CNNsegmentation.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "T4",
      "authorship_tag": "ABX9TyPGSAHIWyYNtxfEHimj236c",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/heejjj/Sleeping_Project/blob/main/CNNsegmentation.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "QL5-fnQNO6wS"
      },
      "outputs": [],
      "source": []
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YzZmzyDiPIx0",
        "outputId": "edd5b9f9-c0d4-4bdf-ea08-74660eea6731"
      },
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install scipy\n",
        "!pip install tqdm"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4ERe9LO8RfV8",
        "outputId": "b7c5b66e-96dd-4f58-eae1-e8aa61404e23"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.10/dist-packages (1.10.1)\n",
            "Requirement already satisfied: numpy<1.27.0,>=1.19.5 in /usr/local/lib/python3.10/dist-packages (from scipy) (1.22.4)\n",
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Requirement already satisfied: tqdm in /usr/local/lib/python3.10/dist-packages (4.65.0)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# /content/drive/MyDrive/KDT_project/Hackathon/CNNsegmentation/data/img/weekday"
      ],
      "metadata": {
        "id": "275R1qqAPJqy"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "\n",
        "# 데이터 전처리를 위한 라이브러리\n",
        "import json\n",
        "from scipy.signal import butter, lfilter, medfilt\n",
        "\n",
        "# 모델 학습을 위한 라이브러리\n",
        "from tensorflow.keras.models import Model\n",
        "from tensorflow.keras.layers import Input, Conv1D, BatchNormalization, MaxPooling1D\n",
        "from tensorflow.keras.layers import Conv1DTranspose, concatenate, Activation, SpatialDropout1D\n",
        "\n",
        "# 성능평가 함수를 위한 라이브러리\n",
        "from tensorflow.keras import backend as K\n",
        "\n",
        "# 시각화를 위한 라이브러리\n",
        "from tqdm import tqdm\n",
        "import matplotlib.pyplot as plt"
      ],
      "metadata": {
        "id": "3yLbPEkDQ6ho"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# weekday\n",
        "- 데이터수 : 50개\n",
        "- 채널 수\n",
        "- 데이터 길이 : 7567\n",
        "- \n",
        "\n",
        "# weekend\n",
        "- 데이터수 : 17개\n",
        "- 채널 수\n",
        "- 데이터 길이 : 8247\n",
        "- \n"
      ],
      "metadata": {
        "id": "WIhlJXtpUPsT"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 변수 정의하기\n",
        "# variables\n",
        "FREQUENCY_OF_DATASET = 500   # 심박수 한 사이클. 1/Hz = 주기. 일단 해보고 바꾼다면 2000\n",
        "\n",
        "num_classes = 4 # Core, REM, Deep, Awake"
      ],
      "metadata": {
        "id": "EIKMs3MyQ_Ic"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 데이터 증강"
      ],
      "metadata": {
        "id": "4ij73XR6dk_F"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def augmentation(dataset, ratio = 0.8):\n",
        "  train_idx = int(50 * ratio)  # 80%를 트레인세트로 분류\n",
        "  idx = np.arange(50)        # 0 ~ 49\n",
        "  np.random.shuffle(idx)\n",
        "\n",
        "  x = dataset[:][idx]\n",
        "  y = dataset[:][idx]\n",
        "\n",
        "  # 2. Split into 8:2\n",
        "  tmp_train_x = x[:train_idx]\n",
        "  tmp_valid_x = x[train_idx:]\n",
        "  tmp_train_y = y[:train_idx]\n",
        "  tmp_valid_y = y[train_idx:]\n",
        "\n",
        "  # 3. Augmentation on Train and Valid dataset\n",
        "  # step 설명 : 250 (0.5초) 이동하며 6초 신호를 augmentation한다.\n",
        "  step = 250\n",
        "\n",
        "  N_train = 1500  # 여유있게 범위를 줌\n",
        "  N_valid = 400\n",
        "\n",
        "  train_x, train_y = np.zeros((N_train, 2000, 1), dtype=np.float32), np.zeros((N_train, 2000, 4), dtype=np.float32) # 4초\n",
        "  valid_x, valid_y = np.zeros((N_valid, 2000, 1), dtype=np.float32), np.zeros((N_valid, 2000, 4), dtype=np.float32) \n",
        "  # print(N_train, N_valid)\n",
        "  print(\"Start Augmentation\")\n",
        "\n",
        "  # 4.1. Train\n",
        "  t_cnt, v_cnt = 0, 0\n",
        "  for patient in range(train_idx):\n",
        "    tmp = tmp_train_y[patient]\n",
        "\n",
        "    # tmp = np.argmax(tmp[:, 1:], axis=-1) \n",
        "    tmp = np.nonzero(tmp)[0]            # 0이 아닌 값들의 index 들을 반환해 주는 함수\n",
        "    start_idx, end_idx = tmp[0], tmp[-1]\n",
        "    length = end_idx - start_idx + 1\n",
        "\n",
        "    tmp_x = tmp_train_x[patient, start_idx:end_idx + 1]\n",
        "    tmp_y = tmp_train_y[patient, start_idx:end_idx + 1]\n",
        "    for idx in range(0, length-2000, step):        # 원본 15초에서 6초만큼씩 잘라, step 만큼 이동하여 데이터 증강\n",
        "      train_x[t_cnt] = tmp_x[idx:idx + 2000, :]\n",
        "      train_y[t_cnt] = tmp_y[idx:idx + 2000, :]\n",
        "      t_cnt += 1\n",
        "  print(f\"Number of Train set : {t_cnt}\")\n",
        "\n",
        "  # # 4.2. Valid\n",
        "  for patient in range(50-train_idx):\n",
        "    tmp = tmp_valid_y[patient]\n",
        "    \n",
        "    tmp = np.nonzero(tmp)[0]  \n",
        "    start_idx, end_idx = tmp[0], tmp[-1]\n",
        "    length = end_idx - start_idx + 1\n",
        "\n",
        "    tmp_x = tmp_valid_x[patient, start_idx:end_idx + 1]\n",
        "    tmp_y = tmp_valid_y[patient, start_idx:end_idx + 1]\n",
        "    for idx in range(0, length-2000, step):\n",
        "      valid_x[v_cnt] = tmp_x[idx:idx + 2000, :]\n",
        "      valid_y[v_cnt] = tmp_y[idx:idx + 2000, :]\n",
        "      v_cnt += 1\n",
        "  print(f\"Number of Valid set : {v_cnt}\")\n",
        "\n",
        "  train_x, train_y = train_x[:t_cnt], train_y[:t_cnt]\n",
        "  valid_x, valid_y = valid_x[:v_cnt], valid_y[:v_cnt]\n",
        "\n",
        "  # 5. Shuffle and Save\n",
        "  print(\"\\nStart saving the dataset...\")\n",
        "  idx = np.arange(t_cnt)\n",
        "  np.random.shuffle(idx)\n",
        "  train_x = train_x[idx]\n",
        "  train_y = train_y[idx]\n",
        "\n",
        "  np.save(f\"train_x\", train_x)\n",
        "  np.save(f\"train_y\", train_y)\n",
        "\n",
        "  np.save(f\"valid_x\", valid_x)\n",
        "  np.save(f\"valid_y\", valid_y)\n",
        "  print(\"Complete saving\\n\")"
      ],
      "metadata": {
        "id": "EWRG5ScfQ6fe"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 모델 설계 : U-Net 1D"
