\n",
        "1. Zero Crossing Rate : The rate of sign-changes of the signal during the duration of a particular frame.\n",
        "2. Energy : The sum of squares of the signal values, normalized by the respective frame length.\n",
        "3. Entropy of Energy : The entropy of sub-frames’ normalized energies. It can be interpreted as a measure of abrupt changes.\n",
        "4. Spectral Centroid : The center of gravity of the spectrum.\n",
        "5. Spectral Spread : The second central moment of the spectrum.\n",
        "6. Spectral Entropy :  Entropy of the normalized spectral energies for a set of sub-frames.\n",
        "7. Spectral Flux : The squared difference between the normalized magnitudes of the spectra of the two successive frames.\n",
        "8. Spectral Rolloff : The frequency below which 90% of the magnitude distribution of the spectrum is concentrated.\n",
        "9.  MFCCs Mel Frequency Cepstral Coefficients form a cepstral representation where the frequency bands are not linear but distributed according to the mel-scale.\n",
        "10. Chroma Vector : A 12-element representation of the spectral energy where the bins represent the 12 equal-tempered pitch classes of western-type music (semitone spacing).\n",
        "11. Chroma Deviation : The standard deviation of the 12 chroma coefficients.\n",
        "\n",
        "\n",
        "In this project i am not going deep in feature selection process to check which features are good for our dataset rather i am only extracting 5 features:\n",
        "- Zero Crossing Rate\n",
        "- Chroma_stft\n",
        "- MFCC\n",
        "- RMS(root mean square) value\n",
        "- MelSpectogram to train our model."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Zw1NLIII7fPu"
      },
      "outputs": [],
      "source": [
        "def extract_features(data):\n",
        "    # ZCR\n",
        "    result = np.array([])\n",
        "    zcr = np.mean(librosa.feature.zero_crossing_rate(y=data).T, axis=0)\n",
        "    result=np.hstack((result, zcr)) # stacking horizontally\n",
        "   \n",
        "    # Chroma_stft\n",
        "    stft = np.abs(librosa.stft(data))\n",
        "    chroma_stft = np.mean(librosa.feature.chroma_stft(S=stft, sr=16000).T, axis=0)\n",
        "    result = np.hstack((result, chroma_stft)) # stacking horizontally\n",
        "\n",
        "    # MFCC\n",
        "    mfcc = np.mean(librosa.feature.mfcc(y=data, sr=16000).T, axis=0)\n",
        "    result = np.hstack((result, mfcc)) # stacking horizontally\n",
        "\n",
        "    # Root Mean Square Value\n",
        "    rms = np.mean(librosa.feature.rms(y=data).T, axis=0)\n",
        "    result = np.hstack((result, rms)) # stacking horizontally\n",
        "\n",
        "    # MelSpectogram\n",
        "    mel = np.mean(librosa.feature.melspectrogram(y=data, sr=16000).T, axis=0)\n",
        "    result = np.hstack((result, mel)) # stacking horizontally\n",
        "    \n",
        "\n",
        "    # Spectral Bandwidth\n",
        "    sb = np.mean(librosa.feature.spectral_bandwidth(y=data,sr=16000), axis=0)[:356]\n",
        "    result = np.hstack((result, sb)) # stacking horizontally\n",
        "\n",
        "    # Spectral Centroid\n",
        "    sc = np.mean(librosa.feature.spectral_centroid(y=data,sr=16000), axis=0)[:356]\n",
        "    result = np.hstack((result, sc)) # stacking horizontally\n",
        "   \n",
        "    return result\n",
        "\n",
        "def get_features(path):\n",
        "    # duration and offset are used to take care of the no audio in start and the ending of each audio files as seen above.\n",
        "    data, sample_rate = librosa.load(path, duration=10, offset=0.6)\n",
        "    try:\n",
        "        # without augmentation\n",
        "        res1 = extract_features(data)\n",
        "        result = np.array(res1)\n",
        "        \n",
        "        # data with noise\n",
        "        noise_data = noise(data)\n",
        "        res2 = extract_features(noise_data)\n",
        "        result = np.vstack((result, res2)) # stacking vertically\n",
        "        \n",
        "        # data with stretching and pitching\n",
        "        new_data = stretch(data)\n",
        "        data_stretch_pitch = pitch(new_data, sample_rate)\n",
        "        res3 = extract_features(data_stretch_pitch)\n",
        "        result = np.vstack((result, res3)) # stacking vertically\n",
        "    except Exception as e:\n",
        "        print(e)\n",
        "    return result"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "lSZ81I7y7fPv"
      },
      "outputs": [],
      "source": [
        "if CFG.feature_extract_train:\n",
        "    X, Y = [], []\n",
        "    for i,file in enumerate(os.listdir(files_dir)[:5000]):\n",
        "        feature = get_features(os.path.join(files_dir,file))\n",
        "        for ele in feature:\n",
        "            X.append(ele)\n",
        "            # appending emotion 3 times as we have made 3 augmentation techniques on each audio file.\n",
        "            Y.append(df[\"professional-diagnosis\"].loc[df[\"file_name\"]==file].values[0])\n",
        "        if(i%100==0):\n",
        "            print(i)\n",
        "    len(X), len(Y)\n",
        "    Features = pd.DataFrame(X)\n",
        "    Features['labels'] = Y\n",
        "    Features.to_csv('features.csv', index=False)\n",
        "    Features.head()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "BNLo4bqU7fPw"
      },
      "outputs": [],
      "source": [
        "#Features = pd.read_csv(\"features.csv\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "E-Sq8YEd7fPw"
      },
      "source": [
        "## Data Preparation\n",
        "\n",
        "- As of now we have extracted the data, now we need to normalize and split our data for training and testing."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "e-ZJIGQR7fPx"
      },
      "outputs": [],
      "source": [
        "X = Features.iloc[: ,:-1].values\n",
        "Y = Features['labels'].values"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "AwpJI-yb7fPx",
        "outputId": "910b3635-be49-4cdf-e57c-4999854b3fc1",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((40500, 874), (40500,), (4500, 874), (4500,))"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ],
      "source": [
        "# splitting data\n",
        "x_train, x_val, y_train, y_val = train_test_split(X, Y, random_state=242,test_size=0.1)\n",
        "x_train.shape, y_train.shape, x_val.shape, y_val.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "T2aauRyu7fPx",
        "outputId": "c31b2b32-d9c1-47be-beab-4003bc287213",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((40500, 874), (40500,), (4500, 874), (4500,))"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ],
      "source": [
        "scaler = MinMaxScaler()\n",
        "x_train = scaler.fit_transform(x_train)\n",
        "x_val = scaler.transform(x_val)\n",
        "x_train.shape, y_train.shape, x_val.shape, y_val.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "D1icmIjV7fPy",
        "outputId": "55e48809-f263-4d85-f378-ecf4bd23d0b0",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((40500, 874, 1), (40500,), (4500, 874, 1), (4500,))"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ],
      "source": [
        "# making our data compatible to model.\n",
        "x_train = np.expand_dims(x_train, axis=2)\n",
        "x_val = np.expand_dims(x_val, axis=2)\n",
        "x_train.shape, y_train.shape, x_val.shape, y_val.shape"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "25I3KJTb7fPy"
      },
      "source": [
        "## Modelling\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "_9n_oD317fPy"
      },
      "outputs": [],
      "source": [
        "\n",
        "def conv1d(x, filters, kernel_size, name, stride=1, activation=tf.nn.elu, padding='SAME'):\n",
        "    return tf.keras.layers.Convolution1D( filters, kernel_size, strides=stride, padding='SAME',name=name,kernel_regularizer=regularizers.L2(0.001))(x)\n",
        "\n",
        "def maxpool1d(x, pool_size, strides, name, padding='SAME'):\n",
        "    return tf.keras.layers.MaxPool1D( pool_size, strides, padding='SAME', name=name)(x)\n",
        "\n",
        "def resblock(x, filters, kernel_size, name):\n",
        "    resblock=conv1d(x, filters, kernel_size, stride=1, name=name+'_conv1')\n",
        "    resblock=tf.keras.layers.BatchNormalization( name=name+'_batch1')(resblock)\n",
        "    resblock=tf.nn.relu(resblock, name=name+'_relu')\n",
        "    resblock=conv1d(resblock, filters, kernel_size, stride=1, name=name+'_conv2')\n",
        "    resblock=tf.keras.layers.BatchNormalization( name=name+'_batch2')(resblock)\n",
        "    return tf.nn.relu(resblock+x)\n",
        "\n",
        "def fullylayer(x, units, name):\n",
        "    return tf.keras.layers.dense(x, units, activation=tf.nn.elu, name=name)\n",
        "\n",
        "def conv2ser(x):\n",
        "    return tf.keras.layers.GlobalAveragePooling1D()(x)\n",
        "\n",
        "def resblock_s(x, filters, kernel_size, name):\n",
        "    resblock=conv1d(x, filters, kernel_size, stride=2, name=name+'_conv1')\n",
        "    shorcut=tf.keras.layers.BatchNormalization( name=name+'_batch1')(resblock)\n",
        "    resblock=tf.nn.relu(shorcut, name=name+'_relu')\n",
        "    resblock=conv1d(resblock, filters, kernel_size, stride=1, name=name+'_conv2')\n",
        "    resblock=tf.keras.layers.BatchNormalization( name=name+'_batch2')(resblock)\n",
        "    return tf.nn.relu(resblock+shorcut)  "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "txXEEeS67fPz"
      },
      "outputs": [],
      "source": [
        "class BaseModel:\n",
        "    def __init__(self, num_class):\n",
        "        raise NotImplementedError\n",
        "        \n",
        "    def run(self, X, is_training):\n",
        "        raise NotImplementedError\n",
        "\n",
        "class Resnet1D(BaseModel):\n",
        "    params={}\n",
        "    def __init__(self, num_class, keep_prob):\n",
        "        self.shape=[-1, 74484, 1]\n",
        "        self.keep_prob=keep_prob\n",
        "        \n",
        "    \n",
        "        "
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def get_model():\n",
        "    if CFG.model==1:\n",
        "        inp = Input(shape=(874, 1))\n",
        "        img_1 = Convolution1D(256, kernel_size=5,strides =1, activation=activations.relu, padding=\"same\",kernel_initializer=\"he_normal\")(inp)\n",
        "        img_1 = MaxPool1D(pool_size=5,strides = 2,padding=\"same\")(img_1)\n",
        "        img_1 = BatchNormalization()(img_1)\n",
        "        img_1 = Convolution1D(128, kernel_size=5,strides =1 ,activation=activations.relu, padding=\"same\",kernel_initializer=\"he_normal\")(img_1)\n",
        "        img_1 = MaxPool1D(pool_size=5,strides = 2)(img_1)\n",
        "        img_1 = BatchNormalization()(img_1)\n",
        "        img_1 = Convolution1D(128, kernel_size=5,strides =1 ,activation=activations.relu, padding=\"same\",kernel_initializer=\"he_normal\")(img_1)\n",
