AD_classification.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "5ejnVL5f3Sb1"
      },
      "outputs": [],
      "source": [
        "import os \n",
        "import glob\n",
        "import shutil\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "import matplotlib.pyplot as plt\n",
        "from skimage.io import imread,imsave\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras.utils import plot_model\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "from tensorflow.keras.models import save_model, load_model, Model\n",
        "from tensorflow.keras.layers import Input,InputLayer, Dropout, Flatten, Dense, Activation, MaxPooling2D, Conv2D\n",
        "from sklearn.metrics import classification_report"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RMvVm55-4L4w",
        "outputId": "543c6d19-e902-4af3-876b-c64ef2dd6e09"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Tue May 31 21:52:20 2022       \n",
            "+-----------------------------------------------------------------------------+\n",
            "| NVIDIA-SMI 460.32.03    Driver Version: 460.32.03    CUDA Version: 11.2     |\n",
            "|-------------------------------+----------------------+----------------------+\n",
            "| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
            "| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\n",
            "|                               |                      |               MIG M. |\n",
            "|===============================+======================+======================|\n",
            "|   0  Tesla P100-PCIE...  Off  | 00000000:00:04.0 Off |                    0 |\n",
            "| N/A   41C    P0    30W / 250W |      0MiB / 16280MiB |      0%      Default |\n",
            "|                               |                      |                  N/A |\n",
            "+-------------------------------+----------------------+----------------------+\n",
            "                                                                               \n",
            "+-----------------------------------------------------------------------------+\n",
            "| Processes:                                                                  |\n",
            "|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\n",
            "|        ID   ID                                                   Usage      |\n",
            "|=============================================================================|\n",
            "|  No running processes found                                                 |\n",
            "+-----------------------------------------------------------------------------+\n"
          ]
        }
      ],
      "source": [
        "!nvidia-smi"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6J7LEgrPHLBb"
      },
      "source": [
        "# Downloading and Fixing the data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "U9hIEcvCscB9"
      },
      "outputs": [],
      "source": [
        "def flatten(t):\n",
        "    return [item for sublist in t for item in sublist]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "D83RsbSktzOZ",
        "outputId": "d6bcfed0-b290-4b53-e9a5-628d7da7c16b"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/gdown/cli.py:131: FutureWarning: Option `--id` was deprecated in version 4.3.1 and will be removed in 5.0. You don't need to pass it anymore to use a file ID.\n",
            "  category=FutureWarning,\n"
          ]
        }
      ],
      "source": [
        "!gdown --id '1HuthY7iMaKLWqPffAAQOxydr-fhgQyzl' --o '/content/sample_data/AFFINE_REG_SKS_MASK_2D.zip' --q\n",
        "!unzip -q '/content/sample_data/AFFINE_REG_SKS_MASK_2D.zip' -d '/content/sample_data/'"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "fvX3srgQuxKS"
      },
      "outputs": [],
      "source": [
        "root = '/content/sample_data/AFFINE_REG_SKS_MASK_2D/'\n",
        "data = testing = pd.read_csv('/content/drive/MyDrive/ADNI/MY_ADNI_CLEAN.csv')\n",
        "x_train, x_test, y_train, y_test = train_test_split(data[['Image Data ID']].values, data[['Group']].values, test_size=0.20, random_state=13 )\n",
        "\n",
        "training = flatten(x_train.tolist())\n",
        "testing = flatten(x_test.tolist())\n",
        "data.loc[data['Image Data ID'].isin(training)].to_csv('/content/sample_data/training.csv', index=False, encoding='utf-8-sig')\n",
        "data.loc[data['Image Data ID'].isin(testing)].to_csv('/content/sample_data/testing.csv', index=False, encoding='utf-8-sig')\n",
        "traning = pd.read_csv('/content/sample_data/training.csv')\n",
        "testing = pd.read_csv('/content/sample_data/testing.csv')\n",
        "\n",
        "image_Shape=(436,364,3)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "nmFokic07Isc"
      },
      "outputs": [],
      "source": [
        "l = ['/testing/CN/','/testing/AD/','/testing/MCI/','/training/CN/','/training/AD/','/training/MCI/']\n",
        "for elem in l:\n",
        "  if not os.path.isdir(root+elem):\n",
        "      os.makedirs(root+elem)\n",
        "      \n",
        "for index ,row in traning.iterrows():\n",
        "  shutil.move(f'{root}{row[\"Image Data ID\"]}.png',f'{root}training/{row[\"Group\"]}/{row[\"Image Data ID\"]}.png')\n",
        "  \n",
        "\n",
        "for index ,row in testing.iterrows():\n",
        "  shutil.move(f'{root}{row[\"Image Data ID\"]}.png',f'{root}testing/{row[\"Group\"]}/{row[\"Image Data ID\"]}.png')\n",
        "  "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2-g9yhpLHmOh"
      },
      "source": [
        "# Models"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "A90ruKTAHoaG"
      },
      "outputs": [],
      "source": [
        "def resnet50_pretrained(optimizer, loss,metrics,shape):\n",
        "  image_input = tf.keras.applications.ResNet50(include_top=False,weights='imagenet',input_shape=shape)\n",
        "\n",
        "  for layer in image_input.layers:\n",
        "      layer.trainable = False\n",
        "      layer._name = layer._name  + str('_img')\n",
        "\n",
        "  y = Flatten() (image_input.output)\n",
        "  c = Dropout(0.4) (y)\n",
        "  c = Dense(512) (c)\n",
        "  c = Dense(256) (c)\n",
        "  c = Dense(128) (c)\n",
        "  output_layer = Dense(3, activation='softmax')(c)\n",
        "\n",
        "  model = Model(inputs=image_input.input,outputs=output_layer)\n",
        "  model.compile(optimizer=optimizer, loss=loss, metrics=metrics)\n",
        "  return model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "uH-nbiisAgZz"
      },
      "outputs": [],
      "source": [
        "def resnet50_row(optimizer, loss,metrics,shape):\n",
        "  image_input = tf.keras.applications.ResNet50(include_top=False,weights='imagenet',input_shape=shape)\n",
        "\n",
        "  for layer in image_input.layers:\n",
        "      layer.trainable = True\n",
        "      layer._name = layer._name  + str('_img')\n",
        "\n",
        "  y = Flatten() (image_input.output)\n",
        "  c = Dropout(0.4) (y)\n",
        "  c = Dense(512) (c)\n",
        "  c = Dense(256) (c)\n",
        "  c = Dense(128) (c)\n",
        "  output_layer = Dense(3, activation='softmax')(c)\n",
        "\n",
        "  model = Model(inputs=image_input.input,outputs=output_layer)\n",
        "  model.compile(optimizer=optimizer, loss=loss, metrics=metrics)\n",
        "  return model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "bMOUYwH55Lug"
      },
      "outputs": [],
      "source": [
        "def plot_confusion_matrix(true_lab,predict_lab,classes,path):\n",
        "  matrix_confusion = tf.math.confusion_matrix(labels=true_lab, predictions=predicted_lab, num_classes=len(classes))\n",
        "  ax = plt.subplot()\n",
        "  sns.heatmap(matrix_confusion, square=True, annot=True, cmap='Blues', fmt='d', cbar=True)\n",
        "  ax.set_xlabel('Predicted labels')\n",
        "  ax.set_ylabel('True labels')\n",
        "  ax.set_title('Confusion Matrix')\n",
        "  ax.xaxis.set_ticklabels(classes)\n",
        "  ax.yaxis.set_ticklabels(classes)\n",
        "  plt.savefig(f'{path}confusion_matrix.png')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "iYjB-l_pdnKM"
      },
      "outputs": [],
      "source": [
        "def dump_text(text,path):\n",
        "  text_file = open(path, \"w\")\n",
        "  text_file.write(text)\n",
        "  text_file.close()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "bDzKxJRQjeeI"
      },
      "outputs": [],
      "source": [
        "config = tf.compat.v1.ConfigProto()\n",
        "config.gpu_options.allow_growth = True\n",
        "session = tf.compat.v1.Session(config=config)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "e3KYSPlwND1B",
        "outputId": "c7c6e7e6-7caa-467c-90c0-f5412c339955"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Found 2789 images belonging to 3 classes.\n",
            "Found 698 images belonging to 3 classes.\n"
          ]
        }
      ],
      "source": [
        "training_image_datagen = ImageDataGenerator(horizontal_flip=True).flow_from_directory(\n",
        "                                                  directory=f'{root}/training/',\n",
        "                                                  target_size=(image_Shape[0], image_Shape[1]),\n",
        "                                                  class_mode='categorical',\n",
        "                                                  batch_size=4\n",
        "                                                  )\n",
        "\n",
        "testing_image_datagen = ImageDataGenerator().flow_from_directory(\n",
        "                                                  directory = f'{root}/testing/' ,\n",
        "                                                  target_size = (image_Shape[0] ,image_Shape[1]) ,\n",
        "                                                  class_mode = 'categorical' ,\n",
        "                                                  batch_size=1 ,\n",
        "                                                  shuffle = False\n",
        "                                                  )"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ufi69uuBNUDu",
        "outputId": "1e38e620-f00d-4684-f29b-763215b58204"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "{'AD': 0, 'CN': 1, 'MCI': 2}"
            ]
          },
          "execution_count": 10,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "training_image_datagen.class_indices"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "XSBFpBUgKGdr"
      },
      "outputs": [],
      "source": [
        "EPOCH = 50\n",
        "METRICS = [\n",
        "    tf.keras.metrics.CategoricalAccuracy(name='Accuracy'),\n",
        "    tf.keras.metrics.Precision(name='AD_Precision',class_id=0),\n",
        "    tf.keras.metrics.Precision(name='CN_Precision',class_id=1),\n",
        "    tf.keras.metrics.Precision(name='MCI_Precision',class_id=2),\n",
        "    tf.keras.metrics.Recall(name='AD_Recall',class_id=0),\n",
        "    tf.keras.metrics.Recall(name='CN_Recall',class_id=1),\n",
        "    tf.keras.metrics.Recall(name='MCI_Recall',class_id=2),\n",
        "    ]\n",
        "LOOS = 'categorical_crossentropy'"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zk132K6ABdgs"
      },
      "source": [
        "# **Resnet 50 pretrained**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MqN2_tGOH_IN"
      },
      "source": [
        "### learning_rate=10E-4"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dtCRREYKJ3OK",
        "outputId": "8a1e0346-e519-4f3c-ae78-4653e8863e73"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "698/698 [==============================] - 72s 81ms/step - loss: 12.8523 - Accuracy: 0.4690 - AD_Precision: 0.2717 - CN_Precision: 0.4260 - MCI_Precision: 0.5622 - AD_Recall: 0.2465 - CN_Recall: 0.4140 - MCI_Recall: 0.5634 - val_loss: 1.5681 - val_Accuracy: 0.4040 - val_AD_Precision: 0.7500 - val_CN_Precision: 0.3298 - val_MCI_Precision: 0.7870 - val_AD_Recall: 0.0236 - val_CN_Recall: 0.9442 - val_MCI_Recall: 0.2273\n",
            "Epoch 2/50\n",
            "698/698 [==============================] - 53s 75ms/step - loss: 1.7410 - Accuracy: 0.5342 - AD_Precision: 0.3543 - CN_Precision: 0.5111 - MCI_Precision: 0.6120 - AD_Recall: 0.3116 - CN_Recall: 0.4830 - MCI_Recall: 0.6125 - val_loss: 2.3237 - val_Accuracy: 0.3453 - val_AD_Precision: 0.2332 - val_CN_Precision: 1.0000 - val_MCI_Precision: 0.6279 - val_AD_Recall: 0.9291 - val_CN_Recall: 0.0609 - val_MCI_Recall: 0.2888\n",
            "Epoch 3/50\n",
            "698/698 [==============================] - 52s 75ms/step - loss: 2.4042 - Accuracy: 0.5561 - AD_Precision: 0.3603 - CN_Precision: 0.5528 - MCI_Precision: 0.6374 - AD_Recall: 0.3511 - CN_Recall: 0.5392 - MCI_Recall: 0.6307 - val_loss: 2.1516 - val_Accuracy: 0.5358 - val_AD_Precision: 0.3301 - val_CN_Precision: 0.5280 - val_MCI_Precision: 0.6971 - val_AD_Recall: 0.5433 - val_CN_Recall: 0.4315 - val_MCI_Recall: 0.5722\n",
            "Epoch 4/50\n",
            "698/698 [==============================] - 52s 75ms/step - loss: 2.8818 - Accuracy: 0.5783 - AD_Precision: 0.4020 - CN_Precision: 0.5562 - MCI_Precision: 0.6621 - AD_Recall: 0.4043 - CN_Recall: 0.5556 - MCI_Recall: 0.6496 - val_loss: 2.0154 - val_Accuracy: 0.5888 - val_AD_Precision: 0.6164 - val_CN_Precision: 0.9091 - val_MCI_Precision: 0.5738 - val_AD_Recall: 0.3543 - val_CN_Recall: 0.1015 - val_MCI_Recall: 0.9251\n",
            "Epoch 5/50\n",
            "698/698 [==============================] - 53s 76ms/step - loss: 1.9368 - Accuracy: 0.6605 - AD_Precision: 0.5292 - CN_Precision: 0.6538 - MCI_Precision: 0.7182 - AD_Recall: 0.5187 - CN_Recall: 0.6538 - MCI_Recall: 0.7092 - val_loss: 2.6856 - val_Accuracy: 0.6060 - val_AD_Precision: 0.5588 - val_CN_Precision: 0.9375 - val_MCI_Precision: 0.6052 - val_AD_Recall: 0.4488 - val_CN_Recall: 0.0761 - val_MCI_Recall: 0.9385\n",
            "Epoch 6/50\n",
            "698/698 [==============================] - 52s 75ms/step - loss: 2.6577 - Accuracy: 0.6350 - AD_Precision: 0.4949 - CN_Precision: 0.6234 - MCI_Precision: 0.6927 - AD_Recall: 0.4832 - CN_Recall: 0.6175 - MCI_Recall: 0.6966 - val_loss: 1.6640 - val_Accuracy: 0.7149 - val_AD_Precision: 0.8500 - val_CN_Precision: 0.6368 - val_MCI_Precision: 0.7427 - val_AD_Recall: 0.4016 - val_CN_Recall: 0.7208 - val_MCI_Recall: 0.8182\n",
            "Epoch 7/50\n",
            "698/698 [==============================] - 52s 75ms/step - loss: 3.2114 - Accuracy: 0.6486 - AD_Precision: 0.5329 - CN_Precision: 0.6328 - MCI_Precision: 0.7029 - AD_Recall: 0.5424 - CN_Recall: 0.6269 - MCI_Recall: 0.6980 - val_loss: 3.2290 - val_Accuracy: 0.6103 - val_AD_Precision: 0.8103 - val_CN_Precision: 0.4579 - val_MCI_Precision: 0.8283 - val_AD_Recall: 0.3701 - val_CN_Recall: 0.9391 - val_MCI_Recall: 0.5160\n",
            "Epoch 8/50\n",
            "698/698 [==============================] - 52s 74ms/step - loss: 2.1318 - Accuracy: 0.7261 - AD_Precision: 0.6195 - CN_Precision: 0.7215 - MCI_Precision: 0.7675 - AD_Recall: 0.6134 - CN_Recall: 0.7181 - MCI_Recall: 0.7680 - val_loss: 3.3305 - val_Accuracy: 0.5358 - val_AD_Precision: 0.4107 - val_CN_Precision: 0.5108 - val_MCI_Precision: 0.9167 - val_AD_Recall: 0.7244 - val_CN_Recall: 0.9594 - val_MCI_Recall: 0.2353\n",
            "Epoch 9/50\n",
            "698/698 [==============================] - 52s 74ms/step - loss: 2.3917 - Accuracy: 0.7157 - AD_Precision: 0.5916 - CN_Precision: 0.7244 - MCI_Precision: 0.7565 - AD_Recall: 0.5858 - CN_Recall: 0.7287 - MCI_Recall: 0.7533 - val_loss: 2.3328 - val_Accuracy: 0.7350 - val_AD_Precision: 0.6812 - val_CN_Precision: 0.8623 - val_MCI_Precision: 0.7067 - val_AD_Recall: 0.3701 - val_CN_Recall: 0.6041 - val_MCI_Recall: 0.9278\n",
            "Epoch 10/50\n",
            "698/698 [==============================] - 52s 75ms/step - loss: 2.7379 - Accuracy: 0.7314 - AD_Precision: 0.6403 - CN_Precision: 0.7168 - MCI_Precision: 0.7730 - AD_Recall: 0.6391 - CN_Recall: 0.7193 - MCI_Recall: 0.7708 - val_loss: 6.6638 - val_Accuracy: 0.4441 - val_AD_Precision: 0.3267 - val_CN_Precision: 0.5344 - val_MCI_Precision: 0.9600 - val_AD_Recall: 0.9055 - val_CN_Recall: 0.8680 - val_MCI_Recall: 0.0642\n",
            "Epoch 11/50\n",
            "698/698 [==============================] - 52s 75ms/step - loss: 2.0487 - Accuracy: 0.7695 - AD_Precision: 0.6812 - CN_Precision: 0.7646 - MCI_Precision: 0.8053 - AD_Recall: 0.6785 - CN_Recall: 0.7673 - MCI_Recall: 0.8031 - val_loss: 1.1153 - val_Accuracy: 0.8324 - val_AD_Precision: 0.8667 - val_CN_Precision: 0.8325 - val_MCI_Precision: 0.8262 - val_AD_Recall: 0.5118 - val_CN_Recall: 0.8579 - val_MCI_Recall: 0.9278\n",
            "Epoch 12/50\n",
            "698/698 [==============================] - 52s 75ms/step - loss: 1.8254 - Accuracy: 0.8010 - AD_Precision: 0.6798 - CN_Precision: 0.8236 - MCI_Precision: 0.8313 - AD_Recall: 0.6824 - CN_Recall: 0.8246 - MCI_Recall: 0.8290 - val_loss: 1.1453 - val_Accuracy: 0.8037 - val_AD_Precision: 0.7018 - val_CN_Precision: 0.8547 - val_MCI_Precision: 0.8107 - val_AD_Recall: 0.6299 - val_CN_Recall: 0.7462 - val_MCI_Recall: 0.8930\n",
            "Epoch 13/50\n",
            "698/698 [==============================] - 52s 75ms/step - loss: 1.8968 - Accuracy: 0.7985 - AD_Precision: 0.7194 - CN_Precision: 0.8012 - MCI_Precision: 0.8261 - AD_Recall: 0.7179 - CN_Recall: 0.8012 - MCI_Recall: 0.8255 - val_loss: 1.3095 - val_Accuracy: 0.8367 - val_AD_Precision: 0.6859 - val_CN_Precision: 0.8902 - val_MCI_Precision: 0.8777 - val_AD_Recall: 0.8425 - val_CN_Recall: 0.7817 - val_MCI_Recall: 0.8636\n",
            "Epoch 14/50\n",
            "698/698 [==============================] - 52s 75ms/step - loss: 1.3984 - Accuracy: 0.8440 - AD_Precision: 0.7843 - CN_Precision: 0.8369 - MCI_Precision: 0.8714 - AD_Recall: 0.7890 - CN_Recall: 0.8339 - MCI_Recall: 0.8690 - val_loss: 3.9878 - val_Accuracy: 0.6777 - val_AD_Precision: 0.7822 - val_CN_Precision: 1.0000 - val_MCI_Precision: 0.6412 - val_AD_Recall: 0.6220 - val_CN_Recall: 0.1675 - val_MCI_Recall: 0.9652\n",