      ],
      "metadata": {
        "id": "vZpAYugadMVl"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def UNet1D():\n",
        "    num_filter = 16\n",
        "\n",
        "    inputs = Input((2000, 1))  # length of 4 sec signal is 3,000\n",
        "\n",
        "    # Encoder\n",
        "    x = Conv1D(num_filter, 9, kernel_initializer='he_normal', padding='same')(inputs)\n",
        "    # kernel_initializer : kernel가중치 행렬의 초기화 프로그램입니다( 참조 keras.initializers). 기본값은 'glorot_uniform'입니다.\n",
        "    x = BatchNormalization()(x)\n",
        "    x = SpatialDropout1D(0.3)(x)\n",
        "    x = Activation('relu')(x)\n",
        "    x = Conv1D(num_filter, 9, kernel_initializer='he_normal', padding='same')(x)\n",
        "    x = BatchNormalization()(x)\n",
        "    x = SpatialDropout1D(0.3)(x)\n",
        "    o1 = Activation('relu')(x)\n",
        "    x = MaxPooling1D(2)(o1)\n",
        "\n",
        "    x = Conv1D(num_filter*2, 9, kernel_initializer='he_normal', padding='same')(x)\n",
        "    x = BatchNormalization()(x)\n",
        "    x = SpatialDropout1D(0.3)(x)\n",
        "    x = Activation('relu')(x)\n",
        "    x = Conv1D(num_filter*2, 9, kernel_initializer='he_normal', padding='same')(x)\n",
        "    x = BatchNormalization()(x)\n",
        "    x = SpatialDropout1D(0.3)(x)\n",
        "    o2 = Activation('relu')(x)\n",
        "    x = MaxPooling1D(2)(o2)\n",
        "\n",
        "    x = Conv1D(num_filter*4, 9, kernel_initializer='he_normal', padding='same')(x)\n",
        "    x = BatchNormalization()(x)\n",
        "    x = SpatialDropout1D(0.3)(x)\n",
        "    x = Activation('relu')(x)\n",
        "    x = Conv1D(num_filter*4, 9, kernel_initializer='he_normal', padding='same')(x)\n",
        "    x = BatchNormalization()(x)\n",
        "    x = SpatialDropout1D(0.3)(x)\n",
        "    o3 = Activation('relu')(x)\n",
        "    x = MaxPooling1D(2)(o3)\n",
        "\n",
        "    x = Conv1D(num_filter*8, 9, kernel_initializer='he_normal', padding='same')(x)\n",
        "    x = BatchNormalization()(x)\n",
        "    x = SpatialDropout1D(0.3)(x)\n",
        "    x = Activation('relu')(x)\n",
        "    x = Conv1D(num_filter*8, 9, kernel_initializer='he_normal', padding='same')(x)\n",
        "    x = BatchNormalization()(x)\n",
        "    x = SpatialDropout1D(0.3)(x)\n",
        "    o4 = Activation('relu')(x)\n",
        "    x = MaxPooling1D(2)(o4)\n",
        "\n",
        "    x = Conv1D(num_filter*16, 9, kernel_initializer='he_normal', padding='same')(x)\n",
        "    x = BatchNormalization()(x)\n",
        "    x = SpatialDropout1D(0.3)(x)\n",
        "    x = Activation('relu')(x)\n",
        "    x = Conv1D(num_filter*16, 9, kernel_initializer='he_normal', padding='same')(x)\n",
        "    x = BatchNormalization()(x)\n",
        "    x = SpatialDropout1D(0.3)(x)\n",
        "    x = Activation('relu')(x)  \n",
        "\n",
        "    # Decoder\n",
        "    x = Conv1DTranspose(num_filter*8, 8, strides=2, padding='same')(x)\n",
        "    print(x.shape, o4.shape)\n",
        "    x = concatenate([x, o4])\n",
        "    x = Conv1D(num_filter*8, 9, kernel_initializer='he_normal', padding='same')(x)\n",
        "    x = BatchNormalization()(x)\n",
        "    x = SpatialDropout1D(0.3)(x)\n",
        "    x = Activation('relu')(x)\n",
        "    x = Conv1D(num_filter*8, 9, kernel_initializer='he_normal', padding='same')(x)\n",
        "    x = BatchNormalization()(x)\n",
        "    x = SpatialDropout1D(0.3)(x)\n",
        "    x = Activation('relu')(x)\n",
        "\n",
        "    x = Conv1DTranspose(num_filter*4, 8, strides=2, padding='same')(x)\n",
        "    print(x.shape, o3.shape)\n",
        "    x = concatenate([x, o3])\n",
        "    x = Conv1D(num_filter*4, 9, kernel_initializer='he_normal', padding='same')(x)\n",
        "    x = BatchNormalization()(x)\n",
        "    x = SpatialDropout1D(0.3)(x)\n",
        "    x = Activation('relu')(x)\n",
        "    x = Conv1D(num_filter*4, 9, kernel_initializer='he_normal', padding='same')(x)\n",
        "    x = BatchNormalization()(x)\n",
        "    x = SpatialDropout1D(0.3)(x)\n",
        "    x = Activation('relu')(x)\n",
        "\n",
        "    x = Conv1DTranspose(num_filter*2, 8, strides=2, padding='same')(x)\n",
        "    print(x.shape, o2.shape)\n",
        "    x = concatenate([x, o2])\n",
        "    x = Conv1D(num_filter*2, 9, kernel_initializer='he_normal', padding='same')(x)\n",
        "    x = BatchNormalization()(x)\n",
        "    x = SpatialDropout1D(0.3)(x)\n",
        "    x = Activation('relu')(x)\n",
        "    x = Conv1D(num_filter*2, 9, kernel_initializer='he_normal', padding='same')(x)\n",
        "    x = BatchNormalization()(x)\n",
        "    x = SpatialDropout1D(0.3)(x)\n",
        "    x = Activation('relu')(x)\n",
        "\n",
        "    x = Conv1DTranspose(num_filter, 8, strides=2, padding='same')(x)\n",
        "    print(x.shape, o1.shape)\n",
        "    x = concatenate([x, o1])\n",
        "    x = Conv1D(num_filter, 9, kernel_initializer='he_normal', padding='same')(x)\n",
        "    x = BatchNormalization()(x)\n",
        "    x = SpatialDropout1D(0.3)(x)\n",
        "    x = Activation('relu')(x)\n",
        "    x = Conv1D(num_filter, 9, kernel_initializer='he_normal', padding='same')(x)\n",
        "    x = BatchNormalization()(x)\n",
        "    x = SpatialDropout1D(0.3)(x)\n",
        "    x = Activation('relu')(x)\n",
        "\n",
        "    outputs = Conv1D(num_classes, 1, activation='softmax', kernel_initializer='he_normal')(x)\n",
        "\n",
        "    model = Model(inputs=inputs, outputs=outputs)\n",
        "    return model"
      ],
      "metadata": {
        "id": "PEKM-9ZGQ6dC"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 성능평가함수 정의"
      ],
      "metadata": {
        "id": "mkhR7qmedRt-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 1. F1-Score = Dice score\n",
        "def f1_score(y_true, y_pred):\n",
        "    smooth = 0.0001\n",
        "    intersection = K.sum(y_true * y_pred, axis=[1, 2])\n",
        "    union = K.sum(y_true, axis=[1, 2]) + K.sum(y_pred, axis=[1, 2])\n",
        "    return K.mean( (2. * intersection + smooth) / (union + smooth), axis=0)\n",
        "\n",