        "        img_1 = MaxPool1D(pool_size=5,strides = 2,padding=\"same\")(img_1)\n",
        "        img_1 = BatchNormalization()(img_1)\n",
        "        img_1 = Convolution1D(256, kernel_size=5,strides =1, activation=activations.relu, padding=\"same\",kernel_initializer=\"he_normal\")(img_1)\n",
        "        img_1 = GlobalMaxPool1D()(img_1)\n",
        "        b = BatchNormalization()(img_1)\n",
        "        dense_1 = Dense(64, activation=activations.relu)(b)\n",
        "        dense_1 = Dense(32, activation=activations.relu)(dense_1)\n",
        "        dense_1 = Dense(1, activation=activations.sigmoid)(dense_1)\n",
        "\n",
        "        model = models.Model(inputs=inp, outputs=dense_1)\n",
        "        # opt = optimizers.Adam(0.0001)\n",
        "\n",
        "        model.compile(optimizer=keras.optimizers.Adam(learning_rate=0.001), loss=losses.binary_crossentropy, metrics=['acc'])\n",
        "        return model\n",
        "    elif CFG.model==2: # ResNet\n",
        "        inp = Input(shape=(x_train.shape[1], 1))\n",
        "            \n",
        "        #conv1d\n",
        "        conv1=conv1d(inp, 64, 7, name='conv1')\n",
        "\n",
        "        #maxpool1d\n",
        "        maxpool1=maxpool1d(conv1, 4, strides=2, name='maxpool1',padding='same')\n",
        "        #resblock1,2,3\n",
        "        resblock1=resblock(maxpool1, 64, 3, name='resblock1')\n",
        "        resblock2=resblock(resblock1, 64, 3, name='resblock2')\n",
        "        resblock3=resblock(resblock2, 64, 3, name='resblock3')\n",
        "        maxpool2=maxpool1d(resblock3, 4, strides=4, name='maxpool2', padding='same')\n",
        "        \n",
        "        #resblock4,5,6,7\n",
        "        resblock4=resblock_s(maxpool2, 128, 3, name='resblock4_s2')\n",
        "        resblock5=resblock(resblock4, 128, 3, name='resblock5')\n",
        "        resblock6=resblock(resblock5, 128, 3, name='resblock6')\n",
        "        resblock7=resblock(resblock6, 128, 3, name='resblock7')\n",
        "        maxpool3=maxpool1d(resblock7, 4, strides=4, name='maxpool3', padding='same')\n",
        "        # #resblock8~13\n",
        "        resblock8=resblock_s(maxpool3, 256, 3, name='resblock8_s2')\n",
        "        resblock9=resblock(resblock8, 256, 3, name='resblock9')\n",
        "        resblock10=resblock(resblock9, 256, 3, name='resblock10')\n",
        "        resblock11=resblock(resblock10, 256, 3, name='resblock11')\n",
        "        # resblock12=resblock(resblock11, 256, 3, name='resblock12')\n",
        "        # resblock13=resblock(resblock12, 256, 3, name='resblock13')\n",
        "        maxpool4=maxpool1d(resblock11, 4, strides=4, name='maxpool4', padding='same')\n",
        "    \n",
        "        # #resblock14~16\n",
        "        resblock14=resblock_s(maxpool4, 512, 3, name='resblock14_s2')\n",
        "        resblock15=resblock(resblock14, 512, 3, name='resblock15')\n",
        "        resblock16=resblock(resblock15, 512, 3, name='resblock16')\n",
        "        maxpool5=maxpool1d(resblock16, 4, strides=4, name='maxpool5', padding='same')\n",
        "        \n",
        "        # #resblock17~19\n",
        "        # resblock17=resblock_s(maxpool5, 1024, 3, name='resblock17_s2')\n",
        "        # resblock18=resblock(resblock17, 1024, 3, name='resblock18')\n",
        "        # resblock19=resblock(resblock18, 1024, 3, name='resblock19')\n",
        "        # maxpool6=maxpool1d(resblock19, 4, strides=4, name='maxpool6', padding='VALID')\n",
        "        \n",
        "        # #resblock\n",
        "        # resblock20=resblock_s(maxpool6, 1024, 3, name='resblock20_s2')\n",
        "        # resblock21=resblock_s(resblock20, 1024, 3, name='resblock21')\n",
        "        # resblock22=resblock(resblock21, 1024, 3, name='resblock22')\n",
        "        # resblock23=resblock(resblock22, 1024, 3, name='resblock23')\n",
        "        # maxpool7=maxpool1d(resblock23, 4, strides=4, name='maxpool7', padding='VALID')\n",
        "        # maxpool8=maxpool1d(maxpool7, 3, strides=3, name='maxpool8', padding='VALID')\n",
        "            \n",
        "        avgpool_flat=conv2ser(maxpool5)\n",
        "\n",
        "        #fullyconectedlyaer\n",
        "        dense_1 = Dense(256, activation=activations.relu)(avgpool_flat)\n",
        "        dense_1 = Dense(128, activation=activations.relu)(dense_1)\n",
        "        dense_1 = Dense(1, activation=activations.sigmoid)(dense_1)\n",
        "        model = models.Model(inputs=inp, outputs=dense_1)\n",
        "        # opt = optimizers.Adam(0.0001)\n",
        "\n",
        "        model.compile(optimizer=keras.optimizers.Nadam(learning_rate=0.001), loss=losses.binary_crossentropy, metrics=['acc'])\n",
        "        return model"
      ],
      "metadata": {
        "id": "YdqKHx1RbSV8"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ifcWBlKC7fPz"
      },
      "outputs": [],
      "source": [
        "# def get_model():##\n",
        "#     if CFG.model==1:\n",
        "#         l=0.001\n",
        "#         inp = Input(shape=(874, 1))\n",
        "#         img_1 = Convolution1D(256, kernel_size=5,strides=1, padding=\"same\",kernel_initializer=\"he_normal\")(inp)\n",
        "#         img_1 = BatchNormalization()(img_1)\n",
        "#         img_1 = Activation(\"relu\")(img_1)\n",
        "#         img_1 = MaxPool1D(pool_size=2,strides = 2,padding=\"same\")(img_1)\n",
        "#         img_1 = Convolution1D(128, kernel_size=5,strides=1, padding=\"same\",kernel_initializer=\"he_normal\",\n",
        "#                               kernel_regularizer=l1_l2(l1=l,l2=l))(img_1) \n",
        "#         img_1 = BatchNormalization()(img_1)\n",
        "#         img_1 = Activation(\"relu\")(img_1)\n",
        "#         #img_1 = Dropout(0.2)(img_1)\n",
        "#         img_1 = MaxPool1D(pool_size=2,strides = 2,padding=\"same\")(img_1)\n",
        "#         img_1 = Convolution1D(128, kernel_size=5,strides =1, padding=\"same\",kernel_initializer=\"he_normal\",\n",
        "#                               kernel_regularizer=l1_l2(l1=l,l2=l))(img_1)\n",
        "#         img_1 = BatchNormalization()(img_1)\n",
        "#         img_1 = Activation(\"relu\")(img_1)\n",
        "#         #img_1 = Dropout(0.2)(img_1)\n",
        "#         img_1 = MaxPool1D(pool_size=2,strides = 2,padding=\"same\")(img_1)\n",
        "#         #\n",
        "#         img_1 = Convolution1D(256, kernel_size=5,strides =1, padding=\"same\",kernel_initializer=\"he_normal\", \n",
        "#                               kernel_regularizer=l1_l2(l1=l,l2=l))(img_1)\n",
        "#         img_1 = BatchNormalization()(img_1)\n",
        "#         img_1 = Activation(\"relu\")(img_1)\n",
        "#         #img_1 = Dropout(0.2)(img_1)\n",
        "#         img_1 = MaxPool1D(pool_size=2,strides = 2,padding=\"same\")(img_1)\n",
        "#         # img_1 = Convolution1D(512, kernel_size=1,strides =1, padding=\"same\",kernel_initializer=\"he_normal\")(img_1)\n",
        "#         # img_1 = BatchNormalization()(img_1)\n",
        "#         # img_1 = Activation(\"relu\")(img_1)\n",
        "#         #\n",
        "#         img_1 = GlobalMaxPool1D()(img_1)\n",
        "#         img_1 = BatchNormalization()(img_1)\n",
        "#         dense_1 = Dense(128, activation=activations.relu)(img_1)\n",
        "#         dense_1 = Dense(64, activation=activations.relu)(img_1)\n",
        "#         dense_1 = Dense(32, activation=activations.relu)(dense_1)\n",
        "#         dense_1 = Dense(1, activation=activations.sigmoid)(dense_1)\n",
        "\n",
        "#         model = models.Model(inputs=inp, outputs=dense_1)\n",
        "#         # opt = optimizers.Adam(0.0001)\n",
        "\n",
        "#         model.compile(optimizer=keras.optimizers.Adam(learning_rate=0.001), loss=losses.binary_crossentropy, metrics=['acc'])\n",
        "#         return model\n",
        "#     elif CFG.model==2: # ResNet\n",
        "#         inp = Input(shape=(x_train.shape[1], 1))\n",
        "            \n",
        "#         #conv1d\n",
        "#         conv1=conv1d(inp, 64, 7, name='conv1')\n",
        "\n",
        "#         #maxpool1d\n",
        "#         maxpool1=maxpool1d(conv1, 4, strides=2, name='maxpool1',padding='same')\n",
        "#         #resblock1,2,3\n",
        "#         resblock1=resblock(maxpool1, 64, 3, name='resblock1')\n",
        "#         resblock2=resblock(resblock1, 64, 3, name='resblock2')\n",
        "#         resblock3=resblock(resblock2, 64, 3, name='resblock3')\n",
        "#         maxpool2=maxpool1d(resblock3, 4, strides=4, name='maxpool2', padding='same')\n",
        "        \n",
        "#         #resblock4,5,6,7\n",
        "#         resblock4=resblock_s(maxpool2, 128, 3, name='resblock4_s2')\n",
        "#         resblock5=resblock(resblock4, 128, 3, name='resblock5')\n",
        "#         resblock6=resblock(resblock5, 128, 3, name='resblock6')\n",
        "#         resblock7=resblock(resblock6, 128, 3, name='resblock7')\n",
        "#         maxpool3=maxpool1d(resblock7, 4, strides=4, name='maxpool3', padding='same')\n",
        "#         # #resblock8~13\n",
        "#         resblock8=resblock_s(maxpool3, 256, 3, name='resblock8_s2')\n",
        "#         resblock9=resblock(resblock8, 256, 3, name='resblock9')\n",
        "#         resblock10=resblock(resblock9, 256, 3, name='resblock10')\n",
        "#         resblock11=resblock(resblock10, 256, 3, name='resblock11')\n",
        "#         # resblock12=resblock(resblock11, 256, 3, name='resblock12')\n",
        "#         # resblock13=resblock(resblock12, 256, 3, name='resblock13')\n",
        "#         maxpool4=maxpool1d(resblock11, 4, strides=4, name='maxpool4', padding='same')\n",
        "    \n",
        "#         # #resblock14~16\n",
        "#         resblock14=resblock_s(maxpool4, 512, 3, name='resblock14_s2')\n",
        "#         resblock15=resblock(resblock14, 512, 3, name='resblock15')\n",
        "#         resblock16=resblock(resblock15, 512, 3, name='resblock16')\n",
        "#         maxpool5=maxpool1d(resblock16, 4, strides=4, name='maxpool5', padding='same')\n",
        "        \n",
        "#         # #resblock17~19\n",
        "#         # resblock17=resblock_s(maxpool5, 1024, 3, name='resblock17_s2')\n",
        "#         # resblock18=resblock(resblock17, 1024, 3, name='resblock18')\n",
        "#         # resblock19=resblock(resblock18, 1024, 3, name='resblock19')\n",
        "#         # maxpool6=maxpool1d(resblock19, 4, strides=4, name='maxpool6', padding='VALID')\n",
        "        \n",
        "#         # #resblock\n",
        "#         # resblock20=resblock_s(maxpool6, 1024, 3, name='resblock20_s2')\n",
        "#         # resblock21=resblock_s(resblock20, 1024, 3, name='resblock21')\n",
        "#         # resblock22=resblock(resblock21, 1024, 3, name='resblock22')\n",
        "#         # resblock23=resblock(resblock22, 1024, 3, name='resblock23')\n",
        "#         # maxpool7=maxpool1d(resblock23, 4, strides=4, name='maxpool7', padding='VALID')\n",