            "Epoch 15/50\n",
            "698/698 [==============================] - 52s 74ms/step - loss: 2.1217 - Accuracy: 0.8139 - AD_Precision: 0.7253 - CN_Precision: 0.8110 - MCI_Precision: 0.8476 - AD_Recall: 0.7239 - CN_Recall: 0.8129 - MCI_Recall: 0.8458 - val_loss: 1.9617 - val_Accuracy: 0.8266 - val_AD_Precision: 0.8500 - val_CN_Precision: 0.9363 - val_MCI_Precision: 0.7852 - val_AD_Recall: 0.5354 - val_CN_Recall: 0.7462 - val_MCI_Recall: 0.9679\n",
            "Epoch 16/50\n",
            "698/698 [==============================] - 52s 75ms/step - loss: 1.5391 - Accuracy: 0.8501 - AD_Precision: 0.7897 - CN_Precision: 0.8475 - MCI_Precision: 0.8731 - AD_Recall: 0.7850 - CN_Recall: 0.8515 - MCI_Recall: 0.8725 - val_loss: 3.1007 - val_Accuracy: 0.7564 - val_AD_Precision: 0.6229 - val_CN_Precision: 0.9756 - val_MCI_Precision: 0.7682 - val_AD_Recall: 0.8583 - val_CN_Recall: 0.4061 - val_MCI_Recall: 0.9037\n",
            "Epoch 17/50\n",
            "698/698 [==============================] - 52s 75ms/step - loss: 1.9826 - Accuracy: 0.8315 - AD_Precision: 0.7628 - CN_Precision: 0.8230 - MCI_Precision: 0.8606 - AD_Recall: 0.7613 - CN_Recall: 0.8211 - MCI_Recall: 0.8612 - val_loss: 4.8730 - val_Accuracy: 0.7579 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.8172 - val_MCI_Precision: 0.7262 - val_AD_Recall: 0.1496 - val_CN_Recall: 0.7716 - val_MCI_Recall: 0.9572\n",
            "Epoch 18/50\n",
            "698/698 [==============================] - 53s 75ms/step - loss: 1.4347 - Accuracy: 0.8627 - AD_Precision: 0.7996 - CN_Precision: 0.8653 - MCI_Precision: 0.8833 - AD_Recall: 0.7949 - CN_Recall: 0.8643 - MCI_Recall: 0.8858 - val_loss: 3.6897 - val_Accuracy: 0.6977 - val_AD_Precision: 0.8100 - val_CN_Precision: 0.5147 - val_MCI_Precision: 0.9595 - val_AD_Recall: 0.6378 - val_CN_Recall: 0.9797 - val_MCI_Recall: 0.5695\n",
            "Epoch 19/50\n",
            "698/698 [==============================] - 53s 76ms/step - loss: 1.5031 - Accuracy: 0.8716 - AD_Precision: 0.8209 - CN_Precision: 0.8675 - MCI_Precision: 0.8922 - AD_Recall: 0.8225 - CN_Recall: 0.8655 - MCI_Recall: 0.8928 - val_loss: 4.5030 - val_Accuracy: 0.6691 - val_AD_Precision: 0.6579 - val_CN_Precision: 0.5301 - val_MCI_Precision: 0.9665 - val_AD_Recall: 0.7874 - val_CN_Recall: 0.9848 - val_MCI_Recall: 0.4626\n",
            "Epoch 20/50\n",
            "698/698 [==============================] - 57s 81ms/step - loss: 1.5853 - Accuracy: 0.8623 - AD_Precision: 0.8300 - CN_Precision: 0.8538 - MCI_Precision: 0.8789 - AD_Recall: 0.8284 - CN_Recall: 0.8608 - MCI_Recall: 0.8746 - val_loss: 6.3203 - val_Accuracy: 0.7149 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.9664 - val_MCI_Precision: 0.6567 - val_AD_Recall: 0.0866 - val_CN_Recall: 0.5838 - val_MCI_Recall: 0.9973\n",
            "Epoch 21/50\n",
            "698/698 [==============================] - 58s 82ms/step - loss: 1.7180 - Accuracy: 0.8706 - AD_Precision: 0.8139 - CN_Precision: 0.8735 - MCI_Precision: 0.8888 - AD_Recall: 0.8107 - CN_Recall: 0.8725 - MCI_Recall: 0.8907 - val_loss: 1.8447 - val_Accuracy: 0.8668 - val_AD_Precision: 0.7403 - val_CN_Precision: 0.9398 - val_MCI_Precision: 0.8862 - val_AD_Recall: 0.8976 - val_CN_Recall: 0.7919 - val_MCI_Recall: 0.8957\n",
            "Epoch 22/50\n",
            "698/698 [==============================] - 58s 84ms/step - loss: 1.4614 - Accuracy: 0.8878 - AD_Precision: 0.8330 - CN_Precision: 0.8902 - MCI_Precision: 0.9059 - AD_Recall: 0.8363 - CN_Recall: 0.8912 - MCI_Recall: 0.9040 - val_loss: 1.5569 - val_Accuracy: 0.8768 - val_AD_Precision: 0.8203 - val_CN_Precision: 0.8034 - val_MCI_Precision: 0.9494 - val_AD_Recall: 0.8268 - val_CN_Recall: 0.9543 - val_MCI_Recall: 0.8529\n",
            "Epoch 23/50\n",
            "698/698 [==============================] - 57s 82ms/step - loss: 1.3513 - Accuracy: 0.8856 - AD_Precision: 0.8343 - CN_Precision: 0.8927 - MCI_Precision: 0.9002 - AD_Recall: 0.8343 - CN_Recall: 0.8854 - MCI_Recall: 0.9040 - val_loss: 3.7327 - val_Accuracy: 0.8166 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.8913 - val_MCI_Precision: 0.7731 - val_AD_Recall: 0.2992 - val_CN_Recall: 0.8325 - val_MCI_Recall: 0.9840\n",
            "Epoch 24/50\n",
            "698/698 [==============================] - 68s 98ms/step - loss: 1.5538 - Accuracy: 0.8860 - AD_Precision: 0.8656 - CN_Precision: 0.8699 - MCI_Precision: 0.9030 - AD_Recall: 0.8639 - CN_Recall: 0.8760 - MCI_Recall: 0.8998 - val_loss: 7.7358 - val_Accuracy: 0.6404 - val_AD_Precision: 0.3481 - val_CN_Precision: 0.9259 - val_MCI_Precision: 0.9828 - val_AD_Recall: 0.9921 - val_CN_Recall: 0.7614 - val_MCI_Recall: 0.4572\n",
            "Epoch 25/50\n",
            "698/698 [==============================] - 60s 86ms/step - loss: 0.9922 - Accuracy: 0.9132 - AD_Precision: 0.8730 - CN_Precision: 0.9165 - MCI_Precision: 0.9257 - AD_Recall: 0.8817 - CN_Recall: 0.9111 - MCI_Recall: 0.9257 - val_loss: 6.0907 - val_Accuracy: 0.7851 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.7839 - val_MCI_Precision: 0.7800 - val_AD_Recall: 0.0945 - val_CN_Recall: 0.9391 - val_MCI_Recall: 0.9385\n",
            "Epoch 26/50\n",
            "698/698 [==============================] - 67s 95ms/step - loss: 1.9865 - Accuracy: 0.8788 - AD_Precision: 0.8399 - CN_Precision: 0.8751 - MCI_Precision: 0.8948 - AD_Recall: 0.8383 - CN_Recall: 0.8772 - MCI_Recall: 0.8942 - val_loss: 3.0247 - val_Accuracy: 0.8281 - val_AD_Precision: 0.9324 - val_CN_Precision: 0.9858 - val_MCI_Precision: 0.7660 - val_AD_Recall: 0.5433 - val_CN_Recall: 0.7056 - val_MCI_Recall: 0.9893\n",
            "Epoch 27/50\n",
            "698/698 [==============================] - 74s 106ms/step - loss: 0.9193 - Accuracy: 0.9136 - AD_Precision: 0.8950 - CN_Precision: 0.9076 - MCI_Precision: 0.9244 - AD_Recall: 0.8915 - CN_Recall: 0.9076 - MCI_Recall: 0.9250 - val_loss: 1.6317 - val_Accuracy: 0.8883 - val_AD_Precision: 0.9109 - val_CN_Precision: 0.8578 - val_MCI_Precision: 0.9013 - val_AD_Recall: 0.7244 - val_CN_Recall: 0.9188 - val_MCI_Recall: 0.9278\n",
            "Epoch 28/50\n",
            "698/698 [==============================] - 60s 87ms/step - loss: 1.0713 - Accuracy: 0.9154 - AD_Precision: 0.8720 - CN_Precision: 0.9184 - MCI_Precision: 0.9290 - AD_Recall: 0.8738 - CN_Recall: 0.9216 - MCI_Recall: 0.9264 - val_loss: 1.6693 - val_Accuracy: 0.8825 - val_AD_Precision: 0.9109 - val_CN_Precision: 0.9222 - val_MCI_Precision: 0.8582 - val_AD_Recall: 0.7244 - val_CN_Recall: 0.8426 - val_MCI_Recall: 0.9545\n",
            "Epoch 29/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 1.8334 - Accuracy: 0.8906 - AD_Precision: 0.8304 - CN_Precision: 0.8893 - MCI_Precision: 0.9129 - AD_Recall: 0.8304 - CN_Recall: 0.8924 - MCI_Recall: 0.9110 - val_loss: 5.9325 - val_Accuracy: 0.7937 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.9769 - val_MCI_Precision: 0.7257 - val_AD_Recall: 0.4252 - val_CN_Recall: 0.6447 - val_MCI_Recall: 0.9973\n",
            "Epoch 30/50\n",
            "698/698 [==============================] - 53s 76ms/step - loss: 1.2212 - Accuracy: 0.9229 - AD_Precision: 0.8913 - CN_Precision: 0.9195 - MCI_Precision: 0.9362 - AD_Recall: 0.8895 - CN_Recall: 0.9216 - MCI_Recall: 0.9355 - val_loss: 1.9938 - val_Accuracy: 0.8797 - val_AD_Precision: 0.8932 - val_CN_Precision: 0.8166 - val_MCI_Precision: 0.9153 - val_AD_Recall: 0.7244 - val_CN_Recall: 0.9492 - val_MCI_Recall: 0.8957\n",
            "Epoch 31/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.8607 - Accuracy: 0.9380 - AD_Precision: 0.9073 - CN_Precision: 0.9428 - MCI_Precision: 0.9460 - AD_Recall: 0.9073 - CN_Recall: 0.9439 - MCI_Recall: 0.9453 - val_loss: 2.1831 - val_Accuracy: 0.8797 - val_AD_Precision: 0.9010 - val_CN_Precision: 0.9419 - val_MCI_Precision: 0.8494 - val_AD_Recall: 0.7165 - val_CN_Recall: 0.8223 - val_MCI_Recall: 0.9652\n",
            "Epoch 32/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 1.8589 - Accuracy: 0.8842 - AD_Precision: 0.8234 - CN_Precision: 0.8868 - MCI_Precision: 0.9047 - AD_Recall: 0.8185 - CN_Recall: 0.8889 - MCI_Recall: 0.9047 - val_loss: 2.6665 - val_Accuracy: 0.8367 - val_AD_Precision: 0.8174 - val_CN_Precision: 0.7176 - val_MCI_Precision: 0.9408 - val_AD_Recall: 0.7402 - val_CN_Recall: 0.9543 - val_MCI_Recall: 0.8075\n",
            "Epoch 33/50\n",
            "698/698 [==============================] - 60s 86ms/step - loss: 1.5017 - Accuracy: 0.9096 - AD_Precision: 0.8625 - CN_Precision: 0.9098 - MCI_Precision: 0.9264 - AD_Recall: 0.8659 - CN_Recall: 0.9088 - MCI_Recall: 0.9257 - val_loss: 3.0062 - val_Accuracy: 0.8539 - val_AD_Precision: 0.9342 - val_CN_Precision: 0.8350 - val_MCI_Precision: 0.8486 - val_AD_Recall: 0.5591 - val_CN_Recall: 0.8731 - val_MCI_Recall: 0.9439\n",
            "Epoch 34/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.8362 - Accuracy: 0.9351 - AD_Precision: 0.9253 - CN_Precision: 0.9219 - MCI_Precision: 0.9466 - AD_Recall: 0.9290 - CN_Recall: 0.9251 - MCI_Recall: 0.9432 - val_loss: 4.2336 - val_Accuracy: 0.8424 - val_AD_Precision: 0.9492 - val_CN_Precision: 0.8877 - val_MCI_Precision: 0.8093 - val_AD_Recall: 0.4409 - val_CN_Recall: 0.8426 - val_MCI_Recall: 0.9759\n",
            "Epoch 35/50\n",
            "698/698 [==============================] - 65s 93ms/step - loss: 1.2780 - Accuracy: 0.9168 - AD_Precision: 0.8841 - CN_Precision: 0.9111 - MCI_Precision: 0.9319 - AD_Recall: 0.8876 - CN_Recall: 0.9111 - MCI_Recall: 0.9306 - val_loss: 5.5624 - val_Accuracy: 0.7564 - val_AD_Precision: 0.4664 - val_CN_Precision: 0.8920 - val_MCI_Precision: 0.9405 - val_AD_Recall: 0.9291 - val_CN_Recall: 0.7970 - val_MCI_Recall: 0.6765\n",
            "Epoch 36/50\n",
            "698/698 [==============================] - 59s 84ms/step - loss: 1.0677 - Accuracy: 0.9297 - AD_Precision: 0.8998 - CN_Precision: 0.9346 - MCI_Precision: 0.9375 - AD_Recall: 0.9034 - CN_Recall: 0.9357 - MCI_Recall: 0.9355 - val_loss: 5.0822 - val_Accuracy: 0.8295 - val_AD_Precision: 0.9565 - val_CN_Precision: 0.9153 - val_MCI_Precision: 0.7819 - val_AD_Recall: 0.3465 - val_CN_Recall: 0.8782 - val_MCI_Recall: 0.9679\n",
            "Epoch 37/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 1.2129 - Accuracy: 0.9236 - AD_Precision: 0.8856 - CN_Precision: 0.9190 - MCI_Precision: 0.9399 - AD_Recall: 0.8856 - CN_Recall: 0.9158 - MCI_Recall: 0.9418 - val_loss: 3.1201 - val_Accuracy: 0.8811 - val_AD_Precision: 0.8583 - val_CN_Precision: 0.9512 - val_MCI_Precision: 0.8599 - val_AD_Recall: 0.8110 - val_CN_Recall: 0.7919 - val_MCI_Recall: 0.9519\n",
            "Epoch 38/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.9091 - Accuracy: 0.9426 - AD_Precision: 0.9245 - CN_Precision: 0.9439 - MCI_Precision: 0.9483 - AD_Recall: 0.9172 - CN_Recall: 0.9450 - MCI_Recall: 0.9502 - val_loss: 4.0024 - val_Accuracy: 0.8453 - val_AD_Precision: 0.9688 - val_CN_Precision: 0.9056 - val_MCI_Precision: 0.8040 - val_AD_Recall: 0.4882 - val_CN_Recall: 0.8274 - val_MCI_Recall: 0.9759\n",
            "Epoch 39/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.9923 - Accuracy: 0.9333 - AD_Precision: 0.9077 - CN_Precision: 0.9288 - MCI_Precision: 0.9459 - AD_Recall: 0.9112 - CN_Recall: 0.9310 - MCI_Recall: 0.9425 - val_loss: 2.7148 - val_Accuracy: 0.8825 - val_AD_Precision: 0.9647 - val_CN_Precision: 0.8077 - val_MCI_Precision: 0.9103 - val_AD_Recall: 0.6457 - val_CN_Recall: 0.9594 - val_MCI_Recall: 0.9225\n",
            "Epoch 40/50\n",
            "698/698 [==============================] - 53s 76ms/step - loss: 0.7960 - Accuracy: 0.9394 - AD_Precision: 0.9331 - CN_Precision: 0.9342 - MCI_Precision: 0.9448 - AD_Recall: 0.9349 - CN_Recall: 0.9298 - MCI_Recall: 0.9467 - val_loss: 5.5115 - val_Accuracy: 0.7178 - val_AD_Precision: 0.7661 - val_CN_Precision: 0.5439 - val_MCI_Precision: 0.9683 - val_AD_Recall: 0.7480 - val_CN_Recall: 0.9746 - val_MCI_Recall: 0.5722\n",
            "Epoch 41/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 1.3683 - Accuracy: 0.9276 - AD_Precision: 0.8957 - CN_Precision: 0.9345 - MCI_Precision: 0.9348 - AD_Recall: 0.8974 - CN_Recall: 0.9345 - MCI_Recall: 0.9341 - val_loss: 9.7699 - val_Accuracy: 0.6404 - val_AD_Precision: 0.3475 - val_CN_Precision: 0.9091 - val_MCI_Precision: 0.9809 - val_AD_Recall: 0.9685 - val_CN_Recall: 0.8629 - val_MCI_Recall: 0.4118\n",
            "Epoch 42/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 1.0470 - Accuracy: 0.9333 - AD_Precision: 0.9077 - CN_Precision: 0.9343 - MCI_Precision: 0.9425 - AD_Recall: 0.9112 - CN_Recall: 0.9322 - MCI_Recall: 0.9418 - val_loss: 6.3556 - val_Accuracy: 0.8324 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.8706 - val_MCI_Precision: 0.8026 - val_AD_Recall: 0.2835 - val_CN_Recall: 0.8883 - val_MCI_Recall: 0.9893\n",
            "Epoch 43/50\n",
            "698/698 [==============================] - 53s 77ms/step - loss: 1.4480 - Accuracy: 0.9251 - AD_Precision: 0.9118 - CN_Precision: 0.9123 - MCI_Precision: 0.9375 - AD_Recall: 0.9172 - CN_Recall: 0.9123 - MCI_Recall: 0.9355 - val_loss: 7.8925 - val_Accuracy: 0.8052 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.8069 - val_MCI_Precision: 0.7941 - val_AD_Recall: 0.1811 - val_CN_Recall: 0.9543 - val_MCI_Recall: 0.9385\n",
            "Epoch 44/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 1.0501 - Accuracy: 0.9387 - AD_Precision: 0.9311 - CN_Precision: 0.9252 - MCI_Precision: 0.9495 - AD_Recall: 0.9329 - CN_Recall: 0.9263 - MCI_Recall: 0.9481 - val_loss: 5.0280 - val_Accuracy: 0.8152 - val_AD_Precision: 0.9833 - val_CN_Precision: 0.6690 - val_MCI_Precision: 0.9060 - val_AD_Recall: 0.4646 - val_CN_Recall: 0.9746 - val_MCI_Recall: 0.8503\n",
            "Epoch 45/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.8380 - Accuracy: 0.9408 - AD_Precision: 0.9311 - CN_Precision: 0.9347 - MCI_Precision: 0.9480 - AD_Recall: 0.9329 - CN_Recall: 0.9368 - MCI_Recall: 0.9460 - val_loss: 3.4709 - val_Accuracy: 0.8782 - val_AD_Precision: 0.8661 - val_CN_Precision: 0.9934 - val_MCI_Precision: 0.8414 - val_AD_Recall: 0.7638 - val_CN_Recall: 0.7614 - val_MCI_Recall: 0.9786\n",
            "Epoch 46/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 0.8222 - Accuracy: 0.9491 - AD_Precision: 0.9312 - CN_Precision: 0.9476 - MCI_Precision: 0.9564 - AD_Recall: 0.9349 - CN_Recall: 0.9520 - MCI_Recall: 0.9523 - val_loss: 2.6096 - val_Accuracy: 0.8883 - val_AD_Precision: 0.7278 - val_CN_Precision: 0.9441 - val_MCI_Precision: 0.9307 - val_AD_Recall: 0.9055 - val_CN_Recall: 0.8579 - val_MCI_Recall: 0.8984\n",
            "Epoch 47/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.7580 - Accuracy: 0.9527 - AD_Precision: 0.9312 - CN_Precision: 0.9545 - MCI_Precision: 0.9592 - AD_Recall: 0.9349 - CN_Recall: 0.9567 - MCI_Recall: 0.9566 - val_loss: 3.3186 - val_Accuracy: 0.8883 - val_AD_Precision: 0.8909 - val_CN_Precision: 0.9419 - val_MCI_Precision: 0.8654 - val_AD_Recall: 0.7717 - val_CN_Recall: 0.8223 - val_MCI_Recall: 0.9626\n",
            "Epoch 48/50\n",
            "698/698 [==============================] - 55s 78ms/step - loss: 1.1015 - Accuracy: 0.9383 - AD_Precision: 0.9118 - CN_Precision: 0.9380 - MCI_Precision: 0.9480 - AD_Recall: 0.9172 - CN_Recall: 0.9380 - MCI_Recall: 0.9460 - val_loss: 9.6621 - val_Accuracy: 0.7550 - val_AD_Precision: 0.9459 - val_CN_Precision: 0.9754 - val_MCI_Precision: 0.6920 - val_AD_Recall: 0.2756 - val_CN_Recall: 0.6041 - val_MCI_Recall: 0.9973\n",
            "Epoch 49/50\n",
            "698/698 [==============================] - 58s 83ms/step - loss: 1.0079 - Accuracy: 0.9430 - AD_Precision: 0.9191 - CN_Precision: 0.9426 - MCI_Precision: 0.9517 - AD_Recall: 0.9191 - CN_Recall: 0.9415 - MCI_Recall: 0.9523 - val_loss: 2.8285 - val_Accuracy: 0.9026 - val_AD_Precision: 0.9375 - val_CN_Precision: 0.8732 - val_MCI_Precision: 0.9100 - val_AD_Recall: 0.7087 - val_CN_Recall: 0.9442 - val_MCI_Recall: 0.9465\n",
            "Epoch 50/50\n",
            "698/698 [==============================] - 56s 80ms/step - loss: 1.2515 - Accuracy: 0.9347 - AD_Precision: 0.8978 - CN_Precision: 0.9346 - MCI_Precision: 0.9480 - AD_Recall: 0.9014 - CN_Recall: 0.9357 - MCI_Recall: 0.9460 - val_loss: 2.5600 - val_Accuracy: 0.8897 - val_AD_Precision: 0.8430 - val_CN_Precision: 0.9189 - val_MCI_Precision: 0.8903 - val_AD_Recall: 0.8031 - val_CN_Recall: 0.8629 - val_MCI_Recall: 0.9332\n"
          ]
        }
      ],
      "source": [
        "# learning_rate=10E-4\n",
        "savepath = '/content/drive/MyDrive/ADNI/resnet50_pretrained_10E-4/'\n",
        "lr=0.0001\n",
        "if not os.path.isdir(savepath):\n",
        "      os.makedirs(savepath)\n",
        "csvloger_callback = tf.keras.callbacks.CSVLogger(f'{savepath}history.csv', separator=',', append=False)\n",
        "\n",
        "modelcheckpoint_callback = tf.keras.callbacks.ModelCheckpoint(savepath+'resnet50_pretrained_10E-4_weights-epoch:{epoch:03d}-val_Accuracy:{val_Accuracy:.4f}.h5',monitor='val_Accuracy'\n",
        "    ,verbose=0,save_best_only=True,save_weights_only=True,mode='max',save_freq='epoch' , initial_value_threshold=0.85)\n",
        "\n",
        "OPT = tf.keras.optimizers.Adam(learning_rate=lr)\n",
        "model01 = resnet50_pretrained(OPT, LOOS,METRICS,image_Shape)\n",
        "history01 = model01.fit(training_image_datagen, epochs=EPOCH, validation_data=testing_image_datagen,callbacks=[csvloger_callback,modelcheckpoint_callback])\n",
        "model01 = resnet50_pretrained(OPT, LOOS,METRICS,image_Shape)\n",
        "model01.load_weights(glob.glob(f\"{savepath}resnet50_pretrained_10E-4_weights-epoch:*-val_Accuracy:{max(history01.history['val_Accuracy']):.4f}.h5\")[0])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 477
        },
        "id": "DjeRrTiwot5L",
        "outputId": "c0cb5b30-a9b7-46dc-f37c-e90e5db4e643"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "          AD       0.94      0.71      0.81       127\n",