        "class f1_loss(tf.losses.Loss):\n",
        "    def __init__(self):\n",
        "        super(f1_loss, self).__init__(reduction='auto', name=\"f1_loss\")\n",
        "\n",
        "    def call(self, y_true, y_pred):\n",
        "        return 1 - f1_score(y_true, y_pred)\n",
        "\n",
        "\n",
        "# 2. Jaccard = IoU\n",
        "def Jaccard(y_true, y_pred):\n",
        "    smooth = 0.0001\n",
        "    intersection = K.sum(y_true * y_pred, axis=[1, 2])\n",
        "    union = K.sum(y_true, axis=[1, 2]) + K.sum(y_pred, axis=[1, 2])\n",
        "    return K.mean((intersection + smooth) / (union - intersection + smooth), axis=0)\n",
        "\n",
        "class JaccardLoss(tf.losses.Loss):\n",
        "    def __init__(self):\n",
        "        super(JaccardLoss, self).__init__(reduction='auto', name=\"JaccardLoss\")\n",
        "\n",
        "    def call(self, y_true, y_pred):\n",
        "        return 1 - Jaccard(y_true, y_pred)"
      ],
      "metadata": {
        "id": "9K1Tj3XGQ6as"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 예측 함수 정의"
      ],
      "metadata": {
        "id": "lWK3jU3QdYnx"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def Predict(test, n_sample=5):\n",
        "    '''\n",
        "    Input\n",
        "        - test: [valid_x, valid_y]\n",
        "        - n_sample: 보여줄 샘플 갯수\n",
        "    테스트셋에 대해서 예측한 결과와 실제 labeling 데이터를 시각화한다.    \n",
        "    '''\n",
        "    model = UNet1D()\n",
        "    model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=0.005),\n",
        "                  loss=f1_loss(),\n",
        "                  metrics=[Jaccard, f1_score])\n",
        "    model.load_weights(\"weights.h5\")\n",
        "\n",
        "    for i, test_ecg in enumerate(test[0]): # valid_x\n",
        "        pred = model.predict(np.reshape(test_ecg, (1, 2000, 1)))\n",
        "        true = np.argmax(test[1][i], axis=-1)\n",
        "        pred = np.argmax(pred[0], axis=-1)\n",
        "\n",
        "        fig, ax = plt.subplots(3, 1, figsize=(20, 12))\n",
        "\n",
        "        ax[0].plot(test_ecg[:, 1])\n",
        "        ax[0].set_title(f\"ECG_{i}'s lead II\", fontsize=10)\n",
        "\n",
        "        ax[1].plot(true)\n",
        "        ax[1].set_title(f\"ECG_{i}'s Groud Truth\", fontsize=10)\n",
        "\n",
        "        ax[2].plot(pred)\n",
        "        ax[2].set_title(f\"ECG_{i}'s Prediction\", fontsize=10)\n",
        "\n",
        "        plt.show()\n",
        "        \n",
        "        # 20개의 샘플만 보여준 후 멈춤\n",
        "        if i == n_sample - 1:\n",
        "            break"
      ],
      "metadata": {
        "id": "e44gaO07dZl1"
      },
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "f1jLPv-rda5D"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import warnings\n",
        "warnings.filterwarnings(action='ignore')"
      ],
      "metadata": {
        "id": "GZiTYG2Tfwwj"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "\n",
        "weekday = pd.read_csv('/content/drive/MyDrive/KDT_project/Hackathon/CNNsegmentation/data/weekday.csv')\n",
        "weekend = pd.read_csv('/content/drive/MyDrive/KDT_project/Hackathon/CNNsegmentation/data/weekend.csv')\n",
        "weekday = weekday.drop('Unnamed: 0', axis=1)\n",
        "weekend = weekend.drop('Unnamed: 0', axis=1)\n",
        "print(weekday.shape, weekend.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZZIrIpZ7dTyH",
        "outputId": "d7837d64-ed46-4530-9d51-e1a828e50ba6"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(1434, 7) (575, 7)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "img_weekday = weekday.iloc[:,[2,3]]\n",
        "img_weekday[\"Duration (hr)\"] = img_weekday[\"Duration (hr)\"].astype('float') * 1000\n",
        "img_weekday['Value']= img_weekday['Value'].replace('Core',2)\n",
        "img_weekday['Value']= img_weekday['Value'].replace('REM',3)\n",
        "img_weekday['Value']= img_weekday['Value'].replace('Deep',1)\n",
        "img_weekday['Value']= img_weekday['Value'].replace('Awake',4)\n",
        "img_weekday = img_weekday[[\"Duration (hr)\", \"Value\"]]\n",
        "img_weekday = np.array(img_weekday)\n",
        "\n",
        "x_list = []   # 라벨을 duration 만큼 반복한 array를 list에 append\n",
        "for x in img_weekday:\n",
        "  if x[1] != 0:\n",
        "    timeseries = np.ones([int(x[0])]) * x[1]\n",
        "    x_list.append(timeseries)\n",
        "\n",
        "ts = np.concatenate(x_list)[:-3383]\n",
        "print(ts.shape)\n",
        "\n",
        "tmp = ts.reshape(50, 7500, -1)\n",
        "print(tmp.shape)\n",
        "\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pxEjQPGjeCJH",
        "outputId": "76ba7189-835f-48b3-9679-b3af0a9f4806"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(375000,)\n",
            "(50, 7500, 1)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Augmentation\n",
        "augmentation(tmp)\n",
        "del tmp"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KKYqGj0fHxZz",
        "outputId": "b1b493c8-b6c0-4185-c4fb-28699b26f642"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Start Augmentation\n",
            "Number of Train set : 880\n",
            "Number of Valid set : 220\n",
            "\n",
            "Start saving the dataset...\n",
            "Complete saving\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train = [np.load(\"train_x.npy\"), np.load(\"train_y.npy\")]\n",
        "valid = [np.load(\"valid_x.npy\"), np.load(\"valid_y.npy\")]"
      ],
      "metadata": {
        "id": "48oGbuV2Hrub"
      },
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 모델 학습"
      ],
      "metadata": {
        "id": "unx4Kj6cNvSE"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 4. Model\n",
        "print(\"Start Training\")\n",
        "batch_size = 256\n",
        "lr = 0.005\n",
        "epochs = 2000\n",
        "\n",
        "model = UNet1D()\n",
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "s9GpBefBV7h_",
        "outputId": "bca9cca5-65a2-4304-bdd4-5268cf83b514"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Start Training\n",
            "(None, 250, 128) (None, 250, 128)\n",
            "(None, 500, 64) (None, 500, 64)\n",
            "(None, 1000, 32) (None, 1000, 32)\n",