        "#         # maxpool8=maxpool1d(maxpool7, 3, strides=3, name='maxpool8', padding='VALID')\n",
        "            \n",
        "#         avgpool_flat=conv2ser(maxpool5)\n",
        "\n",
        "#         #fullyconectedlyaer\n",
        "#         dense_1 = Dense(256, activation=activations.relu)(avgpool_flat)\n",
        "#         dense_1 = Dense(128, activation=activations.relu)(dense_1)\n",
        "#         dense_1 = Dense(1, activation=activations.sigmoid)(dense_1)\n",
        "#         model = models.Model(inputs=inp, outputs=dense_1)\n",
        "#         # opt = optimizers.Adam(0.0001)\n",
        "\n",
        "#         return model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "PP4Sm2VY7fP0"
      },
      "outputs": [],
      "source": [
        "def Conv_1D_Block(x, model_width, kernel, strides):\n",
        "    # 1D Convolutional Block with BatchNormalization\n",
        "    x = tf.keras.layers.Conv1D(model_width, kernel, strides=strides, padding=\"same\", kernel_initializer=\"he_normal\")(x)\n",
        "    x = tf.keras.layers.BatchNormalization()(x)\n",
        "    x = tf.keras.layers.Activation('relu')(x)\n",
        "\n",
        "    return x\n",
        "\n",
        "\n",
        "def trans_conv1D(x, model_width, kernel, strides):\n",
        "    # 1D Transposed Convolutional Block, used instead of UpSampling\n",
        "    x = tf.keras.layers.Conv1DTranspose(model_width, kernel, strides=strides, padding='same')(x)  # Stride = 2, Kernel Size = 4\n",
        "    x = tf.keras.layers.BatchNormalization()(x)\n",
        "    x = tf.keras.layers.Activation('relu')(x)\n",
        "\n",
        "    return x\n",
        "\n",
        "def Feature_Extraction_Block(inputs, model_width, feature_number):\n",
        "    # Feature Extraction Block for the AutoEncoder Mode\n",
        "    shape = inputs.shape\n",
        "    latent = tf.keras.layers.Flatten()(inputs)\n",
        "    latent = tf.keras.layers.Dense(feature_number, name='features')(latent)\n",
        "    latent = tf.keras.layers.Dense(model_width * shape[1])(latent)\n",
        "    latent = tf.keras.layers.Reshape((shape[1], model_width))(latent)\n",
        "\n",
        "    return latent\n",
        "\n",
        "\n",
        "def Attention_Block(skip_connection, gating_signal, num_filters):\n",
        "    # Attention Block\n",
        "    conv1x1_1 = tf.keras.layers.Conv1D(num_filters, 1, strides=2)(skip_connection)\n",
        "    conv1x1_1 = tf.keras.layers.BatchNormalization()(conv1x1_1)\n",
        "    conv1x1_2 = tf.keras.layers.Conv1D(num_filters, 1, strides=1)(gating_signal)\n",
        "    conv1x1_2 = tf.keras.layers.BatchNormalization()(conv1x1_2)\n",
        "    conv1_2 = tf.keras.layers.add([conv1x1_1, conv1x1_2])\n",
        "    conv1_2 = tf.keras.layers.Activation('relu')(conv1_2)\n",
        "    conv1_2 = tf.keras.layers.Conv1D(1, 1, strides=1)(conv1_2)\n",
        "    conv1_2 = tf.keras.layers.BatchNormalization()(conv1_2)\n",
        "    conv1_2 = tf.keras.layers.Activation('sigmoid')(conv1_2)\n",
        "    resampler = trans_conv1D(conv1_2, num_filters, 4, 1)\n",
        "    out = skip_connection*resampler\n",
        "\n",
        "    return out\n",
        "\n",
        "\n",
        "def stem(inputs, num_filters):\n",
        "    # Construct the Stem Convolution Group\n",
        "    # inputs : input vector\n",
        "    # First Convolutional layer, where pooled feature maps will be reduced by 75%\n",
        "    conv = Conv_1D_Block(inputs, num_filters, 7, 2)\n",
        "    if conv.shape[1] <= 2:\n",
        "        pool = tf.keras.layers.MaxPooling1D(pool_size=1, strides=2, padding=\"valid\")(conv)\n",
        "    else:\n",
        "        pool = tf.keras.layers.MaxPooling1D(pool_size=2, strides=2, padding=\"valid\")(conv)\n",
        "\n",
        "    return pool\n",
        "\n",
        "\n",
        "def conv_block(inputs, num_filters):\n",
        "    # Construct Block of Convolutions without Pooling\n",
        "    # x        : input into the block\n",
        "    # n_filters: number of filters\n",
        "    conv = Conv_1D_Block(inputs, num_filters, 3, 2)\n",
        "    conv = Conv_1D_Block(conv, num_filters, 3, 1)\n",
        "    conv = Conv_1D_Block(conv, num_filters, 3, 1)\n",
        "\n",
        "    return conv\n",
        "\n",
        "\n",
        "def residual_block(inputs, num_filters):\n",
        "    # Construct a Residual Block of Convolutions\n",
        "    # x        : input into the block\n",
        "    # n_filters: number of filters\n",
        "    shortcut = inputs\n",
        "    #\n",
        "    conv = Conv_1D_Block(inputs, num_filters, 3, 1)\n",
        "    conv = Conv_1D_Block(conv, num_filters, 3, 1)\n",
        "    conv = tf.keras.layers.Add()([conv, shortcut])\n",
        "    out = tf.keras.layers.Activation('relu')(conv)\n",
        "\n",
        "    return out\n",
        "\n",
        "\n",
        "def residual_group(inputs, num_filters, n_blocks, conv=True):\n",
        "    # x        : input to the group\n",
        "    # n_filters: number of filters\n",
        "    # n_blocks : number of blocks in the group\n",
        "    # conv     : flag to include the convolution block connector\n",
        "    out = inputs\n",
        "    for _ in range(n_blocks):\n",
        "        out = residual_block(out, num_filters)\n",
        "\n",
        "    # Double the size of filters and reduce feature maps by 75% (strides=2, 2) to fit the next Residual Group\n",
        "    if conv:\n",
        "        out = conv_block(out, num_filters * 2)\n",
        "\n",
        "    return out\n",
        "\n",
        "\n",
        "def stem_bottleneck(inputs, num_filters):\n",
        "    # Construct the Stem Convolution Group\n",
        "    # inputs : input vector\n",
        "    # First Convolutional layer, where pooled feature maps will be reduced by 75%\n",
        "    conv = Conv_1D_Block(inputs, num_filters, 7, 2)\n",
        "    if conv.shape[1] <= 2:\n",
        "        pool = tf.keras.layers.MaxPooling1D(pool_size=1, strides=2, padding=\"valid\")(conv)\n",
        "    else:\n",
        "        pool = tf.keras.layers.MaxPooling1D(pool_size=2, strides=2, padding=\"valid\")(conv)\n",
        "\n",
        "    return pool\n",
        "\n",
        "\n",
        "def residual_block_bottleneck(inputs, num_filters):\n",
        "    # Construct a Residual Block of Convolutions\n",
        "    # x        : input into the block\n",
        "    # n_filters: number of filters\n",
        "    shortcut = Conv_1D_Block(inputs, num_filters * 4, 1, 1)\n",
        "    #\n",
        "    conv = Conv_1D_Block(inputs, num_filters, 1, 1)\n",
        "    conv = Conv_1D_Block(conv, num_filters, 3, 1)\n",
        "    conv = Conv_1D_Block(conv, num_filters * 4, 1, 1)\n",
        "    conv = tf.keras.layers.Add()([conv, shortcut])\n",
        "    out = tf.keras.layers.Activation('relu')(conv)\n",
        "\n",
        "    return out\n",
        "\n",
        "\n",
        "def residual_group_bottleneck(inputs, num_filters, n_blocks, conv=True):\n",
        "    # x        : input to the group\n",
        "    # n_filters: number of filters\n",
        "    # n_blocks : number of blocks in the group\n",
        "    # conv     : flag to include the convolution block connector\n",
        "    out = inputs\n",
        "    for i in range(n_blocks):\n",
        "        out = residual_block_bottleneck(out, num_filters)\n",
        "\n",
        "    # Double the size of filters and reduce feature maps by 75% (strides=2) to fit the next Residual Group\n",
        "    if conv:\n",
        "        out = conv_block(out, num_filters * 2)\n",
        "\n",
        "    return out\n",
        "\n",
        "\n",
        "def encoder18(inputs, num_filters):\n",
        "    # Construct the Learner\n",
        "    x1 = residual_group(inputs, num_filters, 2)          # First Residual Block Group of 64 filters\n",
        "    x2 = residual_group(x1, num_filters * 2, 1)           # Second Residual Block Group of 128 filters\n",
        "    x3 = residual_group(x2, num_filters * 4, 1)           # Third Residual Block Group of 256 filters\n",
        "    x4 = residual_group(x3, num_filters * 8, 1, False)  # Fourth Residual Block Group of 512 filters\n",
        "\n",
        "    return x1, x2, x3, x4\n",
        "\n",
        "\n",
        "def encoder34(inputs, num_filters):\n",
        "    # Construct the Learner\n",
        "    x1 = residual_group(inputs, num_filters, 3)          # First Residual Block Group of 64 filters\n",
        "    x2 = residual_group(x1, num_filters * 2, 3)           # Second Residual Block Group of 128 filters\n",
        "    x3 = residual_group(x2, num_filters * 4, 5)           # Third Residual Block Group of 256 filters\n",
        "    x4 = residual_group(x3, num_filters * 8, 2, False)  # Fourth Residual Block Group of 512 filters\n",
        "\n",
        "    return x1, x2, x3, x4\n",
        "\n",
        "\n",
        "def encoder50(inputs, num_filters):\n",
        "    # Construct the Learner\n",
        "    x1 = residual_group_bottleneck(inputs, num_filters, 3)  # First Residual Block Group of 64 filters\n",
        "    x2 = residual_group_bottleneck(x1, num_filters * 2, 3)   # Second Residual Block Group of 128 filters\n",
        "    x3 = residual_group_bottleneck(x2, num_filters * 4, 5)   # Third Residual Block Group of 256 filters\n",
        "    x4 = residual_group_bottleneck(x3, num_filters * 8, 2, False)  # Fourth Residual Block Group of 512 filters\n",
        "\n",
        "    return x1, x2, x3, x4\n",
        "\n",
        "\n",
        "def encoder101(inputs, num_filters):\n",
        "    # Construct the Learner\n",
        "    x1 = residual_group_bottleneck(inputs, num_filters, 3)  # First Residual Block Group of 64 filters\n",
        "    x2 = residual_group_bottleneck(x1, num_filters * 2, 3)   # Second Residual Block Group of 128 filters\n",
        "    x3 = residual_group_bottleneck(x2, num_filters * 4, 22)  # Third Residual Block Group of 256 filters\n",
        "    x4 = residual_group_bottleneck(x3, num_filters * 8, 2, False)  # Fourth Residual Block Group of 512 filters\n",
        "\n",
        "    return x1, x2, x3, x4\n",
        "\n",
        "\n",
        "def encoder152(inputs, num_filters):\n",
        "    # Construct the Learner\n",
        "    x1 = residual_group_bottleneck(inputs, num_filters, 3)  # First Residual Block Group of 64 filters\n",
        "    x2 = residual_group_bottleneck(x1, num_filters * 2, 7)   # Second Residual Block Group of 128 filters\n",
        "    x3 = residual_group_bottleneck(x2, num_filters * 4, 35)  # Third Residual Block Group of 256 filters\n",
        "    x4 = residual_group_bottleneck(x3, num_filters * 8, 2, False)  # Fourth Residual Block Group of 512 filters\n",
        "\n",
        "    return x1, x2, x3, x4\n",
        "\n",
        "\n",
        "def decoder_unit(inputs, num_filters):\n",
        "    # Decoder Unit\n",
        "    conv = Conv_1D_Block(inputs, num_filters, 1, 1)\n",
        "    deconv = trans_conv1D(conv, num_filters, 4, 2)\n",