            "          CN       0.87      0.94      0.91       197\n",
            "         MCI       0.91      0.95      0.93       374\n",
            "\n",
            "    accuracy                           0.90       698\n",
            "   macro avg       0.91      0.87      0.88       698\n",
            "weighted avg       0.90      0.90      0.90       698\n",
            "\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "p = model01.predict(testing_image_datagen)\n",
        "predicted_lab = p.argmax(axis=-1)\n",
        "true_lab = testing_image_datagen.classes\n",
        "plot_confusion_matrix(true_lab, predicted_lab, testing_image_datagen.class_indices.keys(),savepath)\n",
        "report = classification_report(true_lab, predicted_lab, target_names=testing_image_datagen.class_indices.keys())\n",
        "print(report)\n",
        "dump_text(report,f'{savepath}classification_report.txt')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9dvVuzvmIKaC"
      },
      "source": [
        "### learning_rate=10E-5"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LmLUs96kB7JQ",
        "outputId": "e31226a1-0487-4e49-aba2-8c3ccae5ebc9"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "698/698 [==============================] - 59s 79ms/step - loss: 7.0761 - Accuracy: 0.5707 - AD_Precision: 0.4087 - CN_Precision: 0.5239 - MCI_Precision: 0.6551 - AD_Recall: 0.3991 - CN_Recall: 0.5200 - MCI_Recall: 0.6602 - val_loss: 4.2220 - val_Accuracy: 0.4298 - val_AD_Precision: 0.2551 - val_CN_Precision: 0.7087 - val_MCI_Precision: 0.7655 - val_AD_Recall: 0.8898 - val_CN_Recall: 0.3706 - val_MCI_Recall: 0.2968\n",
            "Epoch 2/50\n",
            "698/698 [==============================] - 55s 78ms/step - loss: 3.4269 - Accuracy: 0.6052 - AD_Precision: 0.4687 - CN_Precision: 0.5794 - MCI_Precision: 0.6716 - AD_Recall: 0.4576 - CN_Recall: 0.5801 - MCI_Recall: 0.6692 - val_loss: 4.2318 - val_Accuracy: 0.4169 - val_AD_Precision: 0.2532 - val_CN_Precision: 0.9545 - val_MCI_Precision: 0.7260 - val_AD_Recall: 0.9291 - val_CN_Recall: 0.1066 - val_MCI_Recall: 0.4037\n",
            "Epoch 3/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 2.2227 - Accuracy: 0.6737 - AD_Precision: 0.5582 - CN_Precision: 0.6627 - MCI_Precision: 0.7264 - AD_Recall: 0.5582 - CN_Recall: 0.6550 - MCI_Recall: 0.7218 - val_loss: 1.2202 - val_Accuracy: 0.6977 - val_AD_Precision: 0.6475 - val_CN_Precision: 0.5953 - val_MCI_Precision: 0.8507 - val_AD_Recall: 0.6220 - val_CN_Recall: 0.9036 - val_MCI_Recall: 0.6096\n",
            "Epoch 4/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 1.5193 - Accuracy: 0.7250 - AD_Precision: 0.6333 - CN_Precision: 0.7015 - MCI_Precision: 0.7707 - AD_Recall: 0.6233 - CN_Recall: 0.6982 - MCI_Recall: 0.7701 - val_loss: 3.9003 - val_Accuracy: 0.7006 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.8221 - val_MCI_Precision: 0.6629 - val_AD_Recall: 0.0079 - val_CN_Recall: 0.6802 - val_MCI_Recall: 0.9465\n",
            "Epoch 5/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.9801 - Accuracy: 0.7691 - AD_Precision: 0.6918 - CN_Precision: 0.7485 - MCI_Precision: 0.8134 - AD_Recall: 0.6864 - CN_Recall: 0.7485 - MCI_Recall: 0.8066 - val_loss: 0.5219 - val_Accuracy: 0.8309 - val_AD_Precision: 0.7313 - val_CN_Precision: 0.8186 - val_MCI_Precision: 0.8817 - val_AD_Recall: 0.7717 - val_CN_Recall: 0.8477 - val_MCI_Recall: 0.8369\n",
            "Epoch 6/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.7666 - Accuracy: 0.7974 - AD_Precision: 0.7355 - CN_Precision: 0.7976 - MCI_Precision: 0.8240 - AD_Recall: 0.7022 - CN_Recall: 0.7836 - MCI_Recall: 0.8332 - val_loss: 0.6047 - val_Accuracy: 0.7880 - val_AD_Precision: 0.7917 - val_CN_Precision: 0.6790 - val_MCI_Precision: 0.9103 - val_AD_Recall: 0.7480 - val_CN_Recall: 0.9340 - val_MCI_Recall: 0.7059\n",
            "Epoch 7/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.5837 - Accuracy: 0.8222 - AD_Precision: 0.7951 - CN_Precision: 0.8050 - MCI_Precision: 0.8485 - AD_Recall: 0.7732 - CN_Recall: 0.7871 - MCI_Recall: 0.8556 - val_loss: 0.5647 - val_Accuracy: 0.8052 - val_AD_Precision: 0.5736 - val_CN_Precision: 0.9250 - val_MCI_Precision: 0.8939 - val_AD_Recall: 0.8898 - val_CN_Recall: 0.7513 - val_MCI_Recall: 0.7888\n",
            "Epoch 8/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.4750 - Accuracy: 0.8379 - AD_Precision: 0.8016 - CN_Precision: 0.8284 - MCI_Precision: 0.8619 - AD_Recall: 0.7732 - CN_Recall: 0.8187 - MCI_Recall: 0.8662 - val_loss: 1.0972 - val_Accuracy: 0.6074 - val_AD_Precision: 0.8500 - val_CN_Precision: 0.4495 - val_MCI_Precision: 0.9517 - val_AD_Recall: 0.6693 - val_CN_Recall: 0.9949 - val_MCI_Recall: 0.3690\n",
            "Epoch 9/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.3473 - Accuracy: 0.8770 - AD_Precision: 0.8360 - CN_Precision: 0.8808 - MCI_Precision: 0.9006 - AD_Recall: 0.8146 - CN_Recall: 0.8725 - MCI_Recall: 0.8956 - val_loss: 1.3592 - val_Accuracy: 0.6433 - val_AD_Precision: 0.9767 - val_CN_Precision: 1.0000 - val_MCI_Precision: 0.6023 - val_AD_Recall: 0.3307 - val_CN_Recall: 0.1675 - val_MCI_Recall: 1.0000\n",
            "Epoch 10/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.4206 - Accuracy: 0.8591 - AD_Precision: 0.8376 - CN_Precision: 0.8726 - MCI_Precision: 0.8673 - AD_Recall: 0.7830 - CN_Recall: 0.8491 - MCI_Recall: 0.8837 - val_loss: 0.6607 - val_Accuracy: 0.7722 - val_AD_Precision: 0.9811 - val_CN_Precision: 0.9739 - val_MCI_Precision: 0.7126 - val_AD_Recall: 0.4094 - val_CN_Recall: 0.5685 - val_MCI_Recall: 0.9947\n",
            "Epoch 11/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.3474 - Accuracy: 0.8706 - AD_Precision: 0.8562 - CN_Precision: 0.8681 - MCI_Precision: 0.8847 - AD_Recall: 0.8107 - CN_Recall: 0.8620 - MCI_Recall: 0.8872 - val_loss: 1.0286 - val_Accuracy: 0.6476 - val_AD_Precision: 0.8966 - val_CN_Precision: 0.4770 - val_MCI_Precision: 0.9296 - val_AD_Recall: 0.4094 - val_CN_Recall: 1.0000 - val_MCI_Recall: 0.5294\n",
            "Epoch 12/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.3123 - Accuracy: 0.8885 - AD_Precision: 0.8565 - CN_Precision: 0.8969 - MCI_Precision: 0.9009 - AD_Recall: 0.8008 - CN_Recall: 0.8854 - MCI_Recall: 0.9110 - val_loss: 1.5138 - val_Accuracy: 0.5229 - val_AD_Precision: 0.3056 - val_CN_Precision: 0.9783 - val_MCI_Precision: 0.8341 - val_AD_Recall: 0.9843 - val_CN_Recall: 0.2284 - val_MCI_Recall: 0.5107\n",
            "Epoch 13/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.2825 - Accuracy: 0.9014 - AD_Precision: 0.8758 - CN_Precision: 0.9040 - MCI_Precision: 0.9184 - AD_Recall: 0.8481 - CN_Recall: 0.8807 - MCI_Recall: 0.9229 - val_loss: 0.8214 - val_Accuracy: 0.7264 - val_AD_Precision: 0.9048 - val_CN_Precision: 0.9683 - val_MCI_Precision: 0.6740 - val_AD_Recall: 0.5984 - val_CN_Recall: 0.3096 - val_MCI_Recall: 0.9893\n",
            "Epoch 14/50\n",
            "698/698 [==============================] - 55s 78ms/step - loss: 0.2675 - Accuracy: 0.8975 - AD_Precision: 0.8724 - CN_Precision: 0.9031 - MCI_Precision: 0.9107 - AD_Recall: 0.8363 - CN_Recall: 0.8936 - MCI_Recall: 0.9152 - val_loss: 0.6780 - val_Accuracy: 0.7335 - val_AD_Precision: 0.6629 - val_CN_Precision: 0.6481 - val_MCI_Precision: 0.9797 - val_AD_Recall: 0.9291 - val_CN_Recall: 0.9442 - val_MCI_Recall: 0.5160\n",
            "Epoch 15/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.2340 - Accuracy: 0.9139 - AD_Precision: 0.8697 - CN_Precision: 0.9342 - MCI_Precision: 0.9250 - AD_Recall: 0.8422 - CN_Recall: 0.9135 - MCI_Recall: 0.9334 - val_loss: 0.5015 - val_Accuracy: 0.8166 - val_AD_Precision: 0.9487 - val_CN_Precision: 0.7030 - val_MCI_Precision: 0.8964 - val_AD_Recall: 0.5827 - val_CN_Recall: 0.9492 - val_MCI_Recall: 0.8102\n",
            "Epoch 16/50\n",
            "698/698 [==============================] - 58s 83ms/step - loss: 0.2183 - Accuracy: 0.9186 - AD_Precision: 0.9115 - CN_Precision: 0.9265 - MCI_Precision: 0.9205 - AD_Recall: 0.8738 - CN_Recall: 0.8994 - MCI_Recall: 0.9334 - val_loss: 0.3533 - val_Accuracy: 0.8983 - val_AD_Precision: 0.9320 - val_CN_Precision: 0.9153 - val_MCI_Precision: 0.8856 - val_AD_Recall: 0.7559 - val_CN_Recall: 0.8782 - val_MCI_Recall: 0.9519\n",
            "Epoch 17/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 0.2465 - Accuracy: 0.9093 - AD_Precision: 0.8966 - CN_Precision: 0.9099 - MCI_Precision: 0.9186 - AD_Recall: 0.8718 - CN_Recall: 0.9099 - MCI_Recall: 0.9173 - val_loss: 0.3400 - val_Accuracy: 0.8954 - val_AD_Precision: 0.9485 - val_CN_Precision: 0.8732 - val_MCI_Precision: 0.8934 - val_AD_Recall: 0.7244 - val_CN_Recall: 0.9086 - val_MCI_Recall: 0.9412\n",
            "Epoch 18/50\n",
            "698/698 [==============================] - 57s 82ms/step - loss: 0.2050 - Accuracy: 0.9200 - AD_Precision: 0.9037 - CN_Precision: 0.9277 - MCI_Precision: 0.9281 - AD_Recall: 0.8698 - CN_Recall: 0.9158 - MCI_Recall: 0.9313 - val_loss: 0.3175 - val_Accuracy: 0.9083 - val_AD_Precision: 0.8824 - val_CN_Precision: 0.9086 - val_MCI_Precision: 0.9158 - val_AD_Recall: 0.8268 - val_CN_Recall: 0.9086 - val_MCI_Recall: 0.9305\n",
            "Epoch 19/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.2105 - Accuracy: 0.9233 - AD_Precision: 0.9045 - CN_Precision: 0.9304 - MCI_Precision: 0.9309 - AD_Recall: 0.8777 - CN_Recall: 0.9228 - MCI_Recall: 0.9341 - val_loss: 0.4573 - val_Accuracy: 0.8596 - val_AD_Precision: 0.9512 - val_CN_Precision: 0.8691 - val_MCI_Precision: 0.8534 - val_AD_Recall: 0.6142 - val_CN_Recall: 0.8426 - val_MCI_Recall: 0.9492\n",
            "Epoch 20/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.2035 - Accuracy: 0.9226 - AD_Precision: 0.9165 - CN_Precision: 0.9264 - MCI_Precision: 0.9315 - AD_Recall: 0.8876 - CN_Recall: 0.9123 - MCI_Recall: 0.9341 - val_loss: 0.4440 - val_Accuracy: 0.8553 - val_AD_Precision: 0.9318 - val_CN_Precision: 0.9732 - val_MCI_Precision: 0.8053 - val_AD_Recall: 0.6457 - val_CN_Recall: 0.7360 - val_MCI_Recall: 0.9840\n",
            "Epoch 21/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.2211 - Accuracy: 0.9211 - AD_Precision: 0.9270 - CN_Precision: 0.9190 - MCI_Precision: 0.9296 - AD_Recall: 0.9014 - CN_Recall: 0.9018 - MCI_Recall: 0.9348 - val_loss: 1.2752 - val_Accuracy: 0.7106 - val_AD_Precision: 0.9057 - val_CN_Precision: 1.0000 - val_MCI_Precision: 0.6578 - val_AD_Recall: 0.7559 - val_CN_Recall: 0.1472 - val_MCI_Recall: 0.9866\n",
            "Epoch 22/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.1629 - Accuracy: 0.9426 - AD_Precision: 0.9328 - CN_Precision: 0.9417 - MCI_Precision: 0.9502 - AD_Recall: 0.9034 - CN_Recall: 0.9439 - MCI_Recall: 0.9502 - val_loss: 0.3845 - val_Accuracy: 0.8854 - val_AD_Precision: 0.8413 - val_CN_Precision: 0.8341 - val_MCI_Precision: 0.9415 - val_AD_Recall: 0.8346 - val_CN_Recall: 0.9442 - val_MCI_Recall: 0.8610\n",
            "Epoch 23/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.1756 - Accuracy: 0.9416 - AD_Precision: 0.9231 - CN_Precision: 0.9492 - MCI_Precision: 0.9458 - AD_Recall: 0.8994 - CN_Recall: 0.9404 - MCI_Recall: 0.9530 - val_loss: 0.5262 - val_Accuracy: 0.8438 - val_AD_Precision: 0.9615 - val_CN_Precision: 0.8531 - val_MCI_Precision: 0.8318 - val_AD_Recall: 0.3937 - val_CN_Recall: 0.9137 - val_MCI_Recall: 0.9519\n",
            "Epoch 24/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.1415 - Accuracy: 0.9505 - AD_Precision: 0.9437 - CN_Precision: 0.9522 - MCI_Precision: 0.9544 - AD_Recall: 0.9250 - CN_Recall: 0.9556 - MCI_Recall: 0.9530 - val_loss: 0.3405 - val_Accuracy: 0.8868 - val_AD_Precision: 0.7708 - val_CN_Precision: 0.8846 - val_MCI_Precision: 0.9493 - val_AD_Recall: 0.8740 - val_CN_Recall: 0.9340 - val_MCI_Recall: 0.8503\n",
            "Epoch 25/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.1285 - Accuracy: 0.9548 - AD_Precision: 0.9399 - CN_Precision: 0.9612 - MCI_Precision: 0.9588 - AD_Recall: 0.9250 - CN_Recall: 0.9556 - MCI_Recall: 0.9629 - val_loss: 0.3486 - val_Accuracy: 0.8897 - val_AD_Precision: 0.8942 - val_CN_Precision: 0.8296 - val_MCI_Precision: 0.9290 - val_AD_Recall: 0.7323 - val_CN_Recall: 0.9391 - val_MCI_Recall: 0.9091\n",
            "Epoch 26/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.1558 - Accuracy: 0.9412 - AD_Precision: 0.9357 - CN_Precision: 0.9364 - MCI_Precision: 0.9490 - AD_Recall: 0.9191 - CN_Recall: 0.9298 - MCI_Recall: 0.9516 - val_loss: 0.3454 - val_Accuracy: 0.8983 - val_AD_Precision: 0.8231 - val_CN_Precision: 0.9448 - val_MCI_Precision: 0.9062 - val_AD_Recall: 0.8425 - val_CN_Recall: 0.8680 - val_MCI_Recall: 0.9305\n",
            "Epoch 27/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.1479 - Accuracy: 0.9426 - AD_Precision: 0.9178 - CN_Precision: 0.9542 - MCI_Precision: 0.9475 - AD_Recall: 0.9034 - CN_Recall: 0.9497 - MCI_Recall: 0.9488 - val_loss: 0.5398 - val_Accuracy: 0.8610 - val_AD_Precision: 0.9692 - val_CN_Precision: 0.8230 - val_MCI_Precision: 0.8688 - val_AD_Recall: 0.4961 - val_CN_Recall: 0.9442 - val_MCI_Recall: 0.9385\n",
            "Epoch 28/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.1075 - Accuracy: 0.9652 - AD_Precision: 0.9718 - CN_Precision: 0.9578 - MCI_Precision: 0.9678 - AD_Recall: 0.9527 - CN_Recall: 0.9567 - MCI_Recall: 0.9699 - val_loss: 0.5493 - val_Accuracy: 0.8424 - val_AD_Precision: 0.9423 - val_CN_Precision: 0.9754 - val_MCI_Precision: 0.7872 - val_AD_Recall: 0.7717 - val_CN_Recall: 0.6041 - val_MCI_Recall: 0.9893\n",
            "Epoch 29/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.1285 - Accuracy: 0.9570 - AD_Precision: 0.9287 - CN_Precision: 0.9659 - MCI_Precision: 0.9629 - AD_Recall: 0.9250 - CN_Recall: 0.9614 - MCI_Recall: 0.9643 - val_loss: 0.3369 - val_Accuracy: 0.9026 - val_AD_Precision: 0.9200 - val_CN_Precision: 0.9415 - val_MCI_Precision: 0.8867 - val_AD_Recall: 0.7244 - val_CN_Recall: 0.8985 - val_MCI_Recall: 0.9626\n",
            "Epoch 30/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.1133 - Accuracy: 0.9598 - AD_Precision: 0.9461 - CN_Precision: 0.9591 - MCI_Precision: 0.9657 - AD_Recall: 0.9349 - CN_Recall: 0.9602 - MCI_Recall: 0.9657 - val_loss: 0.3820 - val_Accuracy: 0.8926 - val_AD_Precision: 0.9556 - val_CN_Precision: 0.8756 - val_MCI_Precision: 0.8886 - val_AD_Recall: 0.6772 - val_CN_Recall: 0.9289 - val_MCI_Recall: 0.9385\n",
            "Epoch 31/50\n",
            "698/698 [==============================] - 53s 77ms/step - loss: 0.1196 - Accuracy: 0.9570 - AD_Precision: 0.9401 - CN_Precision: 0.9534 - MCI_Precision: 0.9669 - AD_Recall: 0.9290 - CN_Recall: 0.9579 - MCI_Recall: 0.9636 - val_loss: 0.4647 - val_Accuracy: 0.8825 - val_AD_Precision: 0.9560 - val_CN_Precision: 0.9812 - val_MCI_Precision: 0.8348 - val_AD_Recall: 0.6850 - val_CN_Recall: 0.7970 - val_MCI_Recall: 0.9866\n",
            "Epoch 32/50\n",
            "698/698 [==============================] - 53s 76ms/step - loss: 0.0945 - Accuracy: 0.9652 - AD_Precision: 0.9563 - CN_Precision: 0.9636 - MCI_Precision: 0.9713 - AD_Recall: 0.9487 - CN_Recall: 0.9602 - MCI_Recall: 0.9720 - val_loss: 0.3983 - val_Accuracy: 0.8940 - val_AD_Precision: 0.8015 - val_CN_Precision: 0.9058 - val_MCI_Precision: 0.9241 - val_AD_Recall: 0.8583 - val_CN_Recall: 0.8782 - val_MCI_Recall: 0.9118\n",
            "Epoch 33/50\n",
            "698/698 [==============================] - 53s 76ms/step - loss: 0.1396 - Accuracy: 0.9512 - AD_Precision: 0.9440 - CN_Precision: 0.9484 - MCI_Precision: 0.9574 - AD_Recall: 0.9310 - CN_Recall: 0.9450 - MCI_Recall: 0.9615 - val_loss: 0.8364 - val_Accuracy: 0.7521 - val_AD_Precision: 0.5535 - val_CN_Precision: 0.7562 - val_MCI_Precision: 0.9682 - val_AD_Recall: 0.9370 - val_CN_Recall: 0.9289 - val_MCI_Recall: 0.5695\n",
            "Epoch 34/50\n",
            "698/698 [==============================] - 53s 77ms/step - loss: 0.1163 - Accuracy: 0.9570 - AD_Precision: 0.9365 - CN_Precision: 0.9625 - MCI_Precision: 0.9636 - AD_Recall: 0.9310 - CN_Recall: 0.9602 - MCI_Recall: 0.9636 - val_loss: 0.5615 - val_Accuracy: 0.8668 - val_AD_Precision: 0.9481 - val_CN_Precision: 0.9371 - val_MCI_Precision: 0.8299 - val_AD_Recall: 0.5748 - val_CN_Recall: 0.8325 - val_MCI_Recall: 0.9786\n",
            "Epoch 35/50\n",
            "698/698 [==============================] - 58s 83ms/step - loss: 0.0807 - Accuracy: 0.9753 - AD_Precision: 0.9662 - CN_Precision: 0.9765 - MCI_Precision: 0.9797 - AD_Recall: 0.9586 - CN_Recall: 0.9708 - MCI_Recall: 0.9818 - val_loss: 0.8583 - val_Accuracy: 0.8238 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.8980 - val_MCI_Precision: 0.7876 - val_AD_Recall: 0.2441 - val_CN_Recall: 0.8934 - val_MCI_Recall: 0.9813\n",
            "Epoch 36/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.1232 - Accuracy: 0.9534 - AD_Precision: 0.9384 - CN_Precision: 0.9610 - MCI_Precision: 0.9573 - AD_Recall: 0.9310 - CN_Recall: 0.9509 - MCI_Recall: 0.9594 - val_loss: 0.4486 - val_Accuracy: 0.8926 - val_AD_Precision: 0.9167 - val_CN_Precision: 0.9263 - val_MCI_Precision: 0.8821 - val_AD_Recall: 0.6929 - val_CN_Recall: 0.8934 - val_MCI_Recall: 0.9599\n",
            "Epoch 37/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.0848 - Accuracy: 0.9674 - AD_Precision: 0.9618 - CN_Precision: 0.9728 - MCI_Precision: 0.9674 - AD_Recall: 0.9448 - CN_Recall: 0.9614 - MCI_Recall: 0.9769 - val_loss: 0.4246 - val_Accuracy: 0.8897 - val_AD_Precision: 0.8899 - val_CN_Precision: 0.8217 - val_MCI_Precision: 0.9435 - val_AD_Recall: 0.7638 - val_CN_Recall: 0.9594 - val_MCI_Recall: 0.8930\n",
            "Epoch 38/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.1214 - Accuracy: 0.9552 - AD_Precision: 0.9398 - CN_Precision: 0.9567 - MCI_Precision: 0.9615 - AD_Recall: 0.9231 - CN_Recall: 0.9556 - MCI_Recall: 0.9629 - val_loss: 0.4195 - val_Accuracy: 0.8926 - val_AD_Precision: 0.8168 - val_CN_Precision: 0.8714 - val_MCI_Precision: 0.9432 - val_AD_Recall: 0.8425 - val_CN_Recall: 0.9289 - val_MCI_Recall: 0.8877\n",