            "(None, 2000, 16) (None, 2000, 16)\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            " Layer (type)                   Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            " input_1 (InputLayer)           [(None, 2000, 1)]    0           []                               \n",
            "                                                                                                  \n",
            " conv1d (Conv1D)                (None, 2000, 16)     160         ['input_1[0][0]']                \n",
            "                                                                                                  \n",
            " batch_normalization (BatchNorm  (None, 2000, 16)    64          ['conv1d[0][0]']                 \n",
            " alization)                                                                                       \n",
            "                                                                                                  \n",
            " spatial_dropout1d (SpatialDrop  (None, 2000, 16)    0           ['batch_normalization[0][0]']    \n",
            " out1D)                                                                                           \n",
            "                                                                                                  \n",
            " activation (Activation)        (None, 2000, 16)     0           ['spatial_dropout1d[0][0]']      \n",
            "                                                                                                  \n",
            " conv1d_1 (Conv1D)              (None, 2000, 16)     2320        ['activation[0][0]']             \n",
            "                                                                                                  \n",
            " batch_normalization_1 (BatchNo  (None, 2000, 16)    64          ['conv1d_1[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " spatial_dropout1d_1 (SpatialDr  (None, 2000, 16)    0           ['batch_normalization_1[0][0]']  \n",
            " opout1D)                                                                                         \n",
            "                                                                                                  \n",
            " activation_1 (Activation)      (None, 2000, 16)     0           ['spatial_dropout1d_1[0][0]']    \n",
            "                                                                                                  \n",
            " max_pooling1d (MaxPooling1D)   (None, 1000, 16)     0           ['activation_1[0][0]']           \n",
            "                                                                                                  \n",
            " conv1d_2 (Conv1D)              (None, 1000, 32)     4640        ['max_pooling1d[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_2 (BatchNo  (None, 1000, 32)    128         ['conv1d_2[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " spatial_dropout1d_2 (SpatialDr  (None, 1000, 32)    0           ['batch_normalization_2[0][0]']  \n",
            " opout1D)                                                                                         \n",
            "                                                                                                  \n",
            " activation_2 (Activation)      (None, 1000, 32)     0           ['spatial_dropout1d_2[0][0]']    \n",
            "                                                                                                  \n",
            " conv1d_3 (Conv1D)              (None, 1000, 32)     9248        ['activation_2[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_3 (BatchNo  (None, 1000, 32)    128         ['conv1d_3[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " spatial_dropout1d_3 (SpatialDr  (None, 1000, 32)    0           ['batch_normalization_3[0][0]']  \n",
            " opout1D)                                                                                         \n",
            "                                                                                                  \n",
            " activation_3 (Activation)      (None, 1000, 32)     0           ['spatial_dropout1d_3[0][0]']    \n",
            "                                                                                                  \n",
            " max_pooling1d_1 (MaxPooling1D)  (None, 500, 32)     0           ['activation_3[0][0]']           \n",
            "                                                                                                  \n",
            " conv1d_4 (Conv1D)              (None, 500, 64)      18496       ['max_pooling1d_1[0][0]']        \n",
            "                                                                                                  \n",
            " batch_normalization_4 (BatchNo  (None, 500, 64)     256         ['conv1d_4[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " spatial_dropout1d_4 (SpatialDr  (None, 500, 64)     0           ['batch_normalization_4[0][0]']  \n",
            " opout1D)                                                                                         \n",
            "                                                                                                  \n",
            " activation_4 (Activation)      (None, 500, 64)      0           ['spatial_dropout1d_4[0][0]']    \n",
            "                                                                                                  \n",
            " conv1d_5 (Conv1D)              (None, 500, 64)      36928       ['activation_4[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_5 (BatchNo  (None, 500, 64)     256         ['conv1d_5[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " spatial_dropout1d_5 (SpatialDr  (None, 500, 64)     0           ['batch_normalization_5[0][0]']  \n",
            " opout1D)                                                                                         \n",
            "                                                                                                  \n",
            " activation_5 (Activation)      (None, 500, 64)      0           ['spatial_dropout1d_5[0][0]']    \n",