        "    conv = Conv_1D_Block(deconv, num_filters, 1, 1)\n",
        "\n",
        "    return conv\n",
        "\n",
        "\n",
        "def decoder_block(x1, x2, x3, x4, num_filters, D_S, A_G):\n",
        "    # Decoder Block\n",
        "    levels = []\n",
        "    if D_S == 1:\n",
        "        level = tf.keras.layers.Conv1D(1, 1, name=f'level4')(x4)\n",
        "        levels.append(level)\n",
        "    decode1 = decoder_unit(x4, num_filters * 8)\n",
        "    shape3 = x3.shape\n",
        "    x3 = Conv_1D_Block(x3, num_filters * 8, 1, shape3[1] + 1)\n",
        "    if A_G == 1:\n",
        "        x3 = Attention_Block(x3, decode1, num_filters * 8)\n",
        "    decode1 = tf.keras.layers.concatenate([decode1, x3], axis=-1)\n",
        "    if D_S == 1:\n",
        "        level = tf.keras.layers.Conv1D(1, 1, name=f'level3')(decode1)\n",
        "        levels.append(level)\n",
        "    decode2 = decoder_unit(decode1, num_filters * 4)\n",
        "    shape2 = x2.shape\n",
        "    x2 = Conv_1D_Block(x2, num_filters * 4, 1, shape2[1] + 1)\n",
        "    if A_G == 1:\n",
        "        x2 = Attention_Block(x2, decode2, num_filters * 4)\n",
        "    decode2 = tf.keras.layers.concatenate([decode2, x2], axis=-1)\n",
        "    if D_S == 1:\n",
        "        level = tf.keras.layers.Conv1D(1, 1, name=f'level2')(decode2)\n",
        "        levels.append(level)\n",
        "    decode3 = decoder_unit(decode2, num_filters * 2)\n",
        "    shape1 = x1.shape\n",
        "    x1 = Conv_1D_Block(x1, num_filters * 2, 1, shape1[1]+1)\n",
        "    if A_G == 1:\n",
        "        x1 = Attention_Block(x1, decode3, num_filters * 2)\n",
        "    decode3 = tf.keras.layers.concatenate([decode3, x1], axis=-1)\n",
        "    if D_S == 1:\n",
        "        level = tf.keras.layers.Conv1D(1, 1, name=f'level1')(decode3)\n",
        "        levels.append(level)\n",
        "    decode4 = decoder_unit(decode3, num_filters)\n",
        "    deconv5 = trans_conv1D(decode4, num_filters, 3, 2)\n",
        "    deconv6 = Conv_1D_Block(deconv5, num_filters, 3, 1)\n",
        "    out = Conv_1D_Block(deconv6, num_filters, 2, 1)\n",
        "    if D_S == 1:\n",
        "        level = tf.keras.layers.Conv1D(1, 1, name=f'level0')(decode4)\n",
        "        levels.append(level)\n",
        "\n",
        "    return out, levels\n",
        "\n",
        "\n",
        "class AlbUNet:\n",
        "    def __init__(self, length, num_channel, num_filters, ds=0, ae=0, ag=0, problem_type='Regression',\n",
        "                 output_nums=1, pooling='avg', feature_number=1024, dropout_rate=False):\n",
        "        # length: Input Signal Length\n",
        "        # model_depth: Depth of the Model\n",
        "        # model_width: Width of the Input Layer of the Model\n",
        "        # num_channel: Number of Channels allowed by the Model\n",
        "        # problem_type: Classification (Binary or Multiclass) or Regression\n",
        "        # output_nums: Output Classes (Classification Mode) or Features (Regression Mode)\n",
        "        # ds: Checks where Deep Supervision is active or not, either 0 or 1 [Default value set as 0]\n",
        "        # ae: Enables or diables the AutoEncoder Mode, either 0 or 1 [Default value set as 0]\n",
        "        # ag: Checks where Attention Guided is active or not, either 0 or 1 [Default value set as 0]\n",
        "        # alpha: This Parameter is only for MultiResUNet, default value is 1\n",
        "        # feature_number: Number of Features or Embeddings to be extracted from the AutoEncoder in the A_E Mode\n",
        "        self.length = length\n",
        "        self.num_channel = num_channel\n",
        "        self.num_filters = num_filters\n",
        "        self.problem_type = problem_type\n",
        "        self.output_nums = output_nums\n",
        "        self.D_S = ds\n",
        "        self.A_E = ae\n",
        "        self.A_G = ag\n",
        "        self.pooling = pooling\n",
        "        self.feature_number = feature_number\n",
        "        self.dropout_rate = dropout_rate\n",
        "\n",
        "    def MLP(self, x):\n",
        "        if self.dropout_rate:\n",
        "            x = tf.keras.layers.Dropout(self.dropout_rate, name='Dropout')(x)\n",
        "        outputs = tf.keras.layers.Dense(self.output_nums, activation='linear')(x)\n",
        "        if self.problem_type == 'Classification':\n",
        "            outputs = tf.keras.layers.Dense(self.output_nums, activation='sigmoid')(x)\n",
        "\n",
        "        return outputs\n",
        "\n",
        "    def AlbUNet18(self):\n",
        "        inputs = tf.keras.Input((self.length, self.num_channel))  # Input\n",
        "\n",
        "        # Encoder\n",
        "        stem_ = stem(inputs, self.num_filters)\n",
        "        x1, x2, x3, x4 = encoder18(stem_, self.num_filters)\n",
        "\n",
        "        # Feature Extraction\n",
        "        if self.A_E == 1:\n",
        "            x4 = Feature_Extraction_Block(x4, self.num_filters, self.feature_number)\n",
        "\n",
        "        # Decoder\n",
        "        decoder_output, levels = decoder_block(x1, x2, x3, x4, self.num_filters, self.D_S, self.A_G)\n",
        "        outputs = self.MLP(decoder_output)\n",
        "        decoder_output =GlobalAveragePooling1D()(decoder_output)\n",
        "        b = BatchNormalization()(decoder_output)\n",
        "        dense_1 = Dense(64, activation=activations.relu)(b)\n",
        "        dense_1 = Dense(32, activation=activations.relu)(dense_1)\n",
        "        dense_1 = Dense(1, activation=activations.sigmoid)(dense_1)\n",
        "        # Instantiate the Model\n",
        "        model = tf.keras.Model(inputs=[inputs], outputs=[dense_1])\n",
        "\n",
        "        if self.D_S == 1:\n",
        "            levels.append(outputs)\n",
        "            levels.reverse()\n",
        "            levels=GlobalAveragePooling1D()(levels)\n",
        "            model = tf.keras.Model(inputs=[inputs], outputs=levels)\n",
        "\n",
        "        return model\n",
        "\n",
        "    def AlbUNet34(self):\n",
        "        inputs = tf.keras.Input((self.length, self.num_channel))  # Input\n",
        "\n",
        "        # Encoder\n",
        "        stem_ = stem(inputs, self.num_filters)\n",
        "        x1, x2, x3, x4 = encoder34(stem_, self.num_filters)\n",
        "\n",
        "        # Feature Extraction\n",
        "        if self.A_E == 1:\n",
        "            x4 = Feature_Extraction_Block(x4, self.num_filters, self.feature_number)\n",
        "\n",
        "        # Decoder\n",
        "        decoder_output, levels = decoder_block(x1, x2, x3, x4, self.num_filters, self.D_S, self.A_G)\n",
        "        outputs = self.MLP(decoder_output)\n",
        "\n",
        "        # Instantiate the Model\n",
        "        model = tf.keras.Model(inputs=[inputs], outputs=[outputs])\n",
        "\n",
        "        if self.D_S == 1:\n",
        "            levels.append(outputs)\n",
        "            levels.reverse()\n",
        "            model = tf.keras.Model(inputs=[inputs], outputs=levels)\n",
        "\n",
        "        return model\n",
        "\n",
        "    def AlbUNet50(self):\n",
        "        inputs = tf.keras.Input((self.length, self.num_channel))  # Input\n",
        "\n",
        "        # Encoder\n",
        "        stem_ = stem(inputs, self.num_filters)\n",
        "        x1, x2, x3, x4 = encoder50(stem_, self.num_filters)\n",
        "\n",
        "        # Feature Extraction\n",
        "        if self.A_E == 1:\n",
        "            x4 = Feature_Extraction_Block(x4, self.num_filters, self.feature_number)\n",
        "\n",
        "        # Decoder\n",
        "        decoder_output, levels = decoder_block(x1, x2, x3, x4, self.num_filters, self.D_S, self.A_G)\n",
        "        outputs = self.MLP(decoder_output)\n",
        "\n",
        "        # Instantiate the Model\n",
        "        model = tf.keras.Model(inputs=[inputs], outputs=[outputs])\n",
        "\n",
        "        if self.D_S == 1:\n",
        "            levels.append(outputs)\n",
        "            levels.reverse()\n",
        "            model = tf.keras.Model(inputs=[inputs], outputs=levels)\n",
        "\n",
        "        return model\n",
        "\n",
        "    def AlbUNet101(self):\n",
        "        inputs = tf.keras.Input((self.length, self.num_channel))  # Input\n",
        "\n",
        "        # Encoder\n",
        "        stem_ = stem(inputs, self.num_filters)\n",
        "        x1, x2, x3, x4 = encoder101(stem_, self.num_filters)\n",
        "\n",
        "        # Feature Extraction\n",
        "        if self.A_E == 1:\n",
        "            x4 = Feature_Extraction_Block(x4, self.num_filters, self.feature_number)\n",
        "\n",
        "        # Decoder\n",
        "        decoder_output, levels = decoder_block(x1, x2, x3, x4, self.num_filters, self.D_S, self.A_G)\n",
        "        outputs = self.MLP(decoder_output)\n",
        "\n",
        "        # Instantiate the Model\n",
        "        model = tf.keras.Model(inputs=[inputs], outputs=[outputs])\n",
        "\n",
        "        if self.D_S == 1:\n",
        "            levels.append(outputs)\n",
        "            levels.reverse()\n",
        "            model = tf.keras.Model(inputs=[inputs], outputs=levels)\n",
        "\n",
        "        return model\n",
        "\n",
        "    def AlbUNet152(self):\n",
        "        inputs = tf.keras.Input((self.length, self.num_channel))  # Input\n",
        "\n",
        "        # Encoder\n",
        "        stem_ = stem(inputs, self.num_filters)\n",
        "        x1, x2, x3, x4 = encoder152(stem_, self.num_filters)\n",
        "\n",
        "        # Feature Extraction\n",
        "        if self.A_E == 1:\n",
        "            x4 = Feature_Extraction_Block(x4, self.num_filters, self.feature_number)\n",
        "\n",
        "        # Decoder\n",
        "        decoder_output, levels = decoder_block(x1, x2, x3, x4, self.num_filters, self.D_S, self.A_G)\n",
        "        decoder_output =GlobalAveragePooling1D()(decoder_output)\n",
        "        levels=GlobalAveragePooling1D()(levels)\n",
        "        outputs = self.MLP(decoder_output)\n",
        "\n",
        "        # Instantiate the Model\n",
        "        model = tf.keras.Model(inputs=[inputs], outputs=[outputs])\n",
        "\n",
        "        if self.D_S == 1:\n",
        "            levels.append(outputs)\n",
        "            levels.reverse()\n",
        "            model = tf.keras.Model(inputs=[inputs], outputs=levels)\n",
        "\n",
        "        return model\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "QSvIBB5M7fP2"
      },
      "outputs": [],
      "source": [
        " # Configurations\n",
        "if CFG.model ==3:\n",
        "    num_channel = 1\n",
        "    length = 874\n",
        "    model_width = 64\n",
        "    D_S = 0\n",
        "    A_E = 0\n",