            "Epoch 39/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.0899 - Accuracy: 0.9688 - AD_Precision: 0.9683 - CN_Precision: 0.9648 - MCI_Precision: 0.9720 - AD_Recall: 0.9645 - CN_Recall: 0.9626 - MCI_Recall: 0.9741 - val_loss: 0.5110 - val_Accuracy: 0.8782 - val_AD_Precision: 0.9348 - val_CN_Precision: 0.9753 - val_MCI_Precision: 0.8326 - val_AD_Recall: 0.6772 - val_CN_Recall: 0.8020 - val_MCI_Recall: 0.9840\n",
            "Epoch 40/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.0839 - Accuracy: 0.9735 - AD_Precision: 0.9626 - CN_Precision: 0.9730 - MCI_Precision: 0.9790 - AD_Recall: 0.9645 - CN_Recall: 0.9696 - MCI_Recall: 0.9783 - val_loss: 1.1929 - val_Accuracy: 0.7034 - val_AD_Precision: 0.7255 - val_CN_Precision: 0.5496 - val_MCI_Precision: 0.9889 - val_AD_Recall: 0.8740 - val_CN_Recall: 0.9848 - val_MCI_Recall: 0.4759\n",
            "Epoch 41/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.0850 - Accuracy: 0.9684 - AD_Precision: 0.9742 - CN_Precision: 0.9592 - MCI_Precision: 0.9726 - AD_Recall: 0.9665 - CN_Recall: 0.9614 - MCI_Recall: 0.9713 - val_loss: 0.5498 - val_Accuracy: 0.8639 - val_AD_Precision: 0.9595 - val_CN_Precision: 0.9483 - val_MCI_Precision: 0.8156 - val_AD_Recall: 0.5591 - val_CN_Recall: 0.8376 - val_MCI_Recall: 0.9813\n",
            "Epoch 42/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.1016 - Accuracy: 0.9667 - AD_Precision: 0.9582 - CN_Precision: 0.9717 - MCI_Precision: 0.9686 - AD_Recall: 0.9487 - CN_Recall: 0.9626 - MCI_Recall: 0.9741 - val_loss: 0.8825 - val_Accuracy: 0.7966 - val_AD_Precision: 0.9688 - val_CN_Precision: 0.6361 - val_MCI_Precision: 0.9169 - val_AD_Recall: 0.4882 - val_CN_Recall: 0.9848 - val_MCI_Recall: 0.7968\n",
            "Epoch 43/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.0725 - Accuracy: 0.9763 - AD_Precision: 0.9641 - CN_Precision: 0.9813 - MCI_Precision: 0.9783 - AD_Recall: 0.9546 - CN_Recall: 0.9813 - MCI_Recall: 0.9797 - val_loss: 0.4465 - val_Accuracy: 0.8897 - val_AD_Precision: 0.7651 - val_CN_Precision: 0.8815 - val_MCI_Precision: 0.9580 - val_AD_Recall: 0.8976 - val_CN_Recall: 0.9442 - val_MCI_Recall: 0.8529\n",
            "Epoch 44/50\n",
            "698/698 [==============================] - 58s 83ms/step - loss: 0.0921 - Accuracy: 0.9681 - AD_Precision: 0.9564 - CN_Precision: 0.9670 - MCI_Precision: 0.9742 - AD_Recall: 0.9527 - CN_Recall: 0.9602 - MCI_Recall: 0.9776 - val_loss: 0.3412 - val_Accuracy: 0.9112 - val_AD_Precision: 0.8710 - val_CN_Precision: 0.9086 - val_MCI_Precision: 0.9253 - val_AD_Recall: 0.8504 - val_CN_Recall: 0.9086 - val_MCI_Recall: 0.9278\n",
            "Epoch 45/50\n",
            "698/698 [==============================] - 56s 80ms/step - loss: 0.0759 - Accuracy: 0.9720 - AD_Precision: 0.9585 - CN_Precision: 0.9765 - MCI_Precision: 0.9748 - AD_Recall: 0.9566 - CN_Recall: 0.9719 - MCI_Recall: 0.9769 - val_loss: 0.5193 - val_Accuracy: 0.8897 - val_AD_Precision: 0.7755 - val_CN_Precision: 0.9806 - val_MCI_Precision: 0.9005 - val_AD_Recall: 0.8976 - val_CN_Recall: 0.7716 - val_MCI_Recall: 0.9439\n",
            "Epoch 46/50\n",
            "698/698 [==============================] - 56s 80ms/step - loss: 0.0872 - Accuracy: 0.9663 - AD_Precision: 0.9701 - CN_Precision: 0.9626 - MCI_Precision: 0.9685 - AD_Recall: 0.9586 - CN_Recall: 0.9626 - MCI_Recall: 0.9699 - val_loss: 0.5312 - val_Accuracy: 0.8868 - val_AD_Precision: 0.9740 - val_CN_Precision: 0.9100 - val_MCI_Precision: 0.8599 - val_AD_Recall: 0.5906 - val_CN_Recall: 0.9239 - val_MCI_Recall: 0.9679\n",
            "Epoch 47/50\n",
            "698/698 [==============================] - 56s 80ms/step - loss: 0.0593 - Accuracy: 0.9817 - AD_Precision: 0.9782 - CN_Precision: 0.9825 - MCI_Precision: 0.9825 - AD_Recall: 0.9744 - CN_Recall: 0.9836 - MCI_Recall: 0.9818 - val_loss: 0.6105 - val_Accuracy: 0.8668 - val_AD_Precision: 0.9714 - val_CN_Precision: 0.9551 - val_MCI_Precision: 0.8210 - val_AD_Recall: 0.5354 - val_CN_Recall: 0.8629 - val_MCI_Recall: 0.9813\n",
            "Epoch 48/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 0.1227 - Accuracy: 0.9606 - AD_Precision: 0.9564 - CN_Precision: 0.9543 - MCI_Precision: 0.9658 - AD_Recall: 0.9527 - CN_Recall: 0.9520 - MCI_Recall: 0.9685 - val_loss: 0.5146 - val_Accuracy: 0.8983 - val_AD_Precision: 0.9300 - val_CN_Precision: 0.9709 - val_MCI_Precision: 0.8656 - val_AD_Recall: 0.7323 - val_CN_Recall: 0.8477 - val_MCI_Recall: 0.9813\n",
            "Epoch 49/50\n",
            "698/698 [==============================] - 56s 80ms/step - loss: 0.0875 - Accuracy: 0.9681 - AD_Precision: 0.9529 - CN_Precision: 0.9706 - MCI_Precision: 0.9720 - AD_Recall: 0.9586 - CN_Recall: 0.9649 - MCI_Recall: 0.9720 - val_loss: 0.3847 - val_Accuracy: 0.9011 - val_AD_Precision: 0.8571 - val_CN_Precision: 0.8714 - val_MCI_Precision: 0.9415 - val_AD_Recall: 0.8504 - val_CN_Recall: 0.9289 - val_MCI_Recall: 0.9037\n",
            "Epoch 50/50\n",
            "698/698 [==============================] - 59s 85ms/step - loss: 0.0802 - Accuracy: 0.9717 - AD_Precision: 0.9741 - CN_Precision: 0.9639 - MCI_Precision: 0.9776 - AD_Recall: 0.9625 - CN_Recall: 0.9673 - MCI_Recall: 0.9769 - val_loss: 0.3780 - val_Accuracy: 0.9198 - val_AD_Precision: 0.9076 - val_CN_Precision: 0.9459 - val_MCI_Precision: 0.9203 - val_AD_Recall: 0.8504 - val_CN_Recall: 0.8883 - val_MCI_Recall: 0.9572\n"
          ]
        }
      ],
      "source": [
        "# learning_rate=10E-5\n",
        "savepath = '/content/drive/MyDrive/ADNI/resnet50_pretrained_10E-5/'\n",
        "lr=0.00001\n",
        "if not os.path.isdir(savepath):\n",
        "      os.makedirs(savepath)\n",
        "csvloger_callback = tf.keras.callbacks.CSVLogger(f'{savepath}history.csv', separator=',', append=False)\n",
        "\n",
        "modelcheckpoint_callback = tf.keras.callbacks.ModelCheckpoint(savepath+'resnet50_pretrained_10E-5_weights-epoch:{epoch:03d}-val_Accuracy:{val_Accuracy:.4f}.h5',monitor='val_Accuracy'\n",
        "    ,verbose=0,save_best_only=True,save_weights_only=True,mode='max',save_freq='epoch' , initial_value_threshold=0.85)\n",
        "\n",
        "OPT = tf.keras.optimizers.Adam(learning_rate=lr)\n",
        "model02 = resnet50_pretrained(OPT, LOOS,METRICS,image_Shape)\n",
        "history02 = model02.fit(training_image_datagen, epochs=EPOCH, validation_data=testing_image_datagen,callbacks=[csvloger_callback,modelcheckpoint_callback])\n",
        "model02 = resnet50_pretrained(OPT, LOOS,METRICS,image_Shape)\n",
        "model02.load_weights(glob.glob(f\"{savepath}resnet50_pretrained_10E-5_weights-epoch:*-val_Accuracy:{max(history02.history['val_Accuracy']):.4f}.h5\")[0])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 477
        },
        "id": "zXvq2LTMB9ou",
        "outputId": "be365ad3-b8be-4778-daa8-e9c55fd94fce"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "          AD       0.90      0.85      0.87       127\n",
            "          CN       0.94      0.89      0.91       197\n",
            "         MCI       0.92      0.96      0.94       374\n",
            "\n",
            "    accuracy                           0.92       698\n",
            "   macro avg       0.92      0.90      0.91       698\n",
            "weighted avg       0.92      0.92      0.92       698\n",
            "\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "p = model02.predict(testing_image_datagen)\n",
        "predicted_lab = p.argmax(axis=-1)\n",
        "true_lab = testing_image_datagen.classes\n",
        "plot_confusion_matrix(true_lab, predicted_lab, testing_image_datagen.class_indices.keys(),savepath)\n",
        "report = classification_report(true_lab, predicted_lab, target_names=testing_image_datagen.class_indices.keys())\n",
        "print(report)\n",
        "dump_text(report,f'{savepath}classification_report.txt')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jIcC7H-pI82Y"
      },
      "source": [
        "### learning_rate=10E-6"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wjUS7A9VCIWX",
        "outputId": "42b99232-404c-454a-e72e-7164c1dfede8"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "698/698 [==============================] - 61s 80ms/step - loss: 3.5724 - Accuracy: 0.5139 - AD_Precision: 0.3558 - CN_Precision: 0.4495 - MCI_Precision: 0.6105 - AD_Recall: 0.3407 - CN_Recall: 0.4354 - MCI_Recall: 0.6135 - val_loss: 2.0704 - val_Accuracy: 0.4585 - val_AD_Precision: 0.6000 - val_CN_Precision: 0.3526 - val_MCI_Precision: 0.7105 - val_AD_Recall: 0.1181 - val_CN_Recall: 0.8376 - val_MCI_Recall: 0.3610\n",
            "Epoch 2/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 2.8304 - Accuracy: 0.5009 - AD_Precision: 0.3221 - CN_Precision: 0.4561 - MCI_Precision: 0.5924 - AD_Recall: 0.3018 - CN_Recall: 0.4433 - MCI_Recall: 0.5929 - val_loss: 1.2083 - val_Accuracy: 0.6218 - val_AD_Precision: 0.5625 - val_CN_Precision: 0.5273 - val_MCI_Precision: 0.6997 - val_AD_Recall: 0.1417 - val_CN_Recall: 0.6853 - val_MCI_Recall: 0.7166\n",
            "Epoch 3/50\n",
            "698/698 [==============================] - 55s 78ms/step - loss: 2.3611 - Accuracy: 0.5489 - AD_Precision: 0.4044 - CN_Precision: 0.5048 - MCI_Precision: 0.6366 - AD_Recall: 0.3964 - CN_Recall: 0.4936 - MCI_Recall: 0.6272 - val_loss: 1.0962 - val_Accuracy: 0.6662 - val_AD_Precision: 0.5060 - val_CN_Precision: 0.7068 - val_MCI_Precision: 0.7500 - val_AD_Recall: 0.6693 - val_CN_Recall: 0.4772 - val_MCI_Recall: 0.7380\n",
            "Epoch 4/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 2.1233 - Accuracy: 0.5859 - AD_Precision: 0.4590 - CN_Precision: 0.5442 - MCI_Precision: 0.6579 - AD_Recall: 0.4418 - CN_Recall: 0.5251 - MCI_Recall: 0.6643 - val_loss: 1.1979 - val_Accuracy: 0.6676 - val_AD_Precision: 0.7778 - val_CN_Precision: 0.7808 - val_MCI_Precision: 0.6420 - val_AD_Recall: 0.3858 - val_CN_Recall: 0.2893 - val_MCI_Recall: 0.9492\n",
            "Epoch 5/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 1.9165 - Accuracy: 0.6207 - AD_Precision: 0.4979 - CN_Precision: 0.5875 - MCI_Precision: 0.6896 - AD_Recall: 0.4773 - CN_Recall: 0.5813 - MCI_Recall: 0.6882 - val_loss: 1.1493 - val_Accuracy: 0.6934 - val_AD_Precision: 0.9118 - val_CN_Precision: 0.7949 - val_MCI_Precision: 0.6587 - val_AD_Recall: 0.2441 - val_CN_Recall: 0.4721 - val_MCI_Recall: 0.9545\n",
            "Epoch 6/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 1.5849 - Accuracy: 0.6662 - AD_Precision: 0.5449 - CN_Precision: 0.6409 - MCI_Precision: 0.7298 - AD_Recall: 0.5266 - CN_Recall: 0.6199 - MCI_Recall: 0.7288 - val_loss: 1.0617 - val_Accuracy: 0.7335 - val_AD_Precision: 0.9118 - val_CN_Precision: 0.8433 - val_MCI_Precision: 0.6927 - val_AD_Recall: 0.2441 - val_CN_Recall: 0.5736 - val_MCI_Recall: 0.9706\n",
            "Epoch 7/50\n",
            "698/698 [==============================] - 55s 78ms/step - loss: 1.3874 - Accuracy: 0.7006 - AD_Precision: 0.5904 - CN_Precision: 0.6972 - MCI_Precision: 0.7484 - AD_Recall: 0.5799 - CN_Recall: 0.6760 - MCI_Recall: 0.7505 - val_loss: 0.7457 - val_Accuracy: 0.7564 - val_AD_Precision: 0.5799 - val_CN_Precision: 0.9252 - val_MCI_Precision: 0.8020 - val_AD_Recall: 0.7717 - val_CN_Recall: 0.5025 - val_MCI_Recall: 0.8770\n",
            "Epoch 8/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 1.2133 - Accuracy: 0.7261 - AD_Precision: 0.6410 - CN_Precision: 0.7138 - MCI_Precision: 0.7684 - AD_Recall: 0.6233 - CN_Recall: 0.6971 - MCI_Recall: 0.7765 - val_loss: 1.6630 - val_Accuracy: 0.7278 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.7290 - val_MCI_Precision: 0.7244 - val_AD_Recall: 0.0394 - val_CN_Recall: 0.7919 - val_MCI_Recall: 0.9278\n",
            "Epoch 9/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 1.1789 - Accuracy: 0.7354 - AD_Precision: 0.6735 - CN_Precision: 0.6976 - MCI_Precision: 0.7827 - AD_Recall: 0.6509 - CN_Recall: 0.7041 - MCI_Recall: 0.7800 - val_loss: 0.5263 - val_Accuracy: 0.8338 - val_AD_Precision: 0.6923 - val_CN_Precision: 0.9114 - val_MCI_Precision: 0.8679 - val_AD_Recall: 0.7795 - val_CN_Recall: 0.7310 - val_MCI_Recall: 0.8957\n",
            "Epoch 10/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 1.0005 - Accuracy: 0.7479 - AD_Precision: 0.6747 - CN_Precision: 0.7291 - MCI_Precision: 0.7932 - AD_Recall: 0.6627 - CN_Recall: 0.7240 - MCI_Recall: 0.7849 - val_loss: 0.8874 - val_Accuracy: 0.7249 - val_AD_Precision: 0.4412 - val_CN_Precision: 0.9097 - val_MCI_Precision: 0.9197 - val_AD_Recall: 0.9449 - val_CN_Recall: 0.6650 - val_MCI_Recall: 0.6738\n",
            "Epoch 11/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.9311 - Accuracy: 0.7784 - AD_Precision: 0.7212 - CN_Precision: 0.7706 - MCI_Precision: 0.8082 - AD_Recall: 0.7041 - CN_Recall: 0.7544 - MCI_Recall: 0.8150 - val_loss: 0.5933 - val_Accuracy: 0.8195 - val_AD_Precision: 0.7734 - val_CN_Precision: 0.7206 - val_MCI_Precision: 0.9216 - val_AD_Recall: 0.7795 - val_CN_Recall: 0.9036 - val_MCI_Recall: 0.7861\n",
            "Epoch 12/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.7626 - Accuracy: 0.8032 - AD_Precision: 0.7455 - CN_Precision: 0.7893 - MCI_Precision: 0.8365 - AD_Recall: 0.7278 - CN_Recall: 0.7801 - MCI_Recall: 0.8353 - val_loss: 0.5779 - val_Accuracy: 0.8195 - val_AD_Precision: 0.6750 - val_CN_Precision: 0.9124 - val_MCI_Precision: 0.8553 - val_AD_Recall: 0.8504 - val_CN_Recall: 0.6345 - val_MCI_Recall: 0.9011\n",
            "Epoch 13/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.7024 - Accuracy: 0.8143 - AD_Precision: 0.7480 - CN_Precision: 0.8067 - MCI_Precision: 0.8510 - AD_Recall: 0.7377 - CN_Recall: 0.7906 - MCI_Recall: 0.8486 - val_loss: 0.6401 - val_Accuracy: 0.8367 - val_AD_Precision: 0.9268 - val_CN_Precision: 0.7500 - val_MCI_Precision: 0.8859 - val_AD_Recall: 0.5984 - val_CN_Recall: 0.9137 - val_MCI_Recall: 0.8717\n",
            "Epoch 14/50\n",
            "698/698 [==============================] - 55s 78ms/step - loss: 0.7266 - Accuracy: 0.8136 - AD_Precision: 0.7419 - CN_Precision: 0.8101 - MCI_Precision: 0.8484 - AD_Recall: 0.7258 - CN_Recall: 0.8035 - MCI_Recall: 0.8472 - val_loss: 0.7585 - val_Accuracy: 0.8123 - val_AD_Precision: 0.9492 - val_CN_Precision: 0.7023 - val_MCI_Precision: 0.8774 - val_AD_Recall: 0.4409 - val_CN_Recall: 0.9340 - val_MCI_Recall: 0.8610\n",
            "Epoch 15/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.6092 - Accuracy: 0.8430 - AD_Precision: 0.8000 - CN_Precision: 0.8385 - MCI_Precision: 0.8638 - AD_Recall: 0.7811 - CN_Recall: 0.8257 - MCI_Recall: 0.8711 - val_loss: 0.7592 - val_Accuracy: 0.8309 - val_AD_Precision: 0.9804 - val_CN_Precision: 0.7902 - val_MCI_Precision: 0.8389 - val_AD_Recall: 0.3937 - val_CN_Recall: 0.8985 - val_MCI_Recall: 0.9332\n",
            "Epoch 16/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.5344 - Accuracy: 0.8559 - AD_Precision: 0.8204 - CN_Precision: 0.8462 - MCI_Precision: 0.8800 - AD_Recall: 0.8107 - CN_Recall: 0.8491 - MCI_Recall: 0.8739 - val_loss: 0.7809 - val_Accuracy: 0.8223 - val_AD_Precision: 0.9437 - val_CN_Precision: 0.9448 - val_MCI_Precision: 0.7715 - val_AD_Recall: 0.5276 - val_CN_Recall: 0.6954 - val_MCI_Recall: 0.9840\n",
            "Epoch 17/50\n",
            "698/698 [==============================] - 58s 83ms/step - loss: 0.4769 - Accuracy: 0.8627 - AD_Precision: 0.8100 - CN_Precision: 0.8560 - MCI_Precision: 0.8890 - AD_Recall: 0.7988 - CN_Recall: 0.8480 - MCI_Recall: 0.8921 - val_loss: 0.5961 - val_Accuracy: 0.8668 - val_AD_Precision: 0.9091 - val_CN_Precision: 0.9310 - val_MCI_Precision: 0.8364 - val_AD_Recall: 0.6299 - val_CN_Recall: 0.8223 - val_MCI_Recall: 0.9706\n",
            "Epoch 18/50\n",
            "698/698 [==============================] - 58s 83ms/step - loss: 0.4712 - Accuracy: 0.8695 - AD_Precision: 0.8370 - CN_Precision: 0.8617 - MCI_Precision: 0.8893 - AD_Recall: 0.8304 - CN_Recall: 0.8456 - MCI_Recall: 0.8949 - val_loss: 0.4067 - val_Accuracy: 0.8911 - val_AD_Precision: 0.8649 - val_CN_Precision: 0.9053 - val_MCI_Precision: 0.9003 - val_AD_Recall: 0.7559 - val_CN_Recall: 0.8731 - val_MCI_Recall: 0.9412\n",
            "Epoch 19/50\n",
            "698/698 [==============================] - 55s 78ms/step - loss: 0.4096 - Accuracy: 0.8777 - AD_Precision: 0.8323 - CN_Precision: 0.8707 - MCI_Precision: 0.8987 - AD_Recall: 0.8323 - CN_Recall: 0.8585 - MCI_Recall: 0.9019 - val_loss: 0.4098 - val_Accuracy: 0.8840 - val_AD_Precision: 0.8421 - val_CN_Precision: 0.9171 - val_MCI_Precision: 0.8806 - val_AD_Recall: 0.7559 - val_CN_Recall: 0.8426 - val_MCI_Recall: 0.9465\n",
            "Epoch 20/50\n",
            "698/698 [==============================] - 54s 77ms/step - loss: 0.3842 - Accuracy: 0.8853 - AD_Precision: 0.8498 - CN_Precision: 0.8768 - MCI_Precision: 0.9048 - AD_Recall: 0.8481 - CN_Recall: 0.8655 - MCI_Recall: 0.9054 - val_loss: 0.5556 - val_Accuracy: 0.8438 - val_AD_Precision: 0.6723 - val_CN_Precision: 0.8157 - val_MCI_Precision: 0.9733 - val_AD_Recall: 0.9370 - val_CN_Recall: 0.8985 - val_MCI_Recall: 0.7807\n",
            "Epoch 21/50\n",
            "698/698 [==============================] - 57s 82ms/step - loss: 0.4267 - Accuracy: 0.8734 - AD_Precision: 0.8406 - CN_Precision: 0.8650 - MCI_Precision: 0.8922 - AD_Recall: 0.8323 - CN_Recall: 0.8620 - MCI_Recall: 0.8928 - val_loss: 0.4044 - val_Accuracy: 0.8926 - val_AD_Precision: 0.8045 - val_CN_Precision: 0.9330 - val_MCI_Precision: 0.9089 - val_AD_Recall: 0.8425 - val_CN_Recall: 0.8477 - val_MCI_Recall: 0.9332\n",
            "Epoch 22/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 0.3465 - Accuracy: 0.8982 - AD_Precision: 0.8755 - CN_Precision: 0.8942 - MCI_Precision: 0.9126 - AD_Recall: 0.8600 - CN_Recall: 0.8795 - MCI_Recall: 0.9222 - val_loss: 0.7892 - val_Accuracy: 0.7708 - val_AD_Precision: 0.8509 - val_CN_Precision: 0.6038 - val_MCI_Precision: 0.9436 - val_AD_Recall: 0.7638 - val_CN_Recall: 0.9594 - val_MCI_Recall: 0.6711\n",
            "Epoch 23/50\n",