            "                                                                                                  \n",
            " max_pooling1d_2 (MaxPooling1D)  (None, 250, 64)     0           ['activation_5[0][0]']           \n",
            "                                                                                                  \n",
            " conv1d_6 (Conv1D)              (None, 250, 128)     73856       ['max_pooling1d_2[0][0]']        \n",
            "                                                                                                  \n",
            " batch_normalization_6 (BatchNo  (None, 250, 128)    512         ['conv1d_6[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " spatial_dropout1d_6 (SpatialDr  (None, 250, 128)    0           ['batch_normalization_6[0][0]']  \n",
            " opout1D)                                                                                         \n",
            "                                                                                                  \n",
            " activation_6 (Activation)      (None, 250, 128)     0           ['spatial_dropout1d_6[0][0]']    \n",
            "                                                                                                  \n",
            " conv1d_7 (Conv1D)              (None, 250, 128)     147584      ['activation_6[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_7 (BatchNo  (None, 250, 128)    512         ['conv1d_7[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " spatial_dropout1d_7 (SpatialDr  (None, 250, 128)    0           ['batch_normalization_7[0][0]']  \n",
            " opout1D)                                                                                         \n",
            "                                                                                                  \n",
            " activation_7 (Activation)      (None, 250, 128)     0           ['spatial_dropout1d_7[0][0]']    \n",
            "                                                                                                  \n",
            " max_pooling1d_3 (MaxPooling1D)  (None, 125, 128)    0           ['activation_7[0][0]']           \n",
            "                                                                                                  \n",
            " conv1d_8 (Conv1D)              (None, 125, 256)     295168      ['max_pooling1d_3[0][0]']        \n",
            "                                                                                                  \n",
            " batch_normalization_8 (BatchNo  (None, 125, 256)    1024        ['conv1d_8[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " spatial_dropout1d_8 (SpatialDr  (None, 125, 256)    0           ['batch_normalization_8[0][0]']  \n",
            " opout1D)                                                                                         \n",
            "                                                                                                  \n",
            " activation_8 (Activation)      (None, 125, 256)     0           ['spatial_dropout1d_8[0][0]']    \n",
            "                                                                                                  \n",
            " conv1d_9 (Conv1D)              (None, 125, 256)     590080      ['activation_8[0][0]']           \n",
            "                                                                                                  \n",
            " batch_normalization_9 (BatchNo  (None, 125, 256)    1024        ['conv1d_9[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " spatial_dropout1d_9 (SpatialDr  (None, 125, 256)    0           ['batch_normalization_9[0][0]']  \n",
            " opout1D)                                                                                         \n",
            "                                                                                                  \n",
            " activation_9 (Activation)      (None, 125, 256)     0           ['spatial_dropout1d_9[0][0]']    \n",
            "                                                                                                  \n",
            " conv1d_transpose (Conv1DTransp  (None, 250, 128)    262272      ['activation_9[0][0]']           \n",
            " ose)                                                                                             \n",
            "                                                                                                  \n",
            " concatenate (Concatenate)      (None, 250, 256)     0           ['conv1d_transpose[0][0]',       \n",
            "                                                                  'activation_7[0][0]']           \n",
            "                                                                                                  \n",
            " conv1d_10 (Conv1D)             (None, 250, 128)     295040      ['concatenate[0][0]']            \n",
            "                                                                                                  \n",
            " batch_normalization_10 (BatchN  (None, 250, 128)    512         ['conv1d_10[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " spatial_dropout1d_10 (SpatialD  (None, 250, 128)    0           ['batch_normalization_10[0][0]'] \n",
            " ropout1D)                                                                                        \n",
            "                                                                                                  \n",
            " activation_10 (Activation)     (None, 250, 128)     0           ['spatial_dropout1d_10[0][0]']   \n",
            "                                                                                                  \n",
            " conv1d_11 (Conv1D)             (None, 250, 128)     147584      ['activation_10[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_11 (BatchN  (None, 250, 128)    512         ['conv1d_11[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " spatial_dropout1d_11 (SpatialD  (None, 250, 128)    0           ['batch_normalization_11[0][0]'] \n",
            " ropout1D)                                                                                        \n",