        "    A_G = 1\n",
        "    feature_number = 874\n",
        "    model_name = 'AlbUNet18'\n",
        "    # Build model for AlbUNet\n",
        "    model = AlbUNet(length, num_channel, model_width, ds=D_S, ae=A_E, ag=A_G, problem_type='Classification', output_nums=1, pooling='avg', dropout_rate=False).AlbUNet18()\n",
        "    model.compile(optimizer=keras.optimizers.Nadam(learning_rate=0.001), loss=losses.binary_crossentropy, metrics=['acc'])\n",
        "    model.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "daA2zGJ87fP2",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6e45ee99-55e8-44c7-ba84-4cb549a74090"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_1 (InputLayer)        [(None, 874, 1)]          0         \n",
            "                                                                 \n",
            " conv1d (Conv1D)             (None, 874, 256)          1536      \n",
            "                                                                 \n",
            " max_pooling1d (MaxPooling1D  (None, 437, 256)         0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " batch_normalization (BatchN  (None, 437, 256)         1024      \n",
            " ormalization)                                                   \n",
            "                                                                 \n",
            " conv1d_1 (Conv1D)           (None, 437, 128)          163968    \n",
            "                                                                 \n",
            " max_pooling1d_1 (MaxPooling  (None, 217, 128)         0         \n",
            " 1D)                                                             \n",
            "                                                                 \n",
            " batch_normalization_1 (Batc  (None, 217, 128)         512       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " conv1d_2 (Conv1D)           (None, 217, 128)          82048     \n",
            "                                                                 \n",
            " max_pooling1d_2 (MaxPooling  (None, 109, 128)         0         \n",
            " 1D)                                                             \n",
            "                                                                 \n",
            " batch_normalization_2 (Batc  (None, 109, 128)         512       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " conv1d_3 (Conv1D)           (None, 109, 256)          164096    \n",
            "                                                                 \n",
            " global_max_pooling1d (Globa  (None, 256)              0         \n",
            " lMaxPooling1D)                                                  \n",
            "                                                                 \n",
            " batch_normalization_3 (Batc  (None, 256)              1024      \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " dense (Dense)               (None, 64)                16448     \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 32)                2080      \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 1)                 33        \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 433,281\n",
            "Trainable params: 431,745\n",
            "Non-trainable params: 1,536\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "if CFG.model ==1 or CFG.model ==2:\n",
        "    model = get_model()\n",
        "    model.summary()"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# model.compile(optimizer=keras.optimizers.Nadam(learning_rate=0.0001), loss=losses.binary_crossentropy, metrics=['acc'])"
      ],
      "metadata": {
        "id": "iHKPRUsm3Z6p"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Si0mZ_QW7fP2",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "dcbd98df-2c1d-4517-eee2-a2dc833ea21d"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "weights loaded\n"
          ]
        }
      ],
      "source": [
        "if CFG.load_weights and CFG.model==1:\n",
        "    \n",
        "    try:\n",
        "       model.load_weights(\"/content/gdrive/MyDrive/CS/third year/semester e/Voice Processing/Code/regular_model87.h5\")\n",
        "       #model.load_weights(\"/content/gdrive/MyDrive/CS/third year/vp/epochs/weights.03.h5\")\n",
        "       print(\"weights loaded\")\n",
        "    except ValueError:\n",
        "    # code to handle the exception\n",
        "      print(\"weights do not fit current model architecture\")\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# batch_size and epochs\n",
        "batch_size = 64\n",
        "epochs = 40\n",
        "patience = 20\n",
        "patience_for_plato = 20\n"
      ],
      "metadata": {
        "id": "ly_wqI7CDA-G"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from keras.callbacks import ModelCheckpoint\n",
        "\n",
        "checkpoint_filepath = '/content/gdrive/MyDrive/CS/third year/vp/epochs/weights.{epoch:02d}.h5'\n",
        "\n",
        "# Define the ModelCheckpoint callback to save the model weights after each epoch\n",
        "checkpoint_callback = ModelCheckpoint(\n",
        "    filepath=checkpoint_filepath,\n",
        "    save_weights_only=True,\n",
        "    save_best_only=False,\n",
        "    monitor='val_acc',\n",
        "    #mode='max', #for save_best_only=True\n",
        "    verbose=1\n",
        ")"
      ],
      "metadata": {
        "id": "rqMXlC7INlIb"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "M9OgRy617fP3",
        "outputId": "ec23b1cb-025b-4170-ce36-61e45f539a9b",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/40\n",
            "1266/1266 [==============================] - ETA: 0s - loss: 0.5104 - acc: 0.7984\n",
            "Epoch 1: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.01.h5\n",
            "1266/1266 [==============================] - 44s 22ms/step - loss: 0.5104 - acc: 0.7984 - val_loss: 0.4295 - val_acc: 0.8129 - lr: 0.0010\n",
            "Epoch 2/40\n",
            "1264/1266 [============================>.] - ETA: 0s - loss: 0.4004 - acc: 0.8199\n",
            "Epoch 2: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.02.h5\n",
            "1266/1266 [==============================] - 27s 21ms/step - loss: 0.4003 - acc: 0.8199 - val_loss: 0.4801 - val_acc: 0.7811 - lr: 0.0010\n",
            "Epoch 3/40\n",
            "1263/1266 [============================>.] - ETA: 0s - loss: 0.3785 - acc: 0.8302\n",
            "Epoch 3: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.03.h5\n",
            "1266/1266 [==============================] - 26s 21ms/step - loss: 0.3786 - acc: 0.8301 - val_loss: 0.7101 - val_acc: 0.6618 - lr: 0.0010\n",
            "Epoch 4/40\n",
            "1265/1266 [============================>.] - ETA: 0s - loss: 0.3597 - acc: 0.8416\n",
            "Epoch 4: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.04.h5\n",
            "1266/1266 [==============================] - 26s 21ms/step - loss: 0.3597 - acc: 0.8416 - val_loss: 0.4427 - val_acc: 0.8100 - lr: 0.0010\n",
            "Epoch 5/40\n",
            "1263/1266 [============================>.] - ETA: 0s - loss: 0.3424 - acc: 0.8502\n",
            "Epoch 5: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.05.h5\n",
            "1266/1266 [==============================] - 28s 22ms/step - loss: 0.3427 - acc: 0.8501 - val_loss: 0.4811 - val_acc: 0.7709 - lr: 0.0010\n",
            "Epoch 6/40\n",
            "1265/1266 [============================>.] - ETA: 0s - loss: 0.3257 - acc: 0.8563\n",
            "Epoch 6: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.06.h5\n",
            "1266/1266 [==============================] - 27s 21ms/step - loss: 0.3256 - acc: 0.8564 - val_loss: 0.4896 - val_acc: 0.7771 - lr: 0.0010\n",
            "Epoch 7/40\n",
            "1265/1266 [============================>.] - ETA: 0s - loss: 0.3024 - acc: 0.8673\n",
            "Epoch 7: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.07.h5\n",
            "1266/1266 [==============================] - 25s 20ms/step - loss: 0.3024 - acc: 0.8673 - val_loss: 1.5551 - val_acc: 0.5273 - lr: 0.0010\n",
            "Epoch 8/40\n",
            "1264/1266 [============================>.] - ETA: 0s - loss: 0.2891 - acc: 0.8738\n",
            "Epoch 8: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.08.h5\n",
            "1266/1266 [==============================] - 25s 20ms/step - loss: 0.2891 - acc: 0.8738 - val_loss: 0.7017 - val_acc: 0.7602 - lr: 0.0010\n",
            "Epoch 9/40\n",
            "1266/1266 [==============================] - ETA: 0s - loss: 0.2722 - acc: 0.8835\n",
            "Epoch 9: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.09.h5\n",
            "1266/1266 [==============================] - 26s 20ms/step - loss: 0.2722 - acc: 0.8835 - val_loss: 1.1716 - val_acc: 0.6240 - lr: 0.0010\n",
            "Epoch 10/40\n",
            "1264/1266 [============================>.] - ETA: 0s - loss: 0.2538 - acc: 0.8919\n",
            "Epoch 10: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.10.h5\n",
            "1266/1266 [==============================] - 25s 20ms/step - loss: 0.2537 - acc: 0.8919 - val_loss: 0.4145 - val_acc: 0.8218 - lr: 0.0010\n",
            "Epoch 11/40\n",
            "1264/1266 [============================>.] - ETA: 0s - loss: 0.2394 - acc: 0.8987\n",
            "Epoch 11: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.11.h5\n",
            "1266/1266 [==============================] - 25s 20ms/step - loss: 0.2395 - acc: 0.8985 - val_loss: 0.6434 - val_acc: 0.7827 - lr: 0.0010\n",
            "Epoch 12/40\n",
            "1263/1266 [============================>.] - ETA: 0s - loss: 0.2256 - acc: 0.9060\n",
            "Epoch 12: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.12.h5\n",
            "1266/1266 [==============================] - 26s 20ms/step - loss: 0.2254 - acc: 0.9060 - val_loss: 0.4591 - val_acc: 0.8051 - lr: 0.0010\n",
            "Epoch 13/40\n",
            "1264/1266 [============================>.] - ETA: 0s - loss: 0.2163 - acc: 0.9091\n",
            "Epoch 13: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.13.h5\n",
            "1266/1266 [==============================] - 26s 20ms/step - loss: 0.2163 - acc: 0.9091 - val_loss: 0.5088 - val_acc: 0.7816 - lr: 0.0010\n",
            "Epoch 14/40\n",
            "1266/1266 [==============================] - ETA: 0s - loss: 0.2021 - acc: 0.9171\n",
            "Epoch 14: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.14.h5\n",
            "1266/1266 [==============================] - 26s 20ms/step - loss: 0.2021 - acc: 0.9171 - val_loss: 0.5473 - val_acc: 0.7922 - lr: 0.0010\n",
            "Epoch 15/40\n",
            "1265/1266 [============================>.] - ETA: 0s - loss: 0.1888 - acc: 0.9208\n",
            "Epoch 15: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.15.h5\n",
            "1266/1266 [==============================] - 26s 20ms/step - loss: 0.1889 - acc: 0.9208 - val_loss: 0.4419 - val_acc: 0.8304 - lr: 0.0010\n",
            "Epoch 16/40\n",