            "698/698 [==============================] - 58s 83ms/step - loss: 0.3337 - Accuracy: 0.8982 - AD_Precision: 0.8617 - CN_Precision: 0.9054 - MCI_Precision: 0.9113 - AD_Recall: 0.8481 - CN_Recall: 0.8959 - MCI_Recall: 0.9138 - val_loss: 0.3995 - val_Accuracy: 0.9026 - val_AD_Precision: 0.8296 - val_CN_Precision: 0.9382 - val_MCI_Precision: 0.9186 - val_AD_Recall: 0.8819 - val_CN_Recall: 0.8477 - val_MCI_Recall: 0.9358\n",
            "Epoch 24/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 0.2635 - Accuracy: 0.9107 - AD_Precision: 0.8902 - CN_Precision: 0.9022 - MCI_Precision: 0.9256 - AD_Recall: 0.8955 - CN_Recall: 0.8959 - MCI_Recall: 0.9236 - val_loss: 0.5130 - val_Accuracy: 0.8768 - val_AD_Precision: 0.7107 - val_CN_Precision: 0.9244 - val_MCI_Precision: 0.9288 - val_AD_Recall: 0.8898 - val_CN_Recall: 0.8071 - val_MCI_Recall: 0.9064\n",
            "Epoch 25/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 0.2663 - Accuracy: 0.9179 - AD_Precision: 0.9093 - CN_Precision: 0.9096 - MCI_Precision: 0.9289 - AD_Recall: 0.9093 - CN_Recall: 0.8947 - MCI_Recall: 0.9334 - val_loss: 1.0639 - val_Accuracy: 0.7278 - val_AD_Precision: 0.5484 - val_CN_Precision: 0.6860 - val_MCI_Precision: 0.9720 - val_AD_Recall: 0.9370 - val_CN_Recall: 0.8985 - val_MCI_Recall: 0.5561\n",
            "Epoch 26/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 0.2941 - Accuracy: 0.9161 - AD_Precision: 0.8815 - CN_Precision: 0.9167 - MCI_Precision: 0.9308 - AD_Recall: 0.8659 - CN_Recall: 0.9135 - MCI_Recall: 0.9327 - val_loss: 0.5745 - val_Accuracy: 0.8696 - val_AD_Precision: 0.8584 - val_CN_Precision: 0.9608 - val_MCI_Precision: 0.8415 - val_AD_Recall: 0.7638 - val_CN_Recall: 0.7462 - val_MCI_Recall: 0.9652\n",
            "Epoch 27/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 0.2418 - Accuracy: 0.9254 - AD_Precision: 0.9096 - CN_Precision: 0.9220 - MCI_Precision: 0.9357 - AD_Recall: 0.9132 - CN_Recall: 0.9123 - MCI_Recall: 0.9376 - val_loss: 0.7314 - val_Accuracy: 0.8438 - val_AD_Precision: 0.9733 - val_CN_Precision: 0.7262 - val_MCI_Precision: 0.9048 - val_AD_Recall: 0.5748 - val_CN_Recall: 0.9695 - val_MCI_Recall: 0.8636\n",
            "Epoch 28/50\n",
            "698/698 [==============================] - 56s 80ms/step - loss: 0.2547 - Accuracy: 0.9229 - AD_Precision: 0.9040 - CN_Precision: 0.9232 - MCI_Precision: 0.9318 - AD_Recall: 0.8915 - CN_Recall: 0.9135 - MCI_Recall: 0.9383 - val_loss: 0.4667 - val_Accuracy: 0.8911 - val_AD_Precision: 0.7762 - val_CN_Precision: 0.8812 - val_MCI_Precision: 0.9430 - val_AD_Recall: 0.8740 - val_CN_Recall: 0.9036 - val_MCI_Recall: 0.8850\n",
            "Epoch 29/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.2145 - Accuracy: 0.9347 - AD_Precision: 0.9281 - CN_Precision: 0.9220 - MCI_Precision: 0.9459 - AD_Recall: 0.9172 - CN_Recall: 0.9263 - MCI_Recall: 0.9425 - val_loss: 0.9323 - val_Accuracy: 0.7507 - val_AD_Precision: 0.6962 - val_CN_Precision: 0.6246 - val_MCI_Precision: 0.9496 - val_AD_Recall: 0.8661 - val_CN_Recall: 0.9289 - val_MCI_Recall: 0.6043\n",
            "Epoch 30/50\n",
            "698/698 [==============================] - 55s 78ms/step - loss: 0.1772 - Accuracy: 0.9405 - AD_Precision: 0.9165 - CN_Precision: 0.9436 - MCI_Precision: 0.9490 - AD_Recall: 0.9093 - CN_Recall: 0.9392 - MCI_Recall: 0.9509 - val_loss: 0.5369 - val_Accuracy: 0.8954 - val_AD_Precision: 0.9192 - val_CN_Precision: 0.9399 - val_MCI_Precision: 0.8699 - val_AD_Recall: 0.7165 - val_CN_Recall: 0.8731 - val_MCI_Recall: 0.9652\n",
            "Epoch 31/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.1919 - Accuracy: 0.9390 - AD_Precision: 0.9149 - CN_Precision: 0.9402 - MCI_Precision: 0.9481 - AD_Recall: 0.9112 - CN_Recall: 0.9380 - MCI_Recall: 0.9467 - val_loss: 0.6232 - val_Accuracy: 0.8754 - val_AD_Precision: 0.9348 - val_CN_Precision: 0.9871 - val_MCI_Precision: 0.8248 - val_AD_Recall: 0.6772 - val_CN_Recall: 0.7766 - val_MCI_Recall: 0.9947\n",
            "Epoch 32/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 0.2187 - Accuracy: 0.9283 - AD_Precision: 0.8955 - CN_Precision: 0.9382 - MCI_Precision: 0.9398 - AD_Recall: 0.8955 - CN_Recall: 0.9240 - MCI_Recall: 0.9411 - val_loss: 0.5057 - val_Accuracy: 0.8983 - val_AD_Precision: 0.9474 - val_CN_Precision: 0.9492 - val_MCI_Precision: 0.8670 - val_AD_Recall: 0.7087 - val_CN_Recall: 0.8528 - val_MCI_Recall: 0.9759\n",
            "Epoch 33/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 0.2151 - Accuracy: 0.9416 - AD_Precision: 0.9118 - CN_Precision: 0.9485 - MCI_Precision: 0.9483 - AD_Recall: 0.8974 - CN_Recall: 0.9485 - MCI_Recall: 0.9502 - val_loss: 0.4541 - val_Accuracy: 0.8983 - val_AD_Precision: 0.9020 - val_CN_Precision: 0.8792 - val_MCI_Precision: 0.9096 - val_AD_Recall: 0.7244 - val_CN_Recall: 0.9239 - val_MCI_Recall: 0.9412\n",
            "Epoch 34/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 0.1912 - Accuracy: 0.9437 - AD_Precision: 0.9274 - CN_Precision: 0.9397 - MCI_Precision: 0.9531 - AD_Recall: 0.9073 - CN_Recall: 0.9474 - MCI_Recall: 0.9537 - val_loss: 0.6543 - val_Accuracy: 0.8395 - val_AD_Precision: 0.5951 - val_CN_Precision: 0.9027 - val_MCI_Precision: 0.9704 - val_AD_Recall: 0.9606 - val_CN_Recall: 0.8477 - val_MCI_Recall: 0.7888\n",
            "Epoch 35/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 0.1674 - Accuracy: 0.9484 - AD_Precision: 0.9308 - CN_Precision: 0.9527 - MCI_Precision: 0.9533 - AD_Recall: 0.9290 - CN_Recall: 0.9427 - MCI_Recall: 0.9580 - val_loss: 0.8673 - val_Accuracy: 0.8453 - val_AD_Precision: 0.9844 - val_CN_Precision: 0.7393 - val_MCI_Precision: 0.8987 - val_AD_Recall: 0.4961 - val_CN_Recall: 0.9645 - val_MCI_Recall: 0.9011\n",
            "Epoch 36/50\n",
            "698/698 [==============================] - 59s 84ms/step - loss: 0.1672 - Accuracy: 0.9408 - AD_Precision: 0.9360 - CN_Precision: 0.9331 - MCI_Precision: 0.9470 - AD_Recall: 0.9231 - CN_Recall: 0.9298 - MCI_Recall: 0.9523 - val_loss: 0.3753 - val_Accuracy: 0.9112 - val_AD_Precision: 0.8947 - val_CN_Precision: 0.9355 - val_MCI_Precision: 0.9040 - val_AD_Recall: 0.8031 - val_CN_Recall: 0.8832 - val_MCI_Recall: 0.9572\n",
            "Epoch 37/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 0.1630 - Accuracy: 0.9491 - AD_Precision: 0.9347 - CN_Precision: 0.9483 - MCI_Precision: 0.9579 - AD_Recall: 0.9310 - CN_Recall: 0.9439 - MCI_Recall: 0.9573 - val_loss: 0.4684 - val_Accuracy: 0.8868 - val_AD_Precision: 0.7467 - val_CN_Precision: 0.8867 - val_MCI_Precision: 0.9504 - val_AD_Recall: 0.8819 - val_CN_Recall: 0.9137 - val_MCI_Recall: 0.8717\n",
            "Epoch 38/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 0.1296 - Accuracy: 0.9555 - AD_Precision: 0.9428 - CN_Precision: 0.9554 - MCI_Precision: 0.9601 - AD_Recall: 0.9428 - CN_Recall: 0.9520 - MCI_Recall: 0.9615 - val_loss: 0.5253 - val_Accuracy: 0.8926 - val_AD_Precision: 0.9074 - val_CN_Precision: 0.8267 - val_MCI_Precision: 0.9311 - val_AD_Recall: 0.7717 - val_CN_Recall: 0.9442 - val_MCI_Recall: 0.9037\n",
            "Epoch 39/50\n",
            "698/698 [==============================] - 54s 78ms/step - loss: 0.1981 - Accuracy: 0.9426 - AD_Precision: 0.9051 - CN_Precision: 0.9532 - MCI_Precision: 0.9509 - AD_Recall: 0.9034 - CN_Recall: 0.9520 - MCI_Recall: 0.9502 - val_loss: 0.4709 - val_Accuracy: 0.8968 - val_AD_Precision: 0.7986 - val_CN_Precision: 0.9110 - val_MCI_Precision: 0.9292 - val_AD_Recall: 0.8740 - val_CN_Recall: 0.8832 - val_MCI_Recall: 0.9118\n",
            "Epoch 40/50\n",
            "698/698 [==============================] - 55s 78ms/step - loss: 0.1224 - Accuracy: 0.9638 - AD_Precision: 0.9447 - CN_Precision: 0.9693 - MCI_Precision: 0.9686 - AD_Recall: 0.9428 - CN_Recall: 0.9614 - MCI_Recall: 0.9727 - val_loss: 0.5881 - val_Accuracy: 0.8926 - val_AD_Precision: 0.9451 - val_CN_Precision: 0.9551 - val_MCI_Precision: 0.8571 - val_AD_Recall: 0.6772 - val_CN_Recall: 0.8629 - val_MCI_Recall: 0.9786\n",
            "Epoch 41/50\n",
            "698/698 [==============================] - 58s 84ms/step - loss: 0.1579 - Accuracy: 0.9512 - AD_Precision: 0.9310 - CN_Precision: 0.9485 - MCI_Precision: 0.9600 - AD_Recall: 0.9310 - CN_Recall: 0.9485 - MCI_Recall: 0.9594 - val_loss: 0.4320 - val_Accuracy: 0.9140 - val_AD_Precision: 0.9434 - val_CN_Precision: 0.8894 - val_MCI_Precision: 0.9241 - val_AD_Recall: 0.7874 - val_CN_Recall: 0.9391 - val_MCI_Recall: 0.9439\n",
            "Epoch 42/50\n",
            "698/698 [==============================] - 57s 81ms/step - loss: 0.1567 - Accuracy: 0.9541 - AD_Precision: 0.9333 - CN_Precision: 0.9541 - MCI_Precision: 0.9621 - AD_Recall: 0.9389 - CN_Recall: 0.9485 - MCI_Recall: 0.9615 - val_loss: 0.5229 - val_Accuracy: 0.8968 - val_AD_Precision: 0.8803 - val_CN_Precision: 0.9581 - val_MCI_Precision: 0.8783 - val_AD_Recall: 0.8110 - val_CN_Recall: 0.8122 - val_MCI_Recall: 0.9652\n",
            "Epoch 43/50\n",
            "698/698 [==============================] - 55s 78ms/step - loss: 0.1213 - Accuracy: 0.9591 - AD_Precision: 0.9446 - CN_Precision: 0.9578 - MCI_Precision: 0.9664 - AD_Recall: 0.9408 - CN_Recall: 0.9567 - MCI_Recall: 0.9671 - val_loss: 0.5703 - val_Accuracy: 0.8625 - val_AD_Precision: 0.8261 - val_CN_Precision: 0.7621 - val_MCI_Precision: 0.9644 - val_AD_Recall: 0.8976 - val_CN_Recall: 0.9594 - val_MCI_Recall: 0.7968\n",
            "Epoch 44/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 0.1246 - Accuracy: 0.9591 - AD_Precision: 0.9562 - CN_Precision: 0.9511 - MCI_Precision: 0.9656 - AD_Recall: 0.9467 - CN_Recall: 0.9556 - MCI_Recall: 0.9650 - val_loss: 0.4643 - val_Accuracy: 0.9083 - val_AD_Precision: 0.8689 - val_CN_Precision: 0.8995 - val_MCI_Precision: 0.9253 - val_AD_Recall: 0.8346 - val_CN_Recall: 0.9086 - val_MCI_Recall: 0.9278\n",
            "Epoch 45/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 0.1691 - Accuracy: 0.9520 - AD_Precision: 0.9347 - CN_Precision: 0.9508 - MCI_Precision: 0.9587 - AD_Recall: 0.9310 - CN_Recall: 0.9497 - MCI_Recall: 0.9608 - val_loss: 0.6793 - val_Accuracy: 0.8825 - val_AD_Precision: 0.9747 - val_CN_Precision: 0.9215 - val_MCI_Precision: 0.8501 - val_AD_Recall: 0.6063 - val_CN_Recall: 0.8934 - val_MCI_Recall: 0.9706\n",
            "Epoch 46/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 0.1584 - Accuracy: 0.9559 - AD_Precision: 0.9332 - CN_Precision: 0.9588 - MCI_Precision: 0.9636 - AD_Recall: 0.9369 - CN_Recall: 0.9532 - MCI_Recall: 0.9636 - val_loss: 0.4967 - val_Accuracy: 0.8940 - val_AD_Precision: 0.8790 - val_CN_Precision: 0.8182 - val_MCI_Precision: 0.9559 - val_AD_Recall: 0.8583 - val_CN_Recall: 0.9594 - val_MCI_Recall: 0.8690\n",
            "Epoch 47/50\n",
            "698/698 [==============================] - 59s 84ms/step - loss: 0.1199 - Accuracy: 0.9627 - AD_Precision: 0.9446 - CN_Precision: 0.9682 - MCI_Precision: 0.9672 - AD_Recall: 0.9408 - CN_Recall: 0.9626 - MCI_Recall: 0.9699 - val_loss: 0.4199 - val_Accuracy: 0.9284 - val_AD_Precision: 0.8880 - val_CN_Precision: 0.9565 - val_MCI_Precision: 0.9304 - val_AD_Recall: 0.8740 - val_CN_Recall: 0.8934 - val_MCI_Recall: 0.9652\n",
            "Epoch 48/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 0.1205 - Accuracy: 0.9598 - AD_Precision: 0.9508 - CN_Precision: 0.9612 - MCI_Precision: 0.9643 - AD_Recall: 0.9527 - CN_Recall: 0.9556 - MCI_Recall: 0.9643 - val_loss: 0.3871 - val_Accuracy: 0.9183 - val_AD_Precision: 0.9000 - val_CN_Precision: 0.8889 - val_MCI_Precision: 0.9458 - val_AD_Recall: 0.8504 - val_CN_Recall: 0.9340 - val_MCI_Recall: 0.9332\n",
            "Epoch 49/50\n",
            "698/698 [==============================] - 55s 79ms/step - loss: 0.1072 - Accuracy: 0.9652 - AD_Precision: 0.9601 - CN_Precision: 0.9660 - MCI_Precision: 0.9693 - AD_Recall: 0.9487 - CN_Recall: 0.9626 - MCI_Recall: 0.9720 - val_loss: 0.3894 - val_Accuracy: 0.9183 - val_AD_Precision: 0.8444 - val_CN_Precision: 0.9219 - val_MCI_Precision: 0.9511 - val_AD_Recall: 0.8976 - val_CN_Recall: 0.8985 - val_MCI_Recall: 0.9358\n",
            "Epoch 50/50\n",
            "698/698 [==============================] - 56s 80ms/step - loss: 0.0842 - Accuracy: 0.9738 - AD_Precision: 0.9507 - CN_Precision: 0.9801 - MCI_Precision: 0.9797 - AD_Recall: 0.9507 - CN_Recall: 0.9778 - MCI_Recall: 0.9797 - val_loss: 0.4605 - val_Accuracy: 0.9183 - val_AD_Precision: 0.9231 - val_CN_Precision: 0.9653 - val_MCI_Precision: 0.8993 - val_AD_Recall: 0.8504 - val_CN_Recall: 0.8477 - val_MCI_Recall: 0.9786\n"
          ]
        }
      ],
      "source": [
        "# learning_rate=10E-6\n",
        "savepath = '/content/drive/MyDrive/ADNI/resnet50_pretrained_10E-6/'\n",
        "lr=0.000001\n",
        "if not os.path.isdir(savepath):\n",
        "      os.makedirs(savepath)\n",
        "csvloger_callback = tf.keras.callbacks.CSVLogger(f'{savepath}history.csv', separator=',', append=False)\n",
        "\n",
        "modelcheckpoint_callback = tf.keras.callbacks.ModelCheckpoint(savepath+'resnet50_pretrained_10E-6_weights-epoch:{epoch:03d}-val_Accuracy:{val_Accuracy:.4f}.h5',monitor='val_Accuracy'\n",
        "    ,verbose=0,save_best_only=True,save_weights_only=True,mode='max',save_freq='epoch' , initial_value_threshold=0.85)\n",
        "\n",
        "OPT = tf.keras.optimizers.Adam(learning_rate=lr)\n",
        "model03 = resnet50_pretrained(OPT, LOOS,METRICS,image_Shape)\n",
        "history03 = model03.fit(training_image_datagen, epochs=EPOCH, validation_data=testing_image_datagen,callbacks=[csvloger_callback,modelcheckpoint_callback])\n",
        "model03 = resnet50_pretrained(OPT, LOOS,METRICS,image_Shape)\n",
        "model03.load_weights(glob.glob(f\"{savepath}resnet50_pretrained_10E-6_weights-epoch:*-val_Accuracy:{max(history03.history['val_Accuracy']):.4f}.h5\")[0])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 477
        },
        "id": "ojk7VwObCM2S",
        "outputId": "e825788b-49d3-42ba-99db-8d99d85e8bad"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "          AD       0.89      0.87      0.88       127\n",
            "          CN       0.96      0.89      0.92       197\n",
            "         MCI       0.93      0.97      0.95       374\n",
            "\n",
            "    accuracy                           0.93       698\n",
            "   macro avg       0.92      0.91      0.92       698\n",
            "weighted avg       0.93      0.93      0.93       698\n",
            "\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "p = model03.predict(testing_image_datagen)\n",
        "predicted_lab = p.argmax(axis=-1)\n",
        "true_lab = testing_image_datagen.classes\n",
        "plot_confusion_matrix(true_lab, predicted_lab, testing_image_datagen.class_indices.keys(),savepath)\n",
        "report = classification_report(true_lab, predicted_lab, target_names=testing_image_datagen.class_indices.keys())\n",
        "print(report)\n",
        "dump_text(report,f'{savepath}classification_report.txt')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uZl2hcpBCve8"
      },
      "source": [
        "**Resnet 50 Row**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "U0LtUfIkC3ku",
        "outputId": "b4991222-8680-4fdc-9ad6-c7f852c3d2e2"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/resnet/resnet50_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "94773248/94765736 [==============================] - 1s 0us/step\n",
            "94781440/94765736 [==============================] - 1s 0us/step\n",
            "Epoch 1/50\n",
            "698/698 [==============================] - 118s 141ms/step - loss: 5.3829 - Accuracy: 0.4991 - AD_Precision: 0.3287 - CN_Precision: 0.4642 - MCI_Precision: 0.5838 - AD_Recall: 0.2821 - CN_Recall: 0.4316 - MCI_Recall: 0.5711 - val_loss: 1.0405 - val_Accuracy: 0.5258 - val_AD_Precision: 0.3484 - val_CN_Precision: 0.7125 - val_MCI_Precision: 0.7100 - val_AD_Recall: 0.7874 - val_CN_Recall: 0.2893 - val_MCI_Recall: 0.5107\n",
            "Epoch 2/50\n",
            "698/698 [==============================] - 96s 138ms/step - loss: 0.9363 - Accuracy: 0.6977 - AD_Precision: 0.6302 - CN_Precision: 0.6880 - MCI_Precision: 0.7481 - AD_Recall: 0.5680 - CN_Recall: 0.6678 - MCI_Recall: 0.7428 - val_loss: 1.2953 - val_Accuracy: 0.5874 - val_AD_Precision: 0.8723 - val_CN_Precision: 0.4463 - val_MCI_Precision: 0.8194 - val_AD_Recall: 0.3228 - val_CN_Recall: 0.9492 - val_MCI_Recall: 0.4733\n",
            "Epoch 3/50\n",
            "698/698 [==============================] - 96s 137ms/step - loss: 0.9512 - Accuracy: 0.7200 - AD_Precision: 0.6477 - CN_Precision: 0.7114 - MCI_Precision: 0.7610 - AD_Recall: 0.5838 - CN_Recall: 0.6947 - MCI_Recall: 0.7610 - val_loss: 0.9160 - val_Accuracy: 0.6805 - val_AD_Precision: 0.5238 - val_CN_Precision: 0.6748 - val_MCI_Precision: 0.8868 - val_AD_Recall: 0.8661 - val_CN_Recall: 0.8426 - val_MCI_Recall: 0.5027\n",
            "Epoch 4/50\n",
            "698/698 [==============================] - 96s 138ms/step - loss: 0.6541 - Accuracy: 0.7831 - AD_Precision: 0.7173 - CN_Precision: 0.7943 - MCI_Precision: 0.8120 - AD_Recall: 0.6805 - CN_Recall: 0.7766 - MCI_Recall: 0.8080 - val_loss: 1.1091 - val_Accuracy: 0.6691 - val_AD_Precision: 0.9565 - val_CN_Precision: 0.8667 - val_MCI_Precision: 0.6282 - val_AD_Recall: 0.1732 - val_CN_Recall: 0.3959 - val_MCI_Recall: 0.9759\n",
            "Epoch 5/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.5985 - Accuracy: 0.8093 - AD_Precision: 0.7531 - CN_Precision: 0.8100 - MCI_Precision: 0.8380 - AD_Recall: 0.7219 - CN_Recall: 0.7977 - MCI_Recall: 0.8374 - val_loss: 0.6838 - val_Accuracy: 0.7650 - val_AD_Precision: 0.7238 - val_CN_Precision: 0.9596 - val_MCI_Precision: 0.7433 - val_AD_Recall: 0.5984 - val_CN_Recall: 0.4822 - val_MCI_Recall: 0.9599\n",
            "Epoch 6/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.4777 - Accuracy: 0.8329 - AD_Precision: 0.7707 - CN_Precision: 0.8503 - MCI_Precision: 0.8582 - AD_Recall: 0.7357 - CN_Recall: 0.8304 - MCI_Recall: 0.8612 - val_loss: 0.7242 - val_Accuracy: 0.7536 - val_AD_Precision: 0.4852 - val_CN_Precision: 0.9384 - val_MCI_Precision: 0.8993 - val_AD_Recall: 0.9055 - val_CN_Recall: 0.6954 - val_MCI_Recall: 0.7166\n",