            "                                                                                                  \n",
            " activation_11 (Activation)     (None, 250, 128)     0           ['spatial_dropout1d_11[0][0]']   \n",
            "                                                                                                  \n",
            " conv1d_transpose_1 (Conv1DTran  (None, 500, 64)     65600       ['activation_11[0][0]']          \n",
            " spose)                                                                                           \n",
            "                                                                                                  \n",
            " concatenate_1 (Concatenate)    (None, 500, 128)     0           ['conv1d_transpose_1[0][0]',     \n",
            "                                                                  'activation_5[0][0]']           \n",
            "                                                                                                  \n",
            " conv1d_12 (Conv1D)             (None, 500, 64)      73792       ['concatenate_1[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_12 (BatchN  (None, 500, 64)     256         ['conv1d_12[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " spatial_dropout1d_12 (SpatialD  (None, 500, 64)     0           ['batch_normalization_12[0][0]'] \n",
            " ropout1D)                                                                                        \n",
            "                                                                                                  \n",
            " activation_12 (Activation)     (None, 500, 64)      0           ['spatial_dropout1d_12[0][0]']   \n",
            "                                                                                                  \n",
            " conv1d_13 (Conv1D)             (None, 500, 64)      36928       ['activation_12[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_13 (BatchN  (None, 500, 64)     256         ['conv1d_13[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " spatial_dropout1d_13 (SpatialD  (None, 500, 64)     0           ['batch_normalization_13[0][0]'] \n",
            " ropout1D)                                                                                        \n",
            "                                                                                                  \n",
            " activation_13 (Activation)     (None, 500, 64)      0           ['spatial_dropout1d_13[0][0]']   \n",
            "                                                                                                  \n",
            " conv1d_transpose_2 (Conv1DTran  (None, 1000, 32)    16416       ['activation_13[0][0]']          \n",
            " spose)                                                                                           \n",
            "                                                                                                  \n",
            " concatenate_2 (Concatenate)    (None, 1000, 64)     0           ['conv1d_transpose_2[0][0]',     \n",
            "                                                                  'activation_3[0][0]']           \n",
            "                                                                                                  \n",
            " conv1d_14 (Conv1D)             (None, 1000, 32)     18464       ['concatenate_2[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_14 (BatchN  (None, 1000, 32)    128         ['conv1d_14[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " spatial_dropout1d_14 (SpatialD  (None, 1000, 32)    0           ['batch_normalization_14[0][0]'] \n",
            " ropout1D)                                                                                        \n",
            "                                                                                                  \n",
            " activation_14 (Activation)     (None, 1000, 32)     0           ['spatial_dropout1d_14[0][0]']   \n",
            "                                                                                                  \n",
            " conv1d_15 (Conv1D)             (None, 1000, 32)     9248        ['activation_14[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_15 (BatchN  (None, 1000, 32)    128         ['conv1d_15[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " spatial_dropout1d_15 (SpatialD  (None, 1000, 32)    0           ['batch_normalization_15[0][0]'] \n",
            " ropout1D)                                                                                        \n",
            "                                                                                                  \n",
            " activation_15 (Activation)     (None, 1000, 32)     0           ['spatial_dropout1d_15[0][0]']   \n",
            "                                                                                                  \n",
            " conv1d_transpose_3 (Conv1DTran  (None, 2000, 16)    4112        ['activation_15[0][0]']          \n",
            " spose)                                                                                           \n",
            "                                                                                                  \n",
            " concatenate_3 (Concatenate)    (None, 2000, 32)     0           ['conv1d_transpose_3[0][0]',     \n",
            "                                                                  'activation_1[0][0]']           \n",
            "                                                                                                  \n",
            " conv1d_16 (Conv1D)             (None, 2000, 16)     4624        ['concatenate_3[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_16 (BatchN  (None, 2000, 16)    64          ['conv1d_16[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " spatial_dropout1d_16 (SpatialD  (None, 2000, 16)    0           ['batch_normalization_16[0][0]'] \n",
            " ropout1D)                                                                                        \n",
            "                                                                                                  \n",
            " activation_16 (Activation)     (None, 2000, 16)     0           ['spatial_dropout1d_16[0][0]']   \n",