            "1266/1266 [==============================] - ETA: 0s - loss: 0.1757 - acc: 0.9274\n",
            "Epoch 16: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.16.h5\n",
            "1266/1266 [==============================] - 26s 20ms/step - loss: 0.1757 - acc: 0.9274 - val_loss: 0.5205 - val_acc: 0.8020 - lr: 0.0010\n",
            "Epoch 17/40\n",
            "1264/1266 [============================>.] - ETA: 0s - loss: 0.1663 - acc: 0.9316\n",
            "Epoch 17: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.17.h5\n",
            "1266/1266 [==============================] - 25s 20ms/step - loss: 0.1663 - acc: 0.9316 - val_loss: 0.5226 - val_acc: 0.8244 - lr: 0.0010\n",
            "Epoch 18/40\n",
            "1265/1266 [============================>.] - ETA: 0s - loss: 0.1563 - acc: 0.9357\n",
            "Epoch 18: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.18.h5\n",
            "1266/1266 [==============================] - 26s 20ms/step - loss: 0.1562 - acc: 0.9358 - val_loss: 0.6210 - val_acc: 0.7891 - lr: 0.0010\n",
            "Epoch 19/40\n",
            "1264/1266 [============================>.] - ETA: 0s - loss: 0.1491 - acc: 0.9396\n",
            "Epoch 19: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.19.h5\n",
            "1266/1266 [==============================] - 25s 20ms/step - loss: 0.1491 - acc: 0.9395 - val_loss: 0.6880 - val_acc: 0.7784 - lr: 0.0010\n",
            "Epoch 20/40\n",
            "1263/1266 [============================>.] - ETA: 0s - loss: 0.1430 - acc: 0.9421\n",
            "Epoch 20: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.20.h5\n",
            "1266/1266 [==============================] - 25s 20ms/step - loss: 0.1431 - acc: 0.9420 - val_loss: 0.5636 - val_acc: 0.8298 - lr: 0.0010\n",
            "Epoch 21/40\n",
            "1264/1266 [============================>.] - ETA: 0s - loss: 0.1330 - acc: 0.9469\n",
            "Epoch 21: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.21.h5\n",
            "1266/1266 [==============================] - 26s 21ms/step - loss: 0.1329 - acc: 0.9469 - val_loss: 0.6310 - val_acc: 0.8098 - lr: 0.0010\n",
            "Epoch 22/40\n",
            "1265/1266 [============================>.] - ETA: 0s - loss: 0.1273 - acc: 0.9487\n",
            "Epoch 22: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.22.h5\n",
            "1266/1266 [==============================] - 26s 20ms/step - loss: 0.1274 - acc: 0.9487 - val_loss: 0.7691 - val_acc: 0.7991 - lr: 0.0010\n",
            "Epoch 23/40\n",
            "1263/1266 [============================>.] - ETA: 0s - loss: 0.1260 - acc: 0.9490\n",
            "Epoch 23: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.23.h5\n",
            "1266/1266 [==============================] - 25s 20ms/step - loss: 0.1261 - acc: 0.9489 - val_loss: 1.0686 - val_acc: 0.7698 - lr: 0.0010\n",
            "Epoch 24/40\n",
            "1264/1266 [============================>.] - ETA: 0s - loss: 0.1198 - acc: 0.9534\n",
            "Epoch 24: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.24.h5\n",
            "1266/1266 [==============================] - 25s 20ms/step - loss: 0.1198 - acc: 0.9534 - val_loss: 1.2150 - val_acc: 0.6524 - lr: 0.0010\n",
            "Epoch 25/40\n",
            "1264/1266 [============================>.] - ETA: 0s - loss: 0.1146 - acc: 0.9537\n",
            "Epoch 25: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.25.h5\n",
            "1266/1266 [==============================] - 26s 20ms/step - loss: 0.1145 - acc: 0.9538 - val_loss: 0.6348 - val_acc: 0.8282 - lr: 0.0010\n",
            "Epoch 26/40\n",
            "1264/1266 [============================>.] - ETA: 0s - loss: 0.1089 - acc: 0.9578\n",
            "Epoch 26: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.26.h5\n",
            "1266/1266 [==============================] - 26s 20ms/step - loss: 0.1089 - acc: 0.9578 - val_loss: 0.4981 - val_acc: 0.8373 - lr: 0.0010\n",
            "Epoch 27/40\n",
            "1265/1266 [============================>.] - ETA: 0s - loss: 0.1029 - acc: 0.9608\n",
            "Epoch 27: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.27.h5\n",
            "1266/1266 [==============================] - 26s 20ms/step - loss: 0.1029 - acc: 0.9608 - val_loss: 0.8517 - val_acc: 0.7673 - lr: 0.0010\n",
            "Epoch 28/40\n",
            "1263/1266 [============================>.] - ETA: 0s - loss: 0.0987 - acc: 0.9603\n",
            "Epoch 28: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.28.h5\n",
            "1266/1266 [==============================] - 26s 20ms/step - loss: 0.0985 - acc: 0.9604 - val_loss: 0.6914 - val_acc: 0.8231 - lr: 0.0010\n",
            "Epoch 29/40\n",
            "1263/1266 [============================>.] - ETA: 0s - loss: 0.0968 - acc: 0.9610\n",
            "Epoch 29: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.29.h5\n",
            "1266/1266 [==============================] - 26s 20ms/step - loss: 0.0968 - acc: 0.9611 - val_loss: 0.5039 - val_acc: 0.8473 - lr: 0.0010\n",
            "Epoch 30/40\n",
            "1264/1266 [============================>.] - ETA: 0s - loss: 0.0918 - acc: 0.9642\n",
            "Epoch 30: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.30.h5\n",
            "1266/1266 [==============================] - 26s 20ms/step - loss: 0.0919 - acc: 0.9641 - val_loss: 0.6366 - val_acc: 0.8269 - lr: 0.0010\n",
            "Epoch 31/40\n",
            "1264/1266 [============================>.] - ETA: 0s - loss: 0.0842 - acc: 0.9669\n",
            "Epoch 31: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.31.h5\n",
            "1266/1266 [==============================] - 25s 20ms/step - loss: 0.0841 - acc: 0.9669 - val_loss: 0.6687 - val_acc: 0.8269 - lr: 0.0010\n",
            "Epoch 32/40\n",
            "1264/1266 [============================>.] - ETA: 0s - loss: 0.0854 - acc: 0.9670\n",
            "Epoch 32: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.32.h5\n",
            "1266/1266 [==============================] - 25s 20ms/step - loss: 0.0853 - acc: 0.9670 - val_loss: 0.6020 - val_acc: 0.8458 - lr: 0.0010\n",
            "Epoch 33/40\n",
            "1264/1266 [============================>.] - ETA: 0s - loss: 0.0879 - acc: 0.9663\n",
            "Epoch 33: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.33.h5\n",
            "1266/1266 [==============================] - 26s 20ms/step - loss: 0.0879 - acc: 0.9663 - val_loss: 0.5939 - val_acc: 0.8307 - lr: 0.0010\n",
            "Epoch 34/40\n",
            "1264/1266 [============================>.] - ETA: 0s - loss: 0.0821 - acc: 0.9690\n",
            "Epoch 34: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.34.h5\n",
            "1266/1266 [==============================] - 26s 20ms/step - loss: 0.0821 - acc: 0.9690 - val_loss: 0.6343 - val_acc: 0.8176 - lr: 0.0010\n",
            "Epoch 35/40\n",
            "1264/1266 [============================>.] - ETA: 0s - loss: 0.0768 - acc: 0.9713\n",
            "Epoch 35: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.35.h5\n",
            "1266/1266 [==============================] - 25s 20ms/step - loss: 0.0769 - acc: 0.9712 - val_loss: 0.6253 - val_acc: 0.8398 - lr: 0.0010\n",
            "Epoch 36/40\n",
            "1263/1266 [============================>.] - ETA: 0s - loss: 0.0748 - acc: 0.9711\n",
            "Epoch 36: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.36.h5\n",
            "1266/1266 [==============================] - 26s 20ms/step - loss: 0.0749 - acc: 0.9711 - val_loss: 0.9293 - val_acc: 0.7618 - lr: 0.0010\n",
            "Epoch 37/40\n",
            "1265/1266 [============================>.] - ETA: 0s - loss: 0.0743 - acc: 0.9723\n",
            "Epoch 37: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.37.h5\n",
            "1266/1266 [==============================] - 25s 20ms/step - loss: 0.0743 - acc: 0.9723 - val_loss: 1.7604 - val_acc: 0.7098 - lr: 0.0010\n",
            "Epoch 38/40\n",
            "1265/1266 [============================>.] - ETA: 0s - loss: 0.0692 - acc: 0.9735\n",
            "Epoch 38: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.38.h5\n",
            "1266/1266 [==============================] - 26s 20ms/step - loss: 0.0692 - acc: 0.9735 - val_loss: 1.4683 - val_acc: 0.7571 - lr: 0.0010\n",
            "Epoch 39/40\n",
            "1263/1266 [============================>.] - ETA: 0s - loss: 0.0678 - acc: 0.9744\n",
            "Epoch 39: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.39.h5\n",
            "1266/1266 [==============================] - 25s 20ms/step - loss: 0.0677 - acc: 0.9744 - val_loss: 0.6483 - val_acc: 0.8242 - lr: 0.0010\n",
            "Epoch 40/40\n",
            "1264/1266 [============================>.] - ETA: 0s - loss: 0.0672 - acc: 0.9741\n",
            "Epoch 40: saving model to /content/gdrive/MyDrive/CS/third year/vp/epochs/weights.40.h5\n",
            "1266/1266 [==============================] - 25s 20ms/step - loss: 0.0671 - acc: 0.9741 - val_loss: 0.6068 - val_acc: 0.8531 - lr: 0.0010\n"
          ]
        }
      ],
      "source": [
        "rlrp = ReduceLROnPlateau(monitor='val_acc', factor=0.4, verbose=0, patience=patience_for_plato, min_lr=0.0000001)\n",
        "\n",
        "history=model.fit(x_train, y_train, batch_size=batch_size, epochs=epochs, validation_data=(x_val, y_val),\n",
        " callbacks=[tf.keras.callbacks.EarlyStopping(monitor='val_acc' ,patience=patience,restore_best_weights=True),rlrp,checkpoint_callback])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Wu0EM6EG7fP3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 404
        },
        "outputId": "7fb9d1b7-9926-4d7c-db61-82b6034fbe42"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1440x432 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "e=history.epoch\n",
        "epochs = [i for i in e]\n",
        "fig , ax = plt.subplots(1,2)\n",
        "train_acc = history.history['acc']\n",
        "train_loss = history.history['loss']\n",
        "test_acc = history.history['val_acc']\n",
        "test_loss = history.history['val_loss']\n",
        "\n",
        "fig.set_size_inches(20,6)\n",
        "ax[0].plot(epochs , train_loss , label = 'Training Loss')\n",
        "ax[0].plot(epochs , test_loss , label = 'Testing Loss')\n",
        "ax[0].set_title('Training & Testing Loss')\n",
        "ax[0].legend()\n",
        "ax[0].set_xlabel(\"Epochs\")\n",
        "\n",
        "ax[1].plot(epochs , train_acc , label = 'Training Accuracy')\n",
        "ax[1].plot(epochs , test_acc , label = 'Testing Accuracy')\n",
        "ax[1].set_title('Training & Testing Accuracy')\n",
        "ax[1].legend()\n",
        "ax[1].set_xlabel(\"Epochs\")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "##choosing the weights from the epoch with the best val_acc before the model started overfitting\n",
        "model.load_weights(\"/content/gdrive/MyDrive/CS/third year/vp/epochs/weights.29.h5\")"
      ],
      "metadata": {