            "Epoch 7/50\n",
            "698/698 [==============================] - 96s 137ms/step - loss: 0.5017 - Accuracy: 0.8387 - AD_Precision: 0.7984 - CN_Precision: 0.8289 - MCI_Precision: 0.8636 - AD_Recall: 0.7653 - CN_Recall: 0.8105 - MCI_Recall: 0.8739 - val_loss: 1.2390 - val_Accuracy: 0.6934 - val_AD_Precision: 0.4953 - val_CN_Precision: 0.9438 - val_MCI_Precision: 0.7405 - val_AD_Recall: 0.8268 - val_CN_Recall: 0.4264 - val_MCI_Recall: 0.7781\n",
            "Epoch 8/50\n",
            "698/698 [==============================] - 96s 137ms/step - loss: 0.4599 - Accuracy: 0.8569 - AD_Precision: 0.8122 - CN_Precision: 0.8568 - MCI_Precision: 0.8787 - AD_Recall: 0.7850 - CN_Recall: 0.8398 - MCI_Recall: 0.8830 - val_loss: 0.6881 - val_Accuracy: 0.7794 - val_AD_Precision: 0.6861 - val_CN_Precision: 0.9444 - val_MCI_Precision: 0.7815 - val_AD_Recall: 0.7402 - val_CN_Recall: 0.5178 - val_MCI_Recall: 0.9278\n",
            "Epoch 9/50\n",
            "698/698 [==============================] - 97s 138ms/step - loss: 0.4062 - Accuracy: 0.8677 - AD_Precision: 0.8100 - CN_Precision: 0.8710 - MCI_Precision: 0.8906 - AD_Recall: 0.7988 - CN_Recall: 0.8526 - MCI_Recall: 0.8956 - val_loss: 1.7329 - val_Accuracy: 0.7221 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.7337 - val_MCI_Precision: 0.7140 - val_AD_Recall: 0.0472 - val_CN_Recall: 0.7411 - val_MCI_Recall: 0.9412\n",
            "Epoch 10/50\n",
            "698/698 [==============================] - 101s 144ms/step - loss: 0.4077 - Accuracy: 0.8688 - AD_Precision: 0.8151 - CN_Precision: 0.8749 - MCI_Precision: 0.8896 - AD_Recall: 0.8087 - CN_Recall: 0.8585 - MCI_Recall: 0.8921 - val_loss: 0.4839 - val_Accuracy: 0.8582 - val_AD_Precision: 0.8681 - val_CN_Precision: 0.8246 - val_MCI_Precision: 0.8903 - val_AD_Recall: 0.6220 - val_CN_Recall: 0.8832 - val_MCI_Recall: 0.9118\n",
            "Epoch 11/50\n",
            "698/698 [==============================] - 96s 138ms/step - loss: 0.3499 - Accuracy: 0.8878 - AD_Precision: 0.8491 - CN_Precision: 0.8859 - MCI_Precision: 0.9081 - AD_Recall: 0.8323 - CN_Recall: 0.8807 - MCI_Recall: 0.9075 - val_loss: 0.9916 - val_Accuracy: 0.7450 - val_AD_Precision: 0.4768 - val_CN_Precision: 0.9683 - val_MCI_Precision: 0.8537 - val_AD_Recall: 0.8898 - val_CN_Recall: 0.6193 - val_MCI_Recall: 0.7487\n",
            "Epoch 12/50\n",
            "698/698 [==============================] - 96s 138ms/step - loss: 0.2737 - Accuracy: 0.9132 - AD_Precision: 0.8998 - CN_Precision: 0.9122 - MCI_Precision: 0.9266 - AD_Recall: 0.8856 - CN_Recall: 0.8994 - MCI_Recall: 0.9292 - val_loss: 0.6905 - val_Accuracy: 0.8281 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.8142 - val_MCI_Precision: 0.8254 - val_AD_Recall: 0.3858 - val_CN_Recall: 0.9340 - val_MCI_Recall: 0.9225\n",
            "Epoch 13/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.2858 - Accuracy: 0.9143 - AD_Precision: 0.8831 - CN_Precision: 0.9111 - MCI_Precision: 0.9333 - AD_Recall: 0.8639 - CN_Recall: 0.9111 - MCI_Recall: 0.9320 - val_loss: 1.8546 - val_Accuracy: 0.6032 - val_AD_Precision: 0.5209 - val_CN_Precision: 0.5424 - val_MCI_Precision: 0.9828 - val_AD_Recall: 0.8819 - val_CN_Recall: 0.9746 - val_MCI_Recall: 0.3048\n",
            "Epoch 14/50\n",
            "698/698 [==============================] - 97s 138ms/step - loss: 0.2692 - Accuracy: 0.9186 - AD_Precision: 0.8788 - CN_Precision: 0.9249 - MCI_Precision: 0.9336 - AD_Recall: 0.8580 - CN_Recall: 0.9216 - MCI_Recall: 0.9362 - val_loss: 0.8102 - val_Accuracy: 0.7120 - val_AD_Precision: 0.4627 - val_CN_Precision: 0.8272 - val_MCI_Precision: 0.9270 - val_AD_Recall: 0.9291 - val_CN_Recall: 0.8020 - val_MCI_Recall: 0.5775\n",
            "Epoch 15/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.2957 - Accuracy: 0.9139 - AD_Precision: 0.8667 - CN_Precision: 0.9281 - MCI_Precision: 0.9268 - AD_Recall: 0.8718 - CN_Recall: 0.9205 - MCI_Recall: 0.9229 - val_loss: 0.8463 - val_Accuracy: 0.7464 - val_AD_Precision: 0.4539 - val_CN_Precision: 0.9632 - val_MCI_Precision: 0.9266 - val_AD_Recall: 0.9685 - val_CN_Recall: 0.6650 - val_MCI_Recall: 0.7086\n",
            "Epoch 16/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.2087 - Accuracy: 0.9355 - AD_Precision: 0.9032 - CN_Precision: 0.9420 - MCI_Precision: 0.9449 - AD_Recall: 0.9014 - CN_Recall: 0.9310 - MCI_Recall: 0.9495 - val_loss: 1.3610 - val_Accuracy: 0.7278 - val_AD_Precision: 0.8824 - val_CN_Precision: 0.5374 - val_MCI_Precision: 0.9610 - val_AD_Recall: 0.7087 - val_CN_Recall: 0.9848 - val_MCI_Recall: 0.5936\n",
            "Epoch 17/50\n",
            "698/698 [==============================] - 97s 140ms/step - loss: 0.2386 - Accuracy: 0.9387 - AD_Precision: 0.9208 - CN_Precision: 0.9385 - MCI_Precision: 0.9456 - AD_Recall: 0.9172 - CN_Recall: 0.9287 - MCI_Recall: 0.9509 - val_loss: 0.5269 - val_Accuracy: 0.8453 - val_AD_Precision: 0.8060 - val_CN_Precision: 0.7266 - val_MCI_Precision: 0.9696 - val_AD_Recall: 0.8504 - val_CN_Recall: 0.9848 - val_MCI_Recall: 0.7674\n",
            "Epoch 18/50\n",
            "698/698 [==============================] - 97s 138ms/step - loss: 0.2483 - Accuracy: 0.9286 - AD_Precision: 0.9078 - CN_Precision: 0.9208 - MCI_Precision: 0.9423 - AD_Recall: 0.8935 - CN_Recall: 0.9251 - MCI_Recall: 0.9390 - val_loss: 0.6070 - val_Accuracy: 0.8295 - val_AD_Precision: 0.9126 - val_CN_Precision: 0.9835 - val_MCI_Precision: 0.7733 - val_AD_Recall: 0.7402 - val_CN_Recall: 0.6041 - val_MCI_Recall: 0.9759\n",
            "Epoch 19/50\n",
            "698/698 [==============================] - 104s 149ms/step - loss: 0.1394 - Accuracy: 0.9555 - AD_Precision: 0.9442 - CN_Precision: 0.9544 - MCI_Precision: 0.9608 - AD_Recall: 0.9349 - CN_Recall: 0.9556 - MCI_Recall: 0.9622 - val_loss: 0.9914 - val_Accuracy: 0.7994 - val_AD_Precision: 0.9342 - val_CN_Precision: 0.6328 - val_MCI_Precision: 0.9272 - val_AD_Recall: 0.5591 - val_CN_Recall: 0.9797 - val_MCI_Recall: 0.7834\n",
            "Epoch 20/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.2364 - Accuracy: 0.9365 - AD_Precision: 0.9195 - CN_Precision: 0.9312 - MCI_Precision: 0.9487 - AD_Recall: 0.9014 - CN_Recall: 0.9345 - MCI_Recall: 0.9467 - val_loss: 1.8058 - val_Accuracy: 0.6777 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.4852 - val_MCI_Precision: 0.9442 - val_AD_Recall: 0.4331 - val_CN_Recall: 1.0000 - val_MCI_Recall: 0.5882\n",
            "Epoch 21/50\n",
            "698/698 [==============================] - 101s 145ms/step - loss: 0.1825 - Accuracy: 0.9444 - AD_Precision: 0.9414 - CN_Precision: 0.9340 - MCI_Precision: 0.9530 - AD_Recall: 0.9507 - CN_Recall: 0.9275 - MCI_Recall: 0.9516 - val_loss: 0.6041 - val_Accuracy: 0.8653 - val_AD_Precision: 0.7744 - val_CN_Precision: 0.7959 - val_MCI_Precision: 0.9590 - val_AD_Recall: 0.8110 - val_CN_Recall: 0.9898 - val_MCI_Recall: 0.8128\n",
            "Epoch 22/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.1836 - Accuracy: 0.9477 - AD_Precision: 0.9193 - CN_Precision: 0.9543 - MCI_Precision: 0.9544 - AD_Recall: 0.9211 - CN_Recall: 0.9520 - MCI_Recall: 0.9530 - val_loss: 0.7317 - val_Accuracy: 0.8352 - val_AD_Precision: 0.7181 - val_CN_Precision: 0.9701 - val_MCI_Precision: 0.8390 - val_AD_Recall: 0.8425 - val_CN_Recall: 0.6599 - val_MCI_Recall: 0.9198\n",
            "Epoch 23/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.1350 - Accuracy: 0.9609 - AD_Precision: 0.9418 - CN_Precision: 0.9639 - MCI_Precision: 0.9677 - AD_Recall: 0.9250 - CN_Recall: 0.9684 - MCI_Recall: 0.9671 - val_loss: 1.2958 - val_Accuracy: 0.8166 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.8953 - val_MCI_Precision: 0.7712 - val_AD_Recall: 0.2756 - val_CN_Recall: 0.8680 - val_MCI_Recall: 0.9733\n",
            "Epoch 24/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.1775 - Accuracy: 0.9512 - AD_Precision: 0.9364 - CN_Precision: 0.9519 - MCI_Precision: 0.9587 - AD_Recall: 0.9290 - CN_Recall: 0.9485 - MCI_Recall: 0.9608 - val_loss: 1.1366 - val_Accuracy: 0.7536 - val_AD_Precision: 0.7801 - val_CN_Precision: 0.5913 - val_MCI_Precision: 0.9780 - val_AD_Recall: 0.8661 - val_CN_Recall: 0.9695 - val_MCI_Recall: 0.5936\n",
            "Epoch 25/50\n",
            "698/698 [==============================] - 96s 138ms/step - loss: 0.1390 - Accuracy: 0.9620 - AD_Precision: 0.9489 - CN_Precision: 0.9635 - MCI_Precision: 0.9664 - AD_Recall: 0.9527 - CN_Recall: 0.9579 - MCI_Recall: 0.9678 - val_loss: 0.7980 - val_Accuracy: 0.8152 - val_AD_Precision: 0.8919 - val_CN_Precision: 0.7469 - val_MCI_Precision: 0.8519 - val_AD_Recall: 0.5197 - val_CN_Recall: 0.9137 - val_MCI_Recall: 0.8610\n",
            "Epoch 26/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.2225 - Accuracy: 0.9380 - AD_Precision: 0.9124 - CN_Precision: 0.9384 - MCI_Precision: 0.9480 - AD_Recall: 0.9034 - CN_Recall: 0.9439 - MCI_Recall: 0.9460 - val_loss: 0.6265 - val_Accuracy: 0.8281 - val_AD_Precision: 0.8602 - val_CN_Precision: 0.6940 - val_MCI_Precision: 0.9495 - val_AD_Recall: 0.6299 - val_CN_Recall: 0.9898 - val_MCI_Recall: 0.8048\n",
            "Epoch 27/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.1380 - Accuracy: 0.9566 - AD_Precision: 0.9328 - CN_Precision: 0.9612 - MCI_Precision: 0.9636 - AD_Recall: 0.9310 - CN_Recall: 0.9567 - MCI_Recall: 0.9650 - val_loss: 0.9986 - val_Accuracy: 0.8223 - val_AD_Precision: 0.9326 - val_CN_Precision: 0.9209 - val_MCI_Precision: 0.7730 - val_AD_Recall: 0.6535 - val_CN_Recall: 0.6497 - val_MCI_Recall: 0.9652\n",
            "Epoch 28/50\n",
            "698/698 [==============================] - 101s 145ms/step - loss: 0.1537 - Accuracy: 0.9613 - AD_Precision: 0.9411 - CN_Precision: 0.9626 - MCI_Precision: 0.9677 - AD_Recall: 0.9448 - CN_Recall: 0.9637 - MCI_Recall: 0.9650 - val_loss: 0.6831 - val_Accuracy: 0.8911 - val_AD_Precision: 0.9060 - val_CN_Precision: 0.9565 - val_MCI_Precision: 0.8640 - val_AD_Recall: 0.8346 - val_CN_Recall: 0.7817 - val_MCI_Recall: 0.9679\n",
            "Epoch 29/50\n",
            "698/698 [==============================] - 101s 145ms/step - loss: 0.1303 - Accuracy: 0.9641 - AD_Precision: 0.9502 - CN_Precision: 0.9614 - MCI_Precision: 0.9713 - AD_Recall: 0.9408 - CN_Recall: 0.9614 - MCI_Recall: 0.9734 - val_loss: 0.4449 - val_Accuracy: 0.8940 - val_AD_Precision: 0.8264 - val_CN_Precision: 0.9010 - val_MCI_Precision: 0.9140 - val_AD_Recall: 0.7874 - val_CN_Recall: 0.9239 - val_MCI_Recall: 0.9091\n",
            "Epoch 30/50\n",
            "698/698 [==============================] - 103s 147ms/step - loss: 0.1592 - Accuracy: 0.9566 - AD_Precision: 0.9421 - CN_Precision: 0.9532 - MCI_Precision: 0.9644 - AD_Recall: 0.9310 - CN_Recall: 0.9532 - MCI_Recall: 0.9678 - val_loss: 0.6264 - val_Accuracy: 0.9054 - val_AD_Precision: 0.8626 - val_CN_Precision: 0.8714 - val_MCI_Precision: 0.9412 - val_AD_Recall: 0.8898 - val_CN_Recall: 0.9289 - val_MCI_Recall: 0.8984\n",
            "Epoch 31/50\n",
            "698/698 [==============================] - 99s 142ms/step - loss: 0.1833 - Accuracy: 0.9566 - AD_Precision: 0.9450 - CN_Precision: 0.9508 - MCI_Precision: 0.9649 - AD_Recall: 0.9487 - CN_Recall: 0.9497 - MCI_Recall: 0.9629 - val_loss: 2.2522 - val_Accuracy: 0.7020 - val_AD_Precision: 0.6688 - val_CN_Precision: 0.5621 - val_MCI_Precision: 0.9701 - val_AD_Recall: 0.8268 - val_CN_Recall: 0.9645 - val_MCI_Recall: 0.5214\n",
            "Epoch 32/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.1803 - Accuracy: 0.9588 - AD_Precision: 0.9303 - CN_Precision: 0.9640 - MCI_Precision: 0.9669 - AD_Recall: 0.9211 - CN_Recall: 0.9696 - MCI_Recall: 0.9636 - val_loss: 0.4507 - val_Accuracy: 0.9011 - val_AD_Precision: 0.8102 - val_CN_Precision: 0.9514 - val_MCI_Precision: 0.9096 - val_AD_Recall: 0.8740 - val_CN_Recall: 0.8934 - val_MCI_Recall: 0.9144\n",
            "Epoch 33/50\n",
            "698/698 [==============================] - 98s 141ms/step - loss: 0.0989 - Accuracy: 0.9735 - AD_Precision: 0.9583 - CN_Precision: 0.9789 - MCI_Precision: 0.9755 - AD_Recall: 0.9507 - CN_Recall: 0.9789 - MCI_Recall: 0.9776 - val_loss: 0.5986 - val_Accuracy: 0.8782 - val_AD_Precision: 0.8972 - val_CN_Precision: 0.7810 - val_MCI_Precision: 0.9398 - val_AD_Recall: 0.7559 - val_CN_Recall: 0.9594 - val_MCI_Recall: 0.8770\n",
            "Epoch 34/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.1379 - Accuracy: 0.9674 - AD_Precision: 0.9604 - CN_Precision: 0.9672 - MCI_Precision: 0.9699 - AD_Recall: 0.9566 - CN_Recall: 0.9649 - MCI_Recall: 0.9720 - val_loss: 1.7570 - val_Accuracy: 0.7966 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.6340 - val_MCI_Precision: 0.9150 - val_AD_Recall: 0.2913 - val_CN_Recall: 0.9848 - val_MCI_Recall: 0.8636\n",
            "Epoch 35/50\n",
            "698/698 [==============================] - 98s 141ms/step - loss: 0.0983 - Accuracy: 0.9710 - AD_Precision: 0.9587 - CN_Precision: 0.9718 - MCI_Precision: 0.9762 - AD_Recall: 0.9606 - CN_Recall: 0.9684 - MCI_Recall: 0.9762 - val_loss: 0.7109 - val_Accuracy: 0.8825 - val_AD_Precision: 0.7041 - val_CN_Precision: 0.9055 - val_MCI_Precision: 0.9604 - val_AD_Recall: 0.9370 - val_CN_Recall: 0.9239 - val_MCI_Recall: 0.8422\n",
            "Epoch 36/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.1415 - Accuracy: 0.9670 - AD_Precision: 0.9605 - CN_Precision: 0.9588 - MCI_Precision: 0.9748 - AD_Recall: 0.9586 - CN_Recall: 0.9532 - MCI_Recall: 0.9769 - val_loss: 1.1121 - val_Accuracy: 0.8553 - val_AD_Precision: 0.9524 - val_CN_Precision: 0.9207 - val_MCI_Precision: 0.8133 - val_AD_Recall: 0.6299 - val_CN_Recall: 0.7665 - val_MCI_Recall: 0.9786\n",
            "Epoch 37/50\n",
            "698/698 [==============================] - 102s 146ms/step - loss: 0.1308 - Accuracy: 0.9681 - AD_Precision: 0.9602 - CN_Precision: 0.9627 - MCI_Precision: 0.9741 - AD_Recall: 0.9527 - CN_Recall: 0.9649 - MCI_Recall: 0.9748 - val_loss: 0.5904 - val_Accuracy: 0.9069 - val_AD_Precision: 0.9545 - val_CN_Precision: 0.8832 - val_MCI_Precision: 0.9089 - val_AD_Recall: 0.6614 - val_CN_Recall: 0.9594 - val_MCI_Recall: 0.9599\n",
            "Epoch 38/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.1402 - Accuracy: 0.9670 - AD_Precision: 0.9524 - CN_Precision: 0.9650 - MCI_Precision: 0.9734 - AD_Recall: 0.9467 - CN_Recall: 0.9661 - MCI_Recall: 0.9741 - val_loss: 0.8699 - val_Accuracy: 0.8854 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.8900 - val_MCI_Precision: 0.8667 - val_AD_Recall: 0.5354 - val_CN_Recall: 0.9442 - val_MCI_Recall: 0.9733\n",
            "Epoch 39/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.0978 - Accuracy: 0.9767 - AD_Precision: 0.9586 - CN_Precision: 0.9836 - MCI_Precision: 0.9797 - AD_Recall: 0.9586 - CN_Recall: 0.9825 - MCI_Recall: 0.9797 - val_loss: 0.9328 - val_Accuracy: 0.8381 - val_AD_Precision: 0.8738 - val_CN_Precision: 0.9200 - val_MCI_Precision: 0.8036 - val_AD_Recall: 0.7087 - val_CN_Recall: 0.7005 - val_MCI_Recall: 0.9519\n",
            "Epoch 40/50\n",
            "698/698 [==============================] - 97s 138ms/step - loss: 0.1413 - Accuracy: 0.9634 - AD_Precision: 0.9483 - CN_Precision: 0.9660 - MCI_Precision: 0.9692 - AD_Recall: 0.9408 - CN_Recall: 0.9649 - MCI_Recall: 0.9706 - val_loss: 0.6695 - val_Accuracy: 0.8954 - val_AD_Precision: 0.9691 - val_CN_Precision: 0.9333 - val_MCI_Precision: 0.8622 - val_AD_Recall: 0.7402 - val_CN_Recall: 0.8528 - val_MCI_Recall: 0.9706\n",
            "Epoch 41/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.1114 - Accuracy: 0.9720 - AD_Precision: 0.9564 - CN_Precision: 0.9732 - MCI_Precision: 0.9775 - AD_Recall: 0.9527 - CN_Recall: 0.9754 - MCI_Recall: 0.9762 - val_loss: 0.5311 - val_Accuracy: 0.8883 - val_AD_Precision: 0.8154 - val_CN_Precision: 0.9382 - val_MCI_Precision: 0.8897 - val_AD_Recall: 0.8346 - val_CN_Recall: 0.8477 - val_MCI_Recall: 0.9278\n",
            "Epoch 42/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0762 - Accuracy: 0.9796 - AD_Precision: 0.9744 - CN_Precision: 0.9755 - MCI_Precision: 0.9846 - AD_Recall: 0.9763 - CN_Recall: 0.9766 - MCI_Recall: 0.9825 - val_loss: 1.0648 - val_Accuracy: 0.8768 - val_AD_Precision: 0.9610 - val_CN_Precision: 0.9194 - val_MCI_Precision: 0.8437 - val_AD_Recall: 0.5827 - val_CN_Recall: 0.8680 - val_MCI_Recall: 0.9813\n",
            "Epoch 43/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.1946 - Accuracy: 0.9573 - AD_Precision: 0.9386 - CN_Precision: 0.9566 - MCI_Precision: 0.9657 - AD_Recall: 0.9349 - CN_Recall: 0.9544 - MCI_Recall: 0.9664 - val_loss: 0.7659 - val_Accuracy: 0.8711 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.8348 - val_MCI_Precision: 0.8703 - val_AD_Recall: 0.5591 - val_CN_Recall: 0.9492 - val_MCI_Recall: 0.9332\n",
            "Epoch 44/50\n",
            "698/698 [==============================] - 101s 145ms/step - loss: 0.0796 - Accuracy: 0.9774 - AD_Precision: 0.9742 - CN_Precision: 0.9766 - MCI_Precision: 0.9804 - AD_Recall: 0.9665 - CN_Recall: 0.9754 - MCI_Recall: 0.9818 - val_loss: 0.5374 - val_Accuracy: 0.9112 - val_AD_Precision: 0.8507 - val_CN_Precision: 0.9759 - val_MCI_Precision: 0.9091 - val_AD_Recall: 0.8976 - val_CN_Recall: 0.8223 - val_MCI_Recall: 0.9626\n",
            "Epoch 45/50\n",
            "698/698 [==============================] - 98s 141ms/step - loss: 0.0887 - Accuracy: 0.9774 - AD_Precision: 0.9743 - CN_Precision: 0.9778 - MCI_Precision: 0.9790 - AD_Recall: 0.9724 - CN_Recall: 0.9766 - MCI_Recall: 0.9790 - val_loss: 0.5242 - val_Accuracy: 0.9083 - val_AD_Precision: 0.9273 - val_CN_Precision: 0.8514 - val_MCI_Precision: 0.9448 - val_AD_Recall: 0.8031 - val_CN_Recall: 0.9594 - val_MCI_Recall: 0.9144\n",
            "Epoch 46/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.1115 - Accuracy: 0.9724 - AD_Precision: 0.9664 - CN_Precision: 0.9707 - MCI_Precision: 0.9755 - AD_Recall: 0.9645 - CN_Recall: 0.9696 - MCI_Recall: 0.9769 - val_loss: 0.5299 - val_Accuracy: 0.9040 - val_AD_Precision: 0.8525 - val_CN_Precision: 0.9716 - val_MCI_Precision: 0.8900 - val_AD_Recall: 0.8189 - val_CN_Recall: 0.8680 - val_MCI_Recall: 0.9519\n",