            "                                                                                                  \n",
            " conv1d_17 (Conv1D)             (None, 2000, 16)     2320        ['activation_16[0][0]']          \n",
            "                                                                                                  \n",
            " batch_normalization_17 (BatchN  (None, 2000, 16)    64          ['conv1d_17[0][0]']              \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " spatial_dropout1d_17 (SpatialD  (None, 2000, 16)    0           ['batch_normalization_17[0][0]'] \n",
            " ropout1D)                                                                                        \n",
            "                                                                                                  \n",
            " activation_17 (Activation)     (None, 2000, 16)     0           ['spatial_dropout1d_17[0][0]']   \n",
            "                                                                                                  \n",
            " conv1d_18 (Conv1D)             (None, 2000, 4)      68          ['activation_17[0][0]']          \n",
            "                                                                                                  \n",
            "==================================================================================================\n",
            "Total params: 2,120,836\n",
            "Trainable params: 2,117,892\n",
            "Non-trainable params: 2,944\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 5. Train\n",
        "model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=lr),\n",
        "              loss=f1_loss(),\n",
        "              metrics=[Jaccard, f1_score])\n",
        "\n",
        "callbacks_list = [\n",
        "    tf.keras.callbacks.ModelCheckpoint(\n",
        "        filepath=\"weights.h5\",\n",
        "        monitor=\"val_f1_score\",\n",
        "        mode=\"max\",\n",
        "        save_best_only=True,\n",
        "        save_weights_only=True,\n",
        "        verbose=1),\n",
        "\n",
        "    tf.keras.callbacks.EarlyStopping(\n",
        "        monitor=\"val_f1_score\",\n",
        "        min_delta=0.005,\n",
        "        mode=\"max\",\n",
        "        patience=10,\n",
        "        verbose=1),\n",
        "\n",
        "    tf.keras.callbacks.ReduceLROnPlateau(\n",
        "        monitor='val_f1_score',\n",
        "        factor=0.5,\n",
        "        patience=5,\n",
        "        mode='max',\n",
        "        min_delta=0.01,\n",
        "        cooldown=0,\n",
        "        min_lr=0.0001,\n",
        "        verbose=1)\n",
        "]\n",
        "\n",
        "history = model.fit(train[0], train[1],\n",
        "                    validation_data=(valid[0], valid[1]),\n",
        "                    callbacks=callbacks_list,\n",
        "                    epochs=epochs,\n",
        "                    batch_size=batch_size,\n",
        "                    shuffle=True,\n",
        "                    )\n",
        "\n",
        "\n",
        "loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']\n",
        "Jaccard = history.history[\"Jaccard\"]\n",
        "val_Jaccard = history.history[\"val_Jaccard\"]\n",
        "f1 = history.history[\"f1_score\"]\n",
        "val_f1 = history.history[\"val_f1_score\"]\n",
        "\n",
        "epochs_range = range(len(loss))\n",
        "fig, ax = plt.subplots(1, 3, figsize = (20, 12))\n",
        "ax[0].plot(epochs_range, loss, label='Training Loss')\n",
        "ax[0].plot(epochs_range, val_loss, label='Validation Loss')\n",
        "\n",
        "ax[1].plot(epochs_range, Jaccard, label='Training Jaccard')\n",
        "ax[1].plot(epochs_range, val_Jaccard, label='Validation Jaccard')\n",
        "\n",
        "ax[2].plot(epochs_range, f1, label='Training f1-score')\n",
        "ax[2].plot(epochs_range, val_f1, label='Validation f1-score')\n",
        "plt.show()\n",
        "\n",
        "del model, history, train\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "dfH6L4HaNuGn",
        "outputId": "5af8d1d1-21f8-4029-9a39-cc72674739ed"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/2000\n",
            "4/4 [==============================] - ETA: 0s - loss: 0.5527 - Jaccard: 0.2880 - f1_score: 0.4472\n",
            "Epoch 1: val_f1_score improved from -inf to 0.44687, saving model to weights.h5\n",
            "4/4 [==============================] - 40s 2s/step - loss: 0.5527 - Jaccard: 0.2880 - f1_score: 0.4472 - val_loss: 0.5531 - val_Jaccard: 0.2877 - val_f1_score: 0.4469 - lr: 0.0050\n",
            "Epoch 2/2000\n",
            "4/4 [==============================] - ETA: 0s - loss: 0.5527 - Jaccard: 0.2881 - f1_score: 0.4473\n",
            "Epoch 2: val_f1_score did not improve from 0.44687\n",
            "4/4 [==============================] - 1s 365ms/step - loss: 0.5527 - Jaccard: 0.2881 - f1_score: 0.4473 - val_loss: 0.5531 - val_Jaccard: 0.2877 - val_f1_score: 0.4469 - lr: 0.0050\n",
            "Epoch 3/2000\n",
            "4/4 [==============================] - ETA: 0s - loss: 0.5527 - Jaccard: 0.2881 - f1_score: 0.4473\n",
            "Epoch 3: val_f1_score did not improve from 0.44687\n",
            "4/4 [==============================] - 1s 341ms/step - loss: 0.5527 - Jaccard: 0.2881 - f1_score: 0.4473 - val_loss: 0.5531 - val_Jaccard: 0.2877 - val_f1_score: 0.4469 - lr: 0.0050\n",
            "Epoch 4/2000\n",
            "4/4 [==============================] - ETA: 0s - loss: 0.5527 - Jaccard: 0.2881 - f1_score: 0.4473\n",
            "Epoch 4: val_f1_score did not improve from 0.44687\n",
            "4/4 [==============================] - 1s 359ms/step - loss: 0.5527 - Jaccard: 0.2881 - f1_score: 0.4473 - val_loss: 0.5531 - val_Jaccard: 0.2877 - val_f1_score: 0.4469 - lr: 0.0050\n",
            "Epoch 5/2000\n",
            "4/4 [==============================] - ETA: 0s - loss: 0.5527 - Jaccard: 0.2880 - f1_score: 0.4471\n",
            "Epoch 5: val_f1_score did not improve from 0.44687\n",
            "4/4 [==============================] - 1s 341ms/step - loss: 0.5527 - Jaccard: 0.2880 - f1_score: 0.4471 - val_loss: 0.5531 - val_Jaccard: 0.2877 - val_f1_score: 0.4469 - lr: 0.0050\n",