        "id": "jO8PzCDAXLSx"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Uu3Ow8Oq7fP3",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e5864c9e-3937-4a6d-af3a-5f27d7a93bf8"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1266/1266 [==============================] - 9s 7ms/step - loss: 0.0653 - acc: 0.9761\n",
            "Accuracy of our model on training data :  97.60987758636475 %\n",
            "141/141 [==============================] - 1s 7ms/step - loss: 0.5039 - acc: 0.8473\n",
            "Accuracy of our model on validation data :  84.7333312034607 %\n"
          ]
        }
      ],
      "source": [
        "print(\"Accuracy of our model on training data : \" , model.evaluate(x_train,y_train)[1]*100 , \"%\")\n",
        "print(\"Accuracy of our model on validation data : \" , model.evaluate(x_val,y_val)[1]*100 , \"%\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "SWklCBeg7fP3"
      },
      "outputs": [],
      "source": [
        "if CFG.save_weights:\n",
        "    model.save('regular_model87.h5')\n",
        "    print(\"weights saved\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "gNz8FLHo7fP4",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "38e6029f-56e4-48b6-a1f5-23edc187b02b"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "141/141 [==============================] - 1s 6ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "4500"
            ]
          },
          "metadata": {},
          "execution_count": 84
        }
      ],
      "source": [
        "# predicting on val data.\n",
        "y_pred = model.predict(x_val)\n",
        "len(y_pred)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "v1HLzi-C7fP4"
      },
      "outputs": [],
      "source": [
        "for i in range(len(y_pred)):\n",
        "    y_pred[i][0] =round(y_pred[i][0])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Iqhy91F77fP4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 362
        },
        "outputId": "25c4491f-71ee-4ab8-feb1-211b5a3f51d5"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   Predicted Labels  Actual Labels\n",
              "0               0.0          False\n",
              "1               1.0           True\n",
              "2               0.0          False\n",
              "3               1.0           True\n",
              "4               0.0          False\n",
              "5               1.0           True\n",
              "6               0.0          False\n",
              "7               0.0          False\n",
              "8               0.0          False\n",
              "9               1.0           True"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-60044318-0532-49b7-b38e-15f4dbc9659a\">\n",
              "    <div 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>Predicted Labels</th>\n",
              "      <th>Actual Labels</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0.0</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1.0</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0.0</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1.0</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0.0</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>1.0</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>0.0</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>0.0</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>0.0</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>1.0</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-60044318-0532-49b7-b38e-15f4dbc9659a')\"\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 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\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",
              "    [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-60044318-0532-49b7-b38e-15f4dbc9659a 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-60044318-0532-49b7-b38e-15f4dbc9659a');\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",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 86
        }
      ],
      "source": [
        "df = pd.DataFrame(columns=['Predicted Labels', 'Actual Labels'])\n",
        "df['Predicted Labels'] = y_pred.flatten()\n",
        "df['Actual Labels'] = y_val.flatten()\n",
        "\n",
        "df.head(10)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "fLSRpOus7fP5"
      },
      "outputs": [],
      "source": [
        "import seaborn as sns\n",
        "from sklearn.metrics import confusion_matrix, classification_report\n",
        "from sklearn.metrics import roc_curve\n",
        "from sklearn.metrics import roc_auc_score\n",
        "from matplotlib import pyplot"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "vq8dq_kJ7fP5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 359
        },
        "outputId": "ae6ec362-bd1c-44c4-b8a7-2064161ea74d"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x360 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "cm = confusion_matrix(y_val, y_pred)\n",
        "plt.figure(figsize = (6, 5))\n",
        "sns.heatmap(cm, linecolor='white', cmap='Blues', linewidth=1, annot=True, fmt='')\n",
        "plt.title('Confusion Matrix', size=20)\n",
        "plt.xlabel('Predicted Labels', size=14)\n",
        "plt.ylabel('Actual Labels', size=14)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "j-M73rme7fP5",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "dd6a4744-2a45-46e0-e9da-48bb918f82e5"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "141/141 [==============================] - 1s 6ms/step\n",
            "141/141 [==============================] - 1s 6ms/step\n",
            "141/141 [==============================] - 1s 6ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.8788146057505732, 0.8682467758800976, 0.8473333333333334)"
            ]
          },
          "metadata": {},
          "execution_count": 89
        }
      ],
      "source": [
        "f1_score(y_val, np.round(model.predict(x_val))),recall_score(y_val, np.round(model.predict(x_val))),accuracy_score(y_val, np.round(model.predict(x_val)))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "cfVZ8hcC7fP5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "e242b883-bb90-4fd3-9acf-114f3822a2d5"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 90
        }
      ],
      "source": [
        "tf.keras.utils.plot_model(model, show_shapes=True, show_dtype=True, show_layer_names=True, expand_nested=False)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "NbIO-6M_7fP6"
      },
      "outputs": [],
      "source": [
        "if CFG.feature_extract_test:\n",
        "    X, Y = [], []\n",
        "    for i,file in enumerate(os.listdir(files_dir)[5000:5500]):\n",
        "        feature = get_features(os.path.join(files_dir,file))\n",
        "        X.append(feature)\n",
        "            # appending emotion 3 times as we have made 3 augmentation techniques on each audio file.\n",
        "        Y.append(df[\"professional-diagnosis\"].loc[df[\"file_name\"]==file].values[0])\n",
        "        if(i%100==0):\n",
        "            print(i)\n",
        "    test_features = pd.DataFrame(X)\n",
        "    test_features['labels'] = Y\n",
        "    test_features.to_csv('test_features.csv', index=False)\n",
        "    test_features.head()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "O2VMy9Wf7fP6"
      },
      "outputs": [],
      "source": [
        "test_features = pd.read_csv(\"/content/gdrive/MyDrive/CS/third year/semester e/Voice Processing/Data/test_features.csv\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "J1DG3kLx7fP6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "outputId": "535d2bd3-4847-4ce1-e744-d17793a79f92"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "          0         1         2         3         4         5         6  \\\n",
              "0  0.083333  0.196593  0.232184  0.410434  0.959038  0.459528  0.370831   \n",
              "1  0.099846  0.574448  0.640963  0.567182  0.617855  0.461433  0.420900   \n",
              "2  0.074958  0.480079  0.331349  0.359254  0.241088  0.491897  0.350522   \n",
              "3  0.064965  0.999996  0.349046  0.084231  0.108177  0.221526  0.065852   \n",
              "4  0.048586  0.099845  0.143571  0.472318  0.989016  0.396835  0.114040   \n",
              "\n",
              "          7         8         9  ...          865          866          867  \\\n",
              "0  0.447733  0.559178  0.223871  ...  1295.430370  1188.427835  1131.367443   \n",
              "1  0.507860  0.720867  0.768377  ...  1008.941380   985.372975  1011.440082   \n",
              "2  0.515516  0.456438  0.636179  ...   916.940687   899.465795   898.200540   \n",
              "3  0.160567  0.561014  0.178519  ...   787.001283   766.899693   767.517604   \n",
              "4  0.138395  0.237208  0.092520  ...  1017.574598  1044.652662  1023.418930   \n",
              "\n",
              "           868          869          870          871          872  \\\n",
              "0  1212.189642  1357.534814  1440.050351  1394.131030  1295.153809   \n",
              "1   981.509127   989.256712  1014.821007  1048.221391  1075.135023   \n",
              "2   899.802221   910.635415   889.459211   910.151568   924.586913   \n",
              "3   742.531608   730.434145   738.419267   734.284443   768.021960   \n",
              "4  1028.292075  1154.380415  1162.902760  1190.370686  1212.655064   \n",
              "\n",
              "           873  labels  \n",
              "0  1173.510194    True  \n",
              "1  1020.765866   False  \n",
              "2   900.287018    True  \n",
              "3   769.633828    True  \n",
              "4  1097.642249    True  \n",
              "\n",
              "[5 rows x 875 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-082fb1d7-b209-4a35-ada9-fc875fa9d356\">\n",
              "    <div 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>0</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "      <th>5</th>\n",
              "      <th>6</th>\n",
              "      <th>7</th>\n",
              "      <th>8</th>\n",
              "      <th>9</th>\n",
              "      <th>...</th>\n",
              "      <th>865</th>\n",
              "      <th>866</th>\n",
              "      <th>867</th>\n",
              "      <th>868</th>\n",
              "      <th>869</th>\n",
              "      <th>870</th>\n",
              "      <th>871</th>\n",
              "      <th>872</th>\n",
              "      <th>873</th>\n",
              "      <th>labels</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0.083333</td>\n",
              "      <td>0.196593</td>\n",
              "      <td>0.232184</td>\n",
              "      <td>0.410434</td>\n",
              "      <td>0.959038</td>\n",
              "      <td>0.459528</td>\n",
              "      <td>0.370831</td>\n",
              "      <td>0.447733</td>\n",
              "      <td>0.559178</td>\n",
              "      <td>0.223871</td>\n",
              "      <td>...</td>\n",
              "      <td>1295.430370</td>\n",
              "      <td>1188.427835</td>\n",
              "      <td>1131.367443</td>\n",
              "      <td>1212.189642</td>\n",
              "      <td>1357.534814</td>\n",