            "Epoch 47/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.1396 - Accuracy: 0.9674 - AD_Precision: 0.9623 - CN_Precision: 0.9683 - MCI_Precision: 0.9693 - AD_Recall: 0.9566 - CN_Recall: 0.9661 - MCI_Recall: 0.9720 - val_loss: 1.2398 - val_Accuracy: 0.8754 - val_AD_Precision: 0.9620 - val_CN_Precision: 0.9194 - val_MCI_Precision: 0.8406 - val_AD_Recall: 0.5984 - val_CN_Recall: 0.8680 - val_MCI_Recall: 0.9733\n",
            "Epoch 48/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.1378 - Accuracy: 0.9695 - AD_Precision: 0.9486 - CN_Precision: 0.9707 - MCI_Precision: 0.9762 - AD_Recall: 0.9467 - CN_Recall: 0.9696 - MCI_Recall: 0.9776 - val_loss: 0.8602 - val_Accuracy: 0.8295 - val_AD_Precision: 0.9438 - val_CN_Precision: 0.6844 - val_MCI_Precision: 0.9231 - val_AD_Recall: 0.6614 - val_CN_Recall: 0.9797 - val_MCI_Recall: 0.8021\n",
            "Epoch 49/50\n",
            "698/698 [==============================] - 102s 146ms/step - loss: 0.0432 - Accuracy: 0.9871 - AD_Precision: 0.9802 - CN_Precision: 0.9895 - MCI_Precision: 0.9881 - AD_Recall: 0.9783 - CN_Recall: 0.9918 - MCI_Recall: 0.9874 - val_loss: 0.4977 - val_Accuracy: 0.9140 - val_AD_Precision: 0.8492 - val_CN_Precision: 0.9721 - val_MCI_Precision: 0.9084 - val_AD_Recall: 0.8425 - val_CN_Recall: 0.8832 - val_MCI_Recall: 0.9545\n",
            "Epoch 50/50\n",
            "698/698 [==============================] - 99s 142ms/step - loss: 0.1432 - Accuracy: 0.9688 - AD_Precision: 0.9584 - CN_Precision: 0.9638 - MCI_Precision: 0.9768 - AD_Recall: 0.9546 - CN_Recall: 0.9661 - MCI_Recall: 0.9755 - val_loss: 0.8536 - val_Accuracy: 0.8797 - val_AD_Precision: 0.7586 - val_CN_Precision: 0.9684 - val_MCI_Precision: 0.8886 - val_AD_Recall: 0.8661 - val_CN_Recall: 0.7766 - val_MCI_Recall: 0.9385\n"
          ]
        }
      ],
      "source": [
        "# learning_rate=10E-4\n",
        "savepath = '/content/drive/MyDrive/ADNI/resnet50_row_10E-4/'\n",
        "lr=0.0001\n",
        "if not os.path.isdir(savepath):\n",
        "      os.makedirs(savepath)\n",
        "csvloger_callback = tf.keras.callbacks.CSVLogger(f'{savepath}history.csv', separator=',', append=False)\n",
        "\n",
        "modelcheckpoint_callback = tf.keras.callbacks.ModelCheckpoint(savepath+'resnet50_row_10E-4_weights-epoch:{epoch:03d}-val_Accuracy:{val_Accuracy:.4f}.h5',monitor='val_Accuracy'\n",
        "    ,verbose=0,save_best_only=True,save_weights_only=True,mode='max',save_freq='epoch' , initial_value_threshold=0.85)\n",
        "\n",
        "OPT = tf.keras.optimizers.Adam(learning_rate=lr)\n",
        "model04 = resnet50_row(OPT, LOOS,METRICS,image_Shape)\n",
        "history04 = model04.fit(training_image_datagen, epochs=EPOCH, validation_data=testing_image_datagen,callbacks=[csvloger_callback,modelcheckpoint_callback])\n",
        "model04 = resnet50_row(OPT, LOOS,METRICS,image_Shape)\n",
        "model04.load_weights(glob.glob(f\"{savepath}resnet50_row_10E-4_weights-epoch:*-val_Accuracy:{max(history04.history['val_Accuracy']):.4f}.h5\")[0])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 477
        },
        "id": "fwVOz_T8cuhY",
        "outputId": "22354376-3130-4067-cd04-6ba6cbae4d03"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "          AD       0.85      0.84      0.85       127\n",
            "          CN       0.97      0.88      0.93       197\n",
            "         MCI       0.91      0.95      0.93       374\n",
            "\n",
            "    accuracy                           0.91       698\n",
            "   macro avg       0.91      0.89      0.90       698\n",
            "weighted avg       0.92      0.91      0.91       698\n",
            "\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "p = model04.predict(testing_image_datagen)\n",
        "predicted_lab = p.argmax(axis=-1)\n",
        "true_lab = testing_image_datagen.classes\n",
        "plot_confusion_matrix(true_lab, predicted_lab, testing_image_datagen.class_indices.keys(),savepath)\n",
        "report = classification_report(true_lab, predicted_lab, target_names=testing_image_datagen.class_indices.keys())\n",
        "print(report)\n",
        "dump_text(report,f'{savepath}classification_report.txt')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "J8fZytmUDIGL",
        "outputId": "2ced98e5-b6a6-4540-f5ac-8560ec69e4c0"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "698/698 [==============================] - 106s 142ms/step - loss: 6.6077 - Accuracy: 0.5971 - AD_Precision: 0.4219 - CN_Precision: 0.5789 - MCI_Precision: 0.6687 - AD_Recall: 0.4306 - CN_Recall: 0.5580 - MCI_Recall: 0.6757 - val_loss: 3.3094 - val_Accuracy: 0.6562 - val_AD_Precision: 0.5631 - val_CN_Precision: 0.5633 - val_MCI_Precision: 0.7438 - val_AD_Recall: 0.4567 - val_CN_Recall: 0.6548 - val_MCI_Recall: 0.7219\n",
            "Epoch 2/50\n",
            "698/698 [==============================] - 98s 141ms/step - loss: 2.4900 - Accuracy: 0.7580 - AD_Precision: 0.6833 - CN_Precision: 0.7388 - MCI_Precision: 0.7953 - AD_Recall: 0.6765 - CN_Recall: 0.7310 - MCI_Recall: 0.8031 - val_loss: 1.7916 - val_Accuracy: 0.7493 - val_AD_Precision: 0.5976 - val_CN_Precision: 0.6850 - val_MCI_Precision: 0.9011 - val_AD_Recall: 0.7953 - val_CN_Recall: 0.8832 - val_MCI_Recall: 0.6578\n",
            "Epoch 3/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 1.4012 - Accuracy: 0.8347 - AD_Precision: 0.7737 - CN_Precision: 0.8328 - MCI_Precision: 0.8579 - AD_Recall: 0.7554 - CN_Recall: 0.8386 - MCI_Recall: 0.8591 - val_loss: 4.1153 - val_Accuracy: 0.5759 - val_AD_Precision: 0.3230 - val_CN_Precision: 0.8797 - val_MCI_Precision: 0.9040 - val_AD_Recall: 0.9843 - val_CN_Recall: 0.5939 - val_MCI_Recall: 0.4278\n",
            "Epoch 4/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.9116 - Accuracy: 0.8627 - AD_Precision: 0.8066 - CN_Precision: 0.8625 - MCI_Precision: 0.8840 - AD_Recall: 0.8146 - CN_Recall: 0.8585 - MCI_Recall: 0.8816 - val_loss: 1.5607 - val_Accuracy: 0.8152 - val_AD_Precision: 0.9091 - val_CN_Precision: 0.7291 - val_MCI_Precision: 0.8556 - val_AD_Recall: 0.4724 - val_CN_Recall: 0.9289 - val_MCI_Recall: 0.8717\n",
            "Epoch 5/50\n",
            "698/698 [==============================] - 99s 141ms/step - loss: 0.5543 - Accuracy: 0.9061 - AD_Precision: 0.8668 - CN_Precision: 0.9170 - MCI_Precision: 0.9139 - AD_Recall: 0.8600 - CN_Recall: 0.9170 - MCI_Recall: 0.9145 - val_loss: 2.0253 - val_Accuracy: 0.7708 - val_AD_Precision: 0.8352 - val_CN_Precision: 0.9897 - val_MCI_Precision: 0.7191 - val_AD_Recall: 0.5984 - val_CN_Recall: 0.4873 - val_MCI_Recall: 0.9786\n",
            "Epoch 6/50\n",
            "698/698 [==============================] - 102s 147ms/step - loss: 0.5498 - Accuracy: 0.9068 - AD_Precision: 0.8802 - CN_Precision: 0.9081 - MCI_Precision: 0.9161 - AD_Recall: 0.8836 - CN_Recall: 0.9018 - MCI_Recall: 0.9180 - val_loss: 0.9897 - val_Accuracy: 0.8524 - val_AD_Precision: 0.9032 - val_CN_Precision: 0.8018 - val_MCI_Precision: 0.8691 - val_AD_Recall: 0.6614 - val_CN_Recall: 0.9036 - val_MCI_Recall: 0.8877\n",
            "Epoch 7/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.3267 - Accuracy: 0.9326 - AD_Precision: 0.9098 - CN_Precision: 0.9413 - MCI_Precision: 0.9366 - AD_Recall: 0.8955 - CN_Recall: 0.9380 - MCI_Recall: 0.9425 - val_loss: 0.9410 - val_Accuracy: 0.8395 - val_AD_Precision: 0.6686 - val_CN_Precision: 0.8962 - val_MCI_Precision: 0.8957 - val_AD_Recall: 0.8898 - val_CN_Recall: 0.8325 - val_MCI_Recall: 0.8262\n",
            "Epoch 8/50\n",
            "698/698 [==============================] - 101s 145ms/step - loss: 0.2563 - Accuracy: 0.9441 - AD_Precision: 0.9250 - CN_Precision: 0.9405 - MCI_Precision: 0.9543 - AD_Recall: 0.9250 - CN_Recall: 0.9427 - MCI_Recall: 0.9516 - val_loss: 0.9529 - val_Accuracy: 0.8610 - val_AD_Precision: 0.8203 - val_CN_Precision: 0.9716 - val_MCI_Precision: 0.8364 - val_AD_Recall: 0.8268 - val_CN_Recall: 0.6954 - val_MCI_Recall: 0.9572\n",
            "Epoch 9/50\n",
            "698/698 [==============================] - 102s 146ms/step - loss: 0.2303 - Accuracy: 0.9487 - AD_Precision: 0.9211 - CN_Precision: 0.9544 - MCI_Precision: 0.9552 - AD_Recall: 0.9211 - CN_Recall: 0.9544 - MCI_Recall: 0.9552 - val_loss: 0.7171 - val_Accuracy: 0.8682 - val_AD_Precision: 0.7465 - val_CN_Precision: 0.8531 - val_MCI_Precision: 0.9271 - val_AD_Recall: 0.8346 - val_CN_Recall: 0.9137 - val_MCI_Recall: 0.8503\n",
            "Epoch 10/50\n",
            "698/698 [==============================] - 98s 141ms/step - loss: 0.2321 - Accuracy: 0.9477 - AD_Precision: 0.9267 - CN_Precision: 0.9474 - MCI_Precision: 0.9559 - AD_Recall: 0.9231 - CN_Recall: 0.9474 - MCI_Recall: 0.9559 - val_loss: 1.0083 - val_Accuracy: 0.8009 - val_AD_Precision: 0.6818 - val_CN_Precision: 0.7033 - val_MCI_Precision: 0.9701 - val_AD_Recall: 0.8268 - val_CN_Recall: 0.9746 - val_MCI_Recall: 0.6952\n",
            "Epoch 11/50\n",
            "698/698 [==============================] - 102s 146ms/step - loss: 0.2345 - Accuracy: 0.9423 - AD_Precision: 0.9129 - CN_Precision: 0.9471 - MCI_Precision: 0.9497 - AD_Recall: 0.9093 - CN_Recall: 0.9427 - MCI_Recall: 0.9523 - val_loss: 0.6960 - val_Accuracy: 0.8825 - val_AD_Precision: 0.9255 - val_CN_Precision: 0.8958 - val_MCI_Precision: 0.8707 - val_AD_Recall: 0.6850 - val_CN_Recall: 0.8731 - val_MCI_Recall: 0.9545\n",
            "Epoch 12/50\n",
            "698/698 [==============================] - 102s 146ms/step - loss: 0.1392 - Accuracy: 0.9591 - AD_Precision: 0.9370 - CN_Precision: 0.9670 - MCI_Precision: 0.9630 - AD_Recall: 0.9389 - CN_Recall: 0.9591 - MCI_Recall: 0.9657 - val_loss: 0.4985 - val_Accuracy: 0.8997 - val_AD_Precision: 0.8254 - val_CN_Precision: 0.9543 - val_MCI_Precision: 0.8992 - val_AD_Recall: 0.8189 - val_CN_Recall: 0.8477 - val_MCI_Recall: 0.9545\n",
            "Epoch 13/50\n",
            "698/698 [==============================] - 100s 144ms/step - loss: 0.1157 - Accuracy: 0.9652 - AD_Precision: 0.9544 - CN_Precision: 0.9652 - MCI_Precision: 0.9711 - AD_Recall: 0.9487 - CN_Recall: 0.9719 - MCI_Recall: 0.9664 - val_loss: 0.5223 - val_Accuracy: 0.8968 - val_AD_Precision: 0.9159 - val_CN_Precision: 0.9021 - val_MCI_Precision: 0.8892 - val_AD_Recall: 0.7717 - val_CN_Recall: 0.8883 - val_MCI_Recall: 0.9439\n",
            "Epoch 14/50\n",
            "698/698 [==============================] - 102s 145ms/step - loss: 0.0905 - Accuracy: 0.9724 - AD_Precision: 0.9586 - CN_Precision: 0.9765 - MCI_Precision: 0.9755 - AD_Recall: 0.9586 - CN_Recall: 0.9731 - MCI_Recall: 0.9762 - val_loss: 0.4610 - val_Accuracy: 0.9054 - val_AD_Precision: 0.9266 - val_CN_Precision: 0.9305 - val_MCI_Precision: 0.8925 - val_AD_Recall: 0.7953 - val_CN_Recall: 0.8832 - val_MCI_Recall: 0.9545\n",
            "Epoch 15/50\n",
            "698/698 [==============================] - 99s 142ms/step - loss: 0.1146 - Accuracy: 0.9688 - AD_Precision: 0.9605 - CN_Precision: 0.9695 - MCI_Precision: 0.9734 - AD_Recall: 0.9586 - CN_Recall: 0.9661 - MCI_Recall: 0.9741 - val_loss: 0.5239 - val_Accuracy: 0.8797 - val_AD_Precision: 0.7466 - val_CN_Precision: 0.9441 - val_MCI_Precision: 0.9106 - val_AD_Recall: 0.8583 - val_CN_Recall: 0.8579 - val_MCI_Recall: 0.8984\n",
            "Epoch 16/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0808 - Accuracy: 0.9724 - AD_Precision: 0.9569 - CN_Precision: 0.9732 - MCI_Precision: 0.9782 - AD_Recall: 0.9625 - CN_Recall: 0.9766 - MCI_Recall: 0.9734 - val_loss: 0.9224 - val_Accuracy: 0.8438 - val_AD_Precision: 0.9878 - val_CN_Precision: 0.9645 - val_MCI_Precision: 0.7848 - val_AD_Recall: 0.6378 - val_CN_Recall: 0.6904 - val_MCI_Recall: 0.9947\n",
            "Epoch 17/50\n",
            "698/698 [==============================] - 98s 141ms/step - loss: 0.0787 - Accuracy: 0.9720 - AD_Precision: 0.9664 - CN_Precision: 0.9719 - MCI_Precision: 0.9741 - AD_Recall: 0.9645 - CN_Recall: 0.9708 - MCI_Recall: 0.9755 - val_loss: 0.5668 - val_Accuracy: 0.8739 - val_AD_Precision: 0.9474 - val_CN_Precision: 0.8044 - val_MCI_Precision: 0.8989 - val_AD_Recall: 0.7087 - val_CN_Recall: 0.9188 - val_MCI_Recall: 0.9037\n",
            "Epoch 18/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0616 - Accuracy: 0.9828 - AD_Precision: 0.9744 - CN_Precision: 0.9848 - MCI_Precision: 0.9846 - AD_Recall: 0.9763 - CN_Recall: 0.9860 - MCI_Recall: 0.9832 - val_loss: 0.4258 - val_Accuracy: 0.8983 - val_AD_Precision: 0.8525 - val_CN_Precision: 0.9251 - val_MCI_Precision: 0.9044 - val_AD_Recall: 0.8189 - val_CN_Recall: 0.8782 - val_MCI_Recall: 0.9358\n",
            "Epoch 19/50\n",
            "698/698 [==============================] - 99s 141ms/step - loss: 0.0665 - Accuracy: 0.9788 - AD_Precision: 0.9743 - CN_Precision: 0.9812 - MCI_Precision: 0.9797 - AD_Recall: 0.9704 - CN_Recall: 0.9789 - MCI_Recall: 0.9818 - val_loss: 0.4943 - val_Accuracy: 0.8968 - val_AD_Precision: 0.8468 - val_CN_Precision: 0.9438 - val_MCI_Precision: 0.8914 - val_AD_Recall: 0.8268 - val_CN_Recall: 0.8528 - val_MCI_Recall: 0.9439\n",
            "Epoch 20/50\n",
            "698/698 [==============================] - 99s 141ms/step - loss: 0.0763 - Accuracy: 0.9742 - AD_Precision: 0.9684 - CN_Precision: 0.9730 - MCI_Precision: 0.9776 - AD_Recall: 0.9665 - CN_Recall: 0.9696 - MCI_Recall: 0.9790 - val_loss: 0.5739 - val_Accuracy: 0.8782 - val_AD_Precision: 0.9432 - val_CN_Precision: 0.8404 - val_MCI_Precision: 0.8838 - val_AD_Recall: 0.6535 - val_CN_Recall: 0.9086 - val_MCI_Recall: 0.9358\n",
            "Epoch 21/50\n",
            "698/698 [==============================] - 99s 141ms/step - loss: 0.0520 - Accuracy: 0.9817 - AD_Precision: 0.9862 - CN_Precision: 0.9766 - MCI_Precision: 0.9832 - AD_Recall: 0.9862 - CN_Recall: 0.9754 - MCI_Recall: 0.9839 - val_loss: 0.6873 - val_Accuracy: 0.8410 - val_AD_Precision: 0.8803 - val_CN_Precision: 0.7063 - val_MCI_Precision: 0.9513 - val_AD_Recall: 0.8110 - val_CN_Recall: 0.9645 - val_MCI_Recall: 0.7834\n",
            "Epoch 22/50\n",
            "698/698 [==============================] - 98s 141ms/step - loss: 0.0567 - Accuracy: 0.9792 - AD_Precision: 0.9644 - CN_Precision: 0.9825 - MCI_Precision: 0.9832 - AD_Recall: 0.9625 - CN_Recall: 0.9836 - MCI_Recall: 0.9818 - val_loss: 0.4937 - val_Accuracy: 0.8968 - val_AD_Precision: 0.9245 - val_CN_Precision: 0.8585 - val_MCI_Precision: 0.9153 - val_AD_Recall: 0.7717 - val_CN_Recall: 0.9239 - val_MCI_Recall: 0.9251\n",
            "Epoch 23/50\n",
            "698/698 [==============================] - 99s 141ms/step - loss: 0.0639 - Accuracy: 0.9828 - AD_Precision: 0.9842 - CN_Precision: 0.9790 - MCI_Precision: 0.9853 - AD_Recall: 0.9803 - CN_Recall: 0.9801 - MCI_Recall: 0.9846 - val_loss: 0.5292 - val_Accuracy: 0.8668 - val_AD_Precision: 0.6723 - val_CN_Precision: 0.9239 - val_MCI_Precision: 0.9429 - val_AD_Recall: 0.9370 - val_CN_Recall: 0.8629 - val_MCI_Recall: 0.8396\n",
            "Epoch 24/50\n",
            "698/698 [==============================] - 99s 142ms/step - loss: 0.0479 - Accuracy: 0.9831 - AD_Precision: 0.9743 - CN_Precision: 0.9871 - MCI_Precision: 0.9839 - AD_Recall: 0.9724 - CN_Recall: 0.9836 - MCI_Recall: 0.9867 - val_loss: 0.4735 - val_Accuracy: 0.8997 - val_AD_Precision: 0.8898 - val_CN_Precision: 0.8578 - val_MCI_Precision: 0.9313 - val_AD_Recall: 0.8268 - val_CN_Recall: 0.9188 - val_MCI_Recall: 0.9064\n",
            "Epoch 25/50\n",
            "698/698 [==============================] - 99s 141ms/step - loss: 0.0559 - Accuracy: 0.9799 - AD_Precision: 0.9745 - CN_Precision: 0.9801 - MCI_Precision: 0.9824 - AD_Recall: 0.9783 - CN_Recall: 0.9789 - MCI_Recall: 0.9804 - val_loss: 0.4368 - val_Accuracy: 0.9054 - val_AD_Precision: 0.8919 - val_CN_Precision: 0.9415 - val_MCI_Precision: 0.8945 - val_AD_Recall: 0.7795 - val_CN_Recall: 0.8985 - val_MCI_Recall: 0.9519\n",
            "Epoch 26/50\n",
            "698/698 [==============================] - 99s 141ms/step - loss: 0.0241 - Accuracy: 0.9900 - AD_Precision: 0.9842 - CN_Precision: 0.9941 - MCI_Precision: 0.9895 - AD_Recall: 0.9803 - CN_Recall: 0.9930 - MCI_Recall: 0.9916 - val_loss: 0.5947 - val_Accuracy: 0.8983 - val_AD_Precision: 0.9773 - val_CN_Precision: 0.8685 - val_MCI_Precision: 0.8967 - val_AD_Recall: 0.6772 - val_CN_Recall: 0.9391 - val_MCI_Recall: 0.9519\n",
            "Epoch 27/50\n",
            "698/698 [==============================] - 99s 142ms/step - loss: 0.0179 - Accuracy: 0.9939 - AD_Precision: 0.9941 - CN_Precision: 0.9942 - MCI_Precision: 0.9944 - AD_Recall: 0.9901 - CN_Recall: 0.9965 - MCI_Recall: 0.9937 - val_loss: 0.5235 - val_Accuracy: 0.8940 - val_AD_Precision: 0.7671 - val_CN_Precision: 0.8947 - val_MCI_Precision: 0.9501 - val_AD_Recall: 0.8819 - val_CN_Recall: 0.9492 - val_MCI_Recall: 0.8663\n",
            "Epoch 28/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0610 - Accuracy: 0.9839 - AD_Precision: 0.9761 - CN_Precision: 0.9871 - MCI_Precision: 0.9853 - AD_Recall: 0.9684 - CN_Recall: 0.9860 - MCI_Recall: 0.9874 - val_loss: 0.7304 - val_Accuracy: 0.8668 - val_AD_Precision: 0.8083 - val_CN_Precision: 0.7890 - val_MCI_Precision: 0.9441 - val_AD_Recall: 0.7638 - val_CN_Recall: 0.9492 - val_MCI_Recall: 0.8583\n",
            "Epoch 29/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0589 - Accuracy: 0.9824 - AD_Precision: 0.9801 - CN_Precision: 0.9836 - MCI_Precision: 0.9832 - AD_Recall: 0.9724 - CN_Recall: 0.9836 - MCI_Recall: 0.9853 - val_loss: 0.4830 - val_Accuracy: 0.8868 - val_AD_Precision: 0.7358 - val_CN_Precision: 0.9016 - val_MCI_Precision: 0.9507 - val_AD_Recall: 0.9213 - val_CN_Recall: 0.8832 - val_MCI_Recall: 0.8770\n",
            "Epoch 30/50\n",
            "698/698 [==============================] - 102s 146ms/step - loss: 0.0344 - Accuracy: 0.9864 - AD_Precision: 0.9763 - CN_Precision: 0.9906 - MCI_Precision: 0.9874 - AD_Recall: 0.9744 - CN_Recall: 0.9883 - MCI_Recall: 0.9895 - val_loss: 0.5294 - val_Accuracy: 0.9126 - val_AD_Precision: 0.9596 - val_CN_Precision: 0.9412 - val_MCI_Precision: 0.8883 - val_AD_Recall: 0.7480 - val_CN_Recall: 0.8934 - val_MCI_Recall: 0.9786\n",