            "Epoch 6/2000\n",
            "4/4 [==============================] - ETA: 0s - loss: 0.5527 - Jaccard: 0.2880 - f1_score: 0.4472\n",
            "Epoch 6: val_f1_score did not improve from 0.44687\n",
            "\n",
            "Epoch 6: ReduceLROnPlateau reducing learning rate to 0.0024999999441206455.\n",
            "4/4 [==============================] - 1s 342ms/step - loss: 0.5527 - Jaccard: 0.2880 - f1_score: 0.4472 - val_loss: 0.5531 - val_Jaccard: 0.2877 - val_f1_score: 0.4469 - lr: 0.0050\n",
            "Epoch 7/2000\n",
            "4/4 [==============================] - ETA: 0s - loss: 0.5527 - Jaccard: 0.2881 - f1_score: 0.4473\n",
            "Epoch 7: val_f1_score did not improve from 0.44687\n",
            "4/4 [==============================] - 1s 343ms/step - loss: 0.5527 - Jaccard: 0.2881 - f1_score: 0.4473 - val_loss: 0.5531 - val_Jaccard: 0.2877 - val_f1_score: 0.4469 - lr: 0.0025\n",
            "Epoch 8/2000\n",
            "4/4 [==============================] - ETA: 0s - loss: 0.5527 - Jaccard: 0.2880 - f1_score: 0.4472\n",
            "Epoch 8: val_f1_score did not improve from 0.44687\n",
            "4/4 [==============================] - 1s 359ms/step - loss: 0.5527 - Jaccard: 0.2880 - f1_score: 0.4472 - val_loss: 0.5531 - val_Jaccard: 0.2877 - val_f1_score: 0.4469 - lr: 0.0025\n",
            "Epoch 9/2000\n",
            "4/4 [==============================] - ETA: 0s - loss: 0.5527 - Jaccard: 0.2881 - f1_score: 0.4473\n",
            "Epoch 9: val_f1_score did not improve from 0.44687\n",
            "4/4 [==============================] - 1s 343ms/step - loss: 0.5527 - Jaccard: 0.2881 - f1_score: 0.4473 - val_loss: 0.5531 - val_Jaccard: 0.2877 - val_f1_score: 0.4469 - lr: 0.0025\n",
            "Epoch 10/2000\n",
            "4/4 [==============================] - ETA: 0s - loss: 0.5527 - Jaccard: 0.2880 - f1_score: 0.4472\n",
            "Epoch 10: val_f1_score did not improve from 0.44687\n",
            "4/4 [==============================] - 1s 366ms/step - loss: 0.5527 - Jaccard: 0.2880 - f1_score: 0.4472 - val_loss: 0.5531 - val_Jaccard: 0.2877 - val_f1_score: 0.4469 - lr: 0.0025\n",
            "Epoch 11/2000\n",
            "4/4 [==============================] - ETA: 0s - loss: 0.5527 - Jaccard: 0.2881 - f1_score: 0.4473\n",
            "Epoch 11: val_f1_score did not improve from 0.44687\n",
            "\n",
            "Epoch 11: ReduceLROnPlateau reducing learning rate to 0.0012499999720603228.\n",
            "4/4 [==============================] - 2s 384ms/step - loss: 0.5527 - Jaccard: 0.2881 - f1_score: 0.4473 - val_loss: 0.5531 - val_Jaccard: 0.2877 - val_f1_score: 0.4469 - lr: 0.0025\n",
            "Epoch 11: early stopping\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 2000x1200 with 3 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "valid = [np.load(\"valid_x.npy\"), np.load(\"valid_y.npy\")]\n",
        "print(valid[0].shape)\n",
        "\n",
        "Predict(test=valid)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "BwiTLvMfHqw9",
        "outputId": "37304e53-365a-434c-e567-2eb566041977"
      },
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(220, 2000, 1)\n",
            "(None, 250, 128) (None, 250, 128)\n",
            "(None, 500, 64) (None, 500, 64)\n",
            "(None, 1000, 32) (None, 1000, 32)\n",
            "(None, 2000, 16) (None, 2000, 16)\n",
            "1/1 [==============================] - 1s 1s/step\n"
          ]
        },
        {
          "output_type": "error",
          "ename": "IndexError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mIndexError\u001b[0m                                Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-19-074d3b6d7965>\u001b[0m in \u001b[0;36m<cell line: 4>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvalid\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 4\u001b[0;31m \u001b[0mPredict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtest\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mvalid\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;32m<ipython-input-18-f457db321c9d>\u001b[0m in \u001b[0;36mPredict\u001b[0;34m(test, n_sample)\u001b[0m\n\u001b[1;32m     19\u001b[0m         \u001b[0mfig\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0max\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msubplots\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m3\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfigsize\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m20\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m12\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     20\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 21\u001b[0;31m         \u001b[0max\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mplot\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtest_ecg\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     22\u001b[0m         \u001b[0max\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset_title\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf\"ECG_{i}'s lead II\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfontsize\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m10\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     23\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mIndexError\u001b[0m: index 1 is out of bounds for axis 1 with size 1"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 2000x1200 with 3 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "APuaqkzRZlP8"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "RwC6WI8CHmTL"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "9cte4PKKGnhT"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "MT_dgaI3GlLL"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "6Edxfik3GhIS"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "3kYMY3GWgBSO"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "1yODZAJ5f8rg"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "7lAkl8Qsf5gq"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "ZhizyTosfKBt"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "tSywX23re5E1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "kXYjO0Owe019"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}