              "      <td>1440.050351</td>\n",
              "      <td>1394.131030</td>\n",
              "      <td>1295.153809</td>\n",
              "      <td>1173.510194</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0.099846</td>\n",
              "      <td>0.574448</td>\n",
              "      <td>0.640963</td>\n",
              "      <td>0.567182</td>\n",
              "      <td>0.617855</td>\n",
              "      <td>0.461433</td>\n",
              "      <td>0.420900</td>\n",
              "      <td>0.507860</td>\n",
              "      <td>0.720867</td>\n",
              "      <td>0.768377</td>\n",
              "      <td>...</td>\n",
              "      <td>1008.941380</td>\n",
              "      <td>985.372975</td>\n",
              "      <td>1011.440082</td>\n",
              "      <td>981.509127</td>\n",
              "      <td>989.256712</td>\n",
              "      <td>1014.821007</td>\n",
              "      <td>1048.221391</td>\n",
              "      <td>1075.135023</td>\n",
              "      <td>1020.765866</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0.074958</td>\n",
              "      <td>0.480079</td>\n",
              "      <td>0.331349</td>\n",
              "      <td>0.359254</td>\n",
              "      <td>0.241088</td>\n",
              "      <td>0.491897</td>\n",
              "      <td>0.350522</td>\n",
              "      <td>0.515516</td>\n",
              "      <td>0.456438</td>\n",
              "      <td>0.636179</td>\n",
              "      <td>...</td>\n",
              "      <td>916.940687</td>\n",
              "      <td>899.465795</td>\n",
              "      <td>898.200540</td>\n",
              "      <td>899.802221</td>\n",
              "      <td>910.635415</td>\n",
              "      <td>889.459211</td>\n",
              "      <td>910.151568</td>\n",
              "      <td>924.586913</td>\n",
              "      <td>900.287018</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0.064965</td>\n",
              "      <td>0.999996</td>\n",
              "      <td>0.349046</td>\n",
              "      <td>0.084231</td>\n",
              "      <td>0.108177</td>\n",
              "      <td>0.221526</td>\n",
              "      <td>0.065852</td>\n",
              "      <td>0.160567</td>\n",
              "      <td>0.561014</td>\n",
              "      <td>0.178519</td>\n",
              "      <td>...</td>\n",
              "      <td>787.001283</td>\n",
              "      <td>766.899693</td>\n",
              "      <td>767.517604</td>\n",
              "      <td>742.531608</td>\n",
              "      <td>730.434145</td>\n",
              "      <td>738.419267</td>\n",
              "      <td>734.284443</td>\n",
              "      <td>768.021960</td>\n",
              "      <td>769.633828</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0.048586</td>\n",
              "      <td>0.099845</td>\n",
              "      <td>0.143571</td>\n",
              "      <td>0.472318</td>\n",
              "      <td>0.989016</td>\n",
              "      <td>0.396835</td>\n",
              "      <td>0.114040</td>\n",
              "      <td>0.138395</td>\n",
              "      <td>0.237208</td>\n",
              "      <td>0.092520</td>\n",
              "      <td>...</td>\n",
              "      <td>1017.574598</td>\n",
              "      <td>1044.652662</td>\n",
              "      <td>1023.418930</td>\n",
              "      <td>1028.292075</td>\n",
              "      <td>1154.380415</td>\n",
              "      <td>1162.902760</td>\n",
              "      <td>1190.370686</td>\n",
              "      <td>1212.655064</td>\n",
              "      <td>1097.642249</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5 rows × 875 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-082fb1d7-b209-4a35-ada9-fc875fa9d356')\"\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 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\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",
              "    [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-082fb1d7-b209-4a35-ada9-fc875fa9d356 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-082fb1d7-b209-4a35-ada9-fc875fa9d356');\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",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 93
        }
      ],
      "source": [
        "test_features.head(5)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Qjk_28hH7fP6"
      },
      "outputs": [],
      "source": [
        "x_test = test_features.iloc[: ,:-1].values\n",
        "y_test = test_features['labels'].values"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Fc6vmukQ7fP6",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "bc765034-ef95-468a-8251-eed0d7789e67"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(500, 874)"
            ]
          },
          "metadata": {},
          "execution_count": 95
        }
      ],
      "source": [
        "x_test = scaler.transform(x_test)\n",
        "x_test.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "y9slK_397fP7",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e8851b66-bbbb-4551-a8ce-45eabff91efd"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((500, 874, 1), (500,))"
            ]
          },
          "metadata": {},
          "execution_count": 96
        }
      ],
      "source": [
        "# making our data compatible to model.\n",
        "x_test = np.expand_dims(x_test, axis=2)\n",
        "x_test.shape, y_test.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "PX5R0Y-i7fP7",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "832b670e-9a2f-4743-d9d5-d61b653b5296"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "16/16 [==============================] - 0s 7ms/step - loss: 0.6372 - acc: 0.8300\n",
            "Accuracy of our model on test data :  82.99999833106995 %\n"
          ]
        }
      ],
      "source": [
        "print(\"Accuracy of our model on test data : \" , model.evaluate(x_test,y_test)[1]*100 , \"%\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "a7poobFi7fP7",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d85c04c5-d0e0-4c70-f36f-97a71b451ae4"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "16/16 [==============================] - 0s 5ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "500"
            ]
          },
          "metadata": {},
          "execution_count": 98
        }
      ],
      "source": [
        "# predicting on test data.\n",
        "y_pred_test = model.predict(x_test)\n",
        "len(y_pred_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "UGQkpJG07fP7"
      },
      "outputs": [],
      "source": [
        "for i in range(len(y_pred_test)):\n",
        "    y_pred_test[i][0] =round(y_pred_test[i][0])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "x5P-3Ulm7fP8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 362
        },
        "outputId": "aadddf25-412a-4539-ad44-0154eb835b66"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   Predicted Labels  Actual Labels\n",
              "0               1.0           True\n",
              "1               0.0          False\n",
              "2               1.0           True\n",
              "3               1.0           True\n",
              "4               1.0           True\n",
              "5               0.0          False\n",
              "6               0.0           True\n",
              "7               1.0          False\n",
              "8               0.0           True\n",
              "9               1.0           True"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-55d59cfd-ba31-4432-a5a3-b395adb0e873\">\n",
              "    <div 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>Predicted Labels</th>\n",
              "      <th>Actual Labels</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1.0</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0.0</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1.0</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1.0</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1.0</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>0.0</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>0.0</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>1.0</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>0.0</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>1.0</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-55d59cfd-ba31-4432-a5a3-b395adb0e873')\"\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 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\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",
              "    [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-55d59cfd-ba31-4432-a5a3-b395adb0e873 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-55d59cfd-ba31-4432-a5a3-b395adb0e873');\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",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 100
        }
      ],
      "source": [
        "df_test = pd.DataFrame(columns=['Predicted Labels', 'Actual Labels'])\n",
        "df_test['Predicted Labels'] = y_pred_test.flatten()\n",
        "df_test['Actual Labels'] = y_test.flatten()\n",
        "\n",
        "df_test.head(10)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "LxYErd4D7fP8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 359
        },
        "outputId": "43b3c005-0bb1-4634-8536-312399c80637"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x360 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "cm = confusion_matrix(y_test, y_pred_test)\n",
        "plt.figure(figsize = (6, 5))\n",
        "sns.heatmap(cm, linecolor='white', cmap='Blues', linewidth=1, annot=True, fmt='')\n",
        "plt.title('Confusion Matrix', size=20)\n",
        "plt.xlabel('Predicted Labels', size=14)\n",
        "plt.ylabel('Actual Labels', size=14)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "PK2O1mbg7fP8",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e36957e8-49c6-493d-fd29-09871001868f"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "16/16 [==============================] - 0s 5ms/step\n",
            "16/16 [==============================] - 0s 5ms/step\n",
            "16/16 [==============================] - 0s 5ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.8608837970540099, 0.842948717948718, 0.83)"
            ]
          },
          "metadata": {},
          "execution_count": 102
        }
      ],
      "source": [
        "f1_score(y_test, np.round(model.predict(x_test))),recall_score(y_test, np.round(model.predict(x_test))),accuracy_score(y_test, np.round(model.predict(x_test)))"
      ]
    }
  ],
  "metadata": {
    "kernelspec": {
      "display_name": "TensorFlow",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.9.13"
    },
    "orig_nbformat": 4,
    "vscode": {
      "interpreter": {
        "hash": "7842d1971238c40bc91f0e3c4b1f67715db27be3d2f267cdf8f729dc15e9bd8b"
      }
    },
    "colab": {
      "provenance": []
    },
    "accelerator": "GPU",
    "gpuClass": "standard"
  },
  "nbformat": 4,
  "nbformat_minor": 0
}