            "Epoch 31/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.0412 - Accuracy: 0.9857 - AD_Precision: 0.9783 - CN_Precision: 0.9871 - MCI_Precision: 0.9874 - AD_Recall: 0.9783 - CN_Recall: 0.9883 - MCI_Recall: 0.9867 - val_loss: 0.9112 - val_Accuracy: 0.8252 - val_AD_Precision: 0.6011 - val_CN_Precision: 0.9864 - val_MCI_Precision: 0.8770 - val_AD_Recall: 0.8661 - val_CN_Recall: 0.7360 - val_MCI_Recall: 0.8583\n",
            "Epoch 32/50\n",
            "698/698 [==============================] - 97s 138ms/step - loss: 0.0338 - Accuracy: 0.9878 - AD_Precision: 0.9842 - CN_Precision: 0.9906 - MCI_Precision: 0.9888 - AD_Recall: 0.9842 - CN_Recall: 0.9906 - MCI_Recall: 0.9874 - val_loss: 0.5759 - val_Accuracy: 0.9026 - val_AD_Precision: 0.9882 - val_CN_Precision: 0.9278 - val_MCI_Precision: 0.8732 - val_AD_Recall: 0.6614 - val_CN_Recall: 0.9137 - val_MCI_Recall: 0.9759\n",
            "Epoch 33/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0450 - Accuracy: 0.9842 - AD_Precision: 0.9763 - CN_Precision: 0.9836 - MCI_Precision: 0.9874 - AD_Recall: 0.9744 - CN_Recall: 0.9836 - MCI_Recall: 0.9874 - val_loss: 0.4634 - val_Accuracy: 0.8997 - val_AD_Precision: 0.7847 - val_CN_Precision: 0.8986 - val_MCI_Precision: 0.9564 - val_AD_Recall: 0.8898 - val_CN_Recall: 0.9442 - val_MCI_Recall: 0.8797\n",
            "Epoch 34/50\n",
            "698/698 [==============================] - 100s 144ms/step - loss: 0.0226 - Accuracy: 0.9946 - AD_Precision: 0.9921 - CN_Precision: 0.9976 - MCI_Precision: 0.9937 - AD_Recall: 0.9921 - CN_Recall: 0.9930 - MCI_Recall: 0.9965 - val_loss: 0.2990 - val_Accuracy: 0.9284 - val_AD_Precision: 0.9000 - val_CN_Precision: 0.9628 - val_MCI_Precision: 0.9229 - val_AD_Recall: 0.8504 - val_CN_Recall: 0.9188 - val_MCI_Recall: 0.9599\n",
            "Epoch 35/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.0397 - Accuracy: 0.9849 - AD_Precision: 0.9763 - CN_Precision: 0.9895 - MCI_Precision: 0.9860 - AD_Recall: 0.9744 - CN_Recall: 0.9906 - MCI_Recall: 0.9853 - val_loss: 0.3838 - val_Accuracy: 0.9241 - val_AD_Precision: 0.8417 - val_CN_Precision: 0.9340 - val_MCI_Precision: 0.9503 - val_AD_Recall: 0.9213 - val_CN_Recall: 0.9340 - val_MCI_Recall: 0.9198\n",
            "Epoch 36/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0572 - Accuracy: 0.9839 - AD_Precision: 0.9784 - CN_Precision: 0.9801 - MCI_Precision: 0.9887 - AD_Recall: 0.9842 - CN_Recall: 0.9789 - MCI_Recall: 0.9853 - val_loss: 0.5369 - val_Accuracy: 0.9026 - val_AD_Precision: 0.9450 - val_CN_Precision: 0.8440 - val_MCI_Precision: 0.9268 - val_AD_Recall: 0.8110 - val_CN_Recall: 0.9340 - val_MCI_Recall: 0.9144\n",
            "Epoch 37/50\n",
            "698/698 [==============================] - 97s 140ms/step - loss: 0.0490 - Accuracy: 0.9839 - AD_Precision: 0.9823 - CN_Precision: 0.9801 - MCI_Precision: 0.9874 - AD_Recall: 0.9842 - CN_Recall: 0.9789 - MCI_Recall: 0.9867 - val_loss: 0.5469 - val_Accuracy: 0.8968 - val_AD_Precision: 0.9574 - val_CN_Precision: 0.9709 - val_MCI_Precision: 0.8542 - val_AD_Recall: 0.7087 - val_CN_Recall: 0.8477 - val_MCI_Recall: 0.9866\n",
            "Epoch 38/50\n",
            "698/698 [==============================] - 99s 141ms/step - loss: 0.0159 - Accuracy: 0.9946 - AD_Precision: 0.9941 - CN_Precision: 0.9942 - MCI_Precision: 0.9951 - AD_Recall: 0.9901 - CN_Recall: 0.9965 - MCI_Recall: 0.9951 - val_loss: 0.4218 - val_Accuracy: 0.9097 - val_AD_Precision: 0.9412 - val_CN_Precision: 0.9365 - val_MCI_Precision: 0.8894 - val_AD_Recall: 0.7559 - val_CN_Recall: 0.8985 - val_MCI_Recall: 0.9679\n",
            "Epoch 39/50\n",
            "698/698 [==============================] - 99s 141ms/step - loss: 0.0469 - Accuracy: 0.9871 - AD_Precision: 0.9862 - CN_Precision: 0.9837 - MCI_Precision: 0.9895 - AD_Recall: 0.9862 - CN_Recall: 0.9860 - MCI_Recall: 0.9874 - val_loss: 0.4178 - val_Accuracy: 0.9140 - val_AD_Precision: 0.9400 - val_CN_Precision: 0.9323 - val_MCI_Precision: 0.8988 - val_AD_Recall: 0.7402 - val_CN_Recall: 0.9086 - val_MCI_Recall: 0.9733\n",
            "Epoch 40/50\n",
            "698/698 [==============================] - 98s 141ms/step - loss: 0.0151 - Accuracy: 0.9950 - AD_Precision: 0.9941 - CN_Precision: 0.9942 - MCI_Precision: 0.9958 - AD_Recall: 0.9941 - CN_Recall: 0.9953 - MCI_Recall: 0.9951 - val_loss: 0.7596 - val_Accuracy: 0.8883 - val_AD_Precision: 0.9369 - val_CN_Precision: 0.9801 - val_MCI_Precision: 0.8460 - val_AD_Recall: 0.8189 - val_CN_Recall: 0.7513 - val_MCI_Recall: 0.9840\n",
            "Epoch 41/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.0566 - Accuracy: 0.9842 - AD_Precision: 0.9802 - CN_Precision: 0.9836 - MCI_Precision: 0.9860 - AD_Recall: 0.9783 - CN_Recall: 0.9836 - MCI_Recall: 0.9867 - val_loss: 1.0602 - val_Accuracy: 0.8467 - val_AD_Precision: 0.9182 - val_CN_Precision: 1.0000 - val_MCI_Precision: 0.7910 - val_AD_Recall: 0.7953 - val_CN_Recall: 0.5990 - val_MCI_Recall: 0.9920\n",
            "Epoch 42/50\n",
            "698/698 [==============================] - 99s 141ms/step - loss: 0.0324 - Accuracy: 0.9900 - AD_Precision: 0.9901 - CN_Precision: 0.9894 - MCI_Precision: 0.9902 - AD_Recall: 0.9901 - CN_Recall: 0.9860 - MCI_Recall: 0.9916 - val_loss: 0.6479 - val_Accuracy: 0.8897 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.9529 - val_MCI_Precision: 0.8451 - val_AD_Recall: 0.6929 - val_CN_Recall: 0.8223 - val_MCI_Recall: 0.9920\n",
            "Epoch 43/50\n",
            "698/698 [==============================] - 99s 141ms/step - loss: 0.0222 - Accuracy: 0.9921 - AD_Precision: 0.9882 - CN_Precision: 0.9918 - MCI_Precision: 0.9937 - AD_Recall: 0.9901 - CN_Recall: 0.9918 - MCI_Recall: 0.9930 - val_loss: 0.4445 - val_Accuracy: 0.9198 - val_AD_Precision: 0.8957 - val_CN_Precision: 0.9289 - val_MCI_Precision: 0.9221 - val_AD_Recall: 0.8110 - val_CN_Recall: 0.9289 - val_MCI_Recall: 0.9492\n",
            "Epoch 44/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.0259 - Accuracy: 0.9903 - AD_Precision: 0.9901 - CN_Precision: 0.9883 - MCI_Precision: 0.9916 - AD_Recall: 0.9862 - CN_Recall: 0.9906 - MCI_Recall: 0.9916 - val_loss: 0.4846 - val_Accuracy: 0.8997 - val_AD_Precision: 0.7793 - val_CN_Precision: 0.9095 - val_MCI_Precision: 0.9462 - val_AD_Recall: 0.8898 - val_CN_Recall: 0.9188 - val_MCI_Recall: 0.8930\n",
            "Epoch 45/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.0296 - Accuracy: 0.9928 - AD_Precision: 0.9862 - CN_Precision: 0.9965 - MCI_Precision: 0.9930 - AD_Recall: 0.9862 - CN_Recall: 0.9965 - MCI_Recall: 0.9930 - val_loss: 0.7558 - val_Accuracy: 0.8725 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.9175 - val_MCI_Precision: 0.8341 - val_AD_Recall: 0.5039 - val_CN_Recall: 0.9036 - val_MCI_Recall: 0.9813\n",
            "Epoch 46/50\n",
            "698/698 [==============================] - 98s 141ms/step - loss: 0.0323 - Accuracy: 0.9892 - AD_Precision: 0.9802 - CN_Precision: 0.9918 - MCI_Precision: 0.9909 - AD_Recall: 0.9783 - CN_Recall: 0.9883 - MCI_Recall: 0.9937 - val_loss: 0.6259 - val_Accuracy: 0.8954 - val_AD_Precision: 1.0000 - val_CN_Precision: 0.9215 - val_MCI_Precision: 0.8656 - val_AD_Recall: 0.6457 - val_CN_Recall: 0.8934 - val_MCI_Recall: 0.9813\n",
            "Epoch 47/50\n",
            "698/698 [==============================] - 99s 141ms/step - loss: 0.0308 - Accuracy: 0.9889 - AD_Precision: 0.9861 - CN_Precision: 0.9895 - MCI_Precision: 0.9895 - AD_Recall: 0.9803 - CN_Recall: 0.9906 - MCI_Recall: 0.9909 - val_loss: 0.4867 - val_Accuracy: 0.9241 - val_AD_Precision: 0.9608 - val_CN_Precision: 0.9100 - val_MCI_Precision: 0.9217 - val_AD_Recall: 0.7717 - val_CN_Recall: 0.9239 - val_MCI_Recall: 0.9759\n",
            "Epoch 48/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.0288 - Accuracy: 0.9914 - AD_Precision: 0.9941 - CN_Precision: 0.9883 - MCI_Precision: 0.9923 - AD_Recall: 0.9901 - CN_Recall: 0.9906 - MCI_Recall: 0.9923 - val_loss: 0.4804 - val_Accuracy: 0.9126 - val_AD_Precision: 0.7647 - val_CN_Precision: 0.9482 - val_MCI_Precision: 0.9574 - val_AD_Recall: 0.9213 - val_CN_Recall: 0.9289 - val_MCI_Recall: 0.9011\n",
            "Epoch 49/50\n",
            "698/698 [==============================] - 101s 145ms/step - loss: 0.0403 - Accuracy: 0.9871 - AD_Precision: 0.9802 - CN_Precision: 0.9848 - MCI_Precision: 0.9909 - AD_Recall: 0.9763 - CN_Recall: 0.9883 - MCI_Recall: 0.9895 - val_loss: 0.4362 - val_Accuracy: 0.9341 - val_AD_Precision: 0.8267 - val_CN_Precision: 0.9779 - val_MCI_Precision: 0.9563 - val_AD_Recall: 0.9764 - val_CN_Recall: 0.8985 - val_MCI_Recall: 0.9358\n",
            "Epoch 50/50\n",
            "698/698 [==============================] - 100s 143ms/step - loss: 0.0238 - Accuracy: 0.9925 - AD_Precision: 0.9921 - CN_Precision: 0.9918 - MCI_Precision: 0.9930 - AD_Recall: 0.9901 - CN_Recall: 0.9918 - MCI_Recall: 0.9937 - val_loss: 0.9442 - val_Accuracy: 0.8639 - val_AD_Precision: 0.9474 - val_CN_Precision: 0.9427 - val_MCI_Precision: 0.8184 - val_AD_Recall: 0.7087 - val_CN_Recall: 0.7513 - val_MCI_Recall: 0.9759\n"
          ]
        }
      ],
      "source": [
        "# learning_rate=10E-5\n",
        "savepath = '/content/drive/MyDrive/ADNI/resnet50_row_10E-5/'\n",
        "lr=0.00001\n",
        "if not os.path.isdir(savepath):\n",
        "      os.makedirs(savepath)\n",
        "csvloger_callback = tf.keras.callbacks.CSVLogger(f'{savepath}history.csv', separator=',', append=False)\n",
        "\n",
        "modelcheckpoint_callback = tf.keras.callbacks.ModelCheckpoint(savepath+'resnet50_row_10E-5_weights-epoch:{epoch:03d}-val_Accuracy:{val_Accuracy:.4f}.h5',monitor='val_Accuracy'\n",
        "    ,verbose=0,save_best_only=True,save_weights_only=True,mode='max',save_freq='epoch' , initial_value_threshold=0.85)\n",
        "\n",
        "OPT = tf.keras.optimizers.Adam(learning_rate=lr)\n",
        "model05 = resnet50_row(OPT, LOOS,METRICS,image_Shape)\n",
        "history05 = model05.fit(training_image_datagen, epochs=EPOCH, validation_data=testing_image_datagen,callbacks=[csvloger_callback,modelcheckpoint_callback])\n",
        "model05 = resnet50_row(OPT, LOOS,METRICS,image_Shape)\n",
        "model05.load_weights(glob.glob(f\"{savepath}resnet50_row_10E-5_weights-epoch:*-val_Accuracy:{max(history05.history['val_Accuracy']):.4f}.h5\")[0])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 477
        },
        "id": "-CVZ5XfIDQeq",
        "outputId": "718a237a-5b6e-4513-91c9-3bc51a0b1738"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "          AD       0.83      0.98      0.90       127\n",
            "          CN       0.98      0.90      0.94       197\n",
            "         MCI       0.96      0.94      0.95       374\n",
            "\n",
            "    accuracy                           0.93       698\n",
            "   macro avg       0.92      0.94      0.93       698\n",
            "weighted avg       0.94      0.93      0.93       698\n",
            "\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "p = model05.predict(testing_image_datagen)\n",
        "predicted_lab = p.argmax(axis=-1)\n",
        "true_lab = testing_image_datagen.classes\n",
        "plot_confusion_matrix(true_lab, predicted_lab, testing_image_datagen.class_indices.keys(),savepath)\n",
        "report = classification_report(true_lab, predicted_lab, target_names=testing_image_datagen.class_indices.keys())\n",
        "print(report)\n",
        "dump_text(report,f'{savepath}classification_report.txt')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yVLEEUCXDOsQ",
        "outputId": "e8ad5c51-164d-4eec-e642-cb365f8f3b86"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "698/698 [==============================] - 105s 141ms/step - loss: 2.8186 - Accuracy: 0.5369 - AD_Precision: 0.3495 - CN_Precision: 0.4885 - MCI_Precision: 0.6233 - AD_Recall: 0.3060 - CN_Recall: 0.4648 - MCI_Recall: 0.6541 - val_loss: 1.8575 - val_Accuracy: 0.5473 - val_AD_Precision: 0.4138 - val_CN_Precision: 0.4489 - val_MCI_Precision: 0.6571 - val_AD_Recall: 0.1890 - val_CN_Recall: 0.6244 - val_MCI_Recall: 0.6096\n",
            "Epoch 2/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 1.5280 - Accuracy: 0.6432 - AD_Precision: 0.4882 - CN_Precision: 0.6211 - MCI_Precision: 0.7138 - AD_Recall: 0.4477 - CN_Recall: 0.6117 - MCI_Recall: 0.7148 - val_loss: 1.3118 - val_Accuracy: 0.6877 - val_AD_Precision: 0.6329 - val_CN_Precision: 0.6890 - val_MCI_Precision: 0.6975 - val_AD_Recall: 0.3937 - val_CN_Recall: 0.5736 - val_MCI_Recall: 0.8262\n",
            "Epoch 3/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 1.1136 - Accuracy: 0.7401 - AD_Precision: 0.6203 - CN_Precision: 0.7352 - MCI_Precision: 0.7915 - AD_Recall: 0.5897 - CN_Recall: 0.7240 - MCI_Recall: 0.7877 - val_loss: 1.0185 - val_Accuracy: 0.7120 - val_AD_Precision: 0.5594 - val_CN_Precision: 0.6808 - val_MCI_Precision: 0.8190 - val_AD_Recall: 0.6299 - val_CN_Recall: 0.7360 - val_MCI_Recall: 0.7139\n",
            "Epoch 4/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.7295 - Accuracy: 0.8006 - AD_Precision: 0.7636 - CN_Precision: 0.7849 - MCI_Precision: 0.8301 - AD_Recall: 0.7456 - CN_Recall: 0.7637 - MCI_Recall: 0.8353 - val_loss: 1.2255 - val_Accuracy: 0.7378 - val_AD_Precision: 0.8800 - val_CN_Precision: 0.8138 - val_MCI_Precision: 0.7048 - val_AD_Recall: 0.3465 - val_CN_Recall: 0.5990 - val_MCI_Recall: 0.9385\n",
            "Epoch 5/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.5481 - Accuracy: 0.8551 - AD_Precision: 0.8044 - CN_Precision: 0.8470 - MCI_Precision: 0.8839 - AD_Recall: 0.7870 - CN_Recall: 0.8351 - MCI_Recall: 0.8858 - val_loss: 0.7407 - val_Accuracy: 0.7951 - val_AD_Precision: 0.7596 - val_CN_Precision: 0.7812 - val_MCI_Precision: 0.8231 - val_AD_Recall: 0.6220 - val_CN_Recall: 0.7614 - val_MCI_Recall: 0.8583\n",
            "Epoch 6/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.4553 - Accuracy: 0.8788 - AD_Precision: 0.8395 - CN_Precision: 0.8766 - MCI_Precision: 0.8983 - AD_Recall: 0.8047 - CN_Recall: 0.8725 - MCI_Recall: 0.9040 - val_loss: 0.7871 - val_Accuracy: 0.7894 - val_AD_Precision: 0.5954 - val_CN_Precision: 0.9220 - val_MCI_Precision: 0.8346 - val_AD_Recall: 0.8110 - val_CN_Recall: 0.6599 - val_MCI_Recall: 0.8503\n",
            "Epoch 7/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.3756 - Accuracy: 0.8928 - AD_Precision: 0.8657 - CN_Precision: 0.8857 - MCI_Precision: 0.9106 - AD_Recall: 0.8521 - CN_Recall: 0.8795 - MCI_Recall: 0.9131 - val_loss: 0.8598 - val_Accuracy: 0.7994 - val_AD_Precision: 0.6463 - val_CN_Precision: 0.9147 - val_MCI_Precision: 0.8243 - val_AD_Recall: 0.8346 - val_CN_Recall: 0.5990 - val_MCI_Recall: 0.8904\n",
            "Epoch 8/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.2726 - Accuracy: 0.9154 - AD_Precision: 0.8924 - CN_Precision: 0.9107 - MCI_Precision: 0.9286 - AD_Recall: 0.8836 - CN_Recall: 0.9064 - MCI_Recall: 0.9299 - val_loss: 0.7548 - val_Accuracy: 0.8166 - val_AD_Precision: 0.7876 - val_CN_Precision: 0.8816 - val_MCI_Precision: 0.8042 - val_AD_Recall: 0.7008 - val_CN_Recall: 0.6802 - val_MCI_Recall: 0.9225\n",
            "Epoch 9/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.2563 - Accuracy: 0.9272 - AD_Precision: 0.9094 - CN_Precision: 0.9206 - MCI_Precision: 0.9420 - AD_Recall: 0.9112 - CN_Recall: 0.9216 - MCI_Recall: 0.9341 - val_loss: 0.5986 - val_Accuracy: 0.8410 - val_AD_Precision: 0.8136 - val_CN_Precision: 0.8274 - val_MCI_Precision: 0.8690 - val_AD_Recall: 0.7559 - val_CN_Recall: 0.8274 - val_MCI_Recall: 0.8690\n",
            "Epoch 10/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.1827 - Accuracy: 0.9394 - AD_Precision: 0.9246 - CN_Precision: 0.9379 - MCI_Precision: 0.9475 - AD_Recall: 0.9191 - CN_Recall: 0.9368 - MCI_Recall: 0.9481 - val_loss: 0.7261 - val_Accuracy: 0.8410 - val_AD_Precision: 0.8485 - val_CN_Precision: 0.8922 - val_MCI_Precision: 0.8255 - val_AD_Recall: 0.6614 - val_CN_Recall: 0.7563 - val_MCI_Recall: 0.9358\n",
            "Epoch 11/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.1562 - Accuracy: 0.9523 - AD_Precision: 0.9360 - CN_Precision: 0.9476 - MCI_Precision: 0.9628 - AD_Recall: 0.9231 - CN_Recall: 0.9509 - MCI_Recall: 0.9615 - val_loss: 0.7571 - val_Accuracy: 0.8352 - val_AD_Precision: 0.9481 - val_CN_Precision: 0.8449 - val_MCI_Precision: 0.8205 - val_AD_Recall: 0.5748 - val_CN_Recall: 0.8020 - val_MCI_Recall: 0.9412\n",
            "Epoch 12/50\n",
            "698/698 [==============================] - 101s 145ms/step - loss: 0.1426 - Accuracy: 0.9598 - AD_Precision: 0.9426 - CN_Precision: 0.9636 - MCI_Precision: 0.9643 - AD_Recall: 0.9389 - CN_Recall: 0.9591 - MCI_Recall: 0.9657 - val_loss: 0.6277 - val_Accuracy: 0.8510 - val_AD_Precision: 0.7966 - val_CN_Precision: 0.8048 - val_MCI_Precision: 0.8940 - val_AD_Recall: 0.7402 - val_CN_Recall: 0.8579 - val_MCI_Recall: 0.8797\n",
            "Epoch 13/50\n",
            "698/698 [==============================] - 101s 145ms/step - loss: 0.1358 - Accuracy: 0.9555 - AD_Precision: 0.9482 - CN_Precision: 0.9487 - MCI_Precision: 0.9629 - AD_Recall: 0.9389 - CN_Recall: 0.9520 - MCI_Recall: 0.9629 - val_loss: 0.6001 - val_Accuracy: 0.8596 - val_AD_Precision: 0.8462 - val_CN_Precision: 0.7937 - val_MCI_Precision: 0.9101 - val_AD_Recall: 0.7795 - val_CN_Recall: 0.8985 - val_MCI_Recall: 0.8663\n",
            "Epoch 14/50\n",
            "698/698 [==============================] - 97s 140ms/step - loss: 0.0945 - Accuracy: 0.9706 - AD_Precision: 0.9622 - CN_Precision: 0.9674 - MCI_Precision: 0.9768 - AD_Recall: 0.9546 - CN_Recall: 0.9708 - MCI_Recall: 0.9755 - val_loss: 0.6884 - val_Accuracy: 0.8582 - val_AD_Precision: 0.7895 - val_CN_Precision: 0.9241 - val_MCI_Precision: 0.8571 - val_AD_Recall: 0.8268 - val_CN_Recall: 0.7411 - val_MCI_Recall: 0.9305\n",
            "Epoch 15/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0976 - Accuracy: 0.9670 - AD_Precision: 0.9562 - CN_Precision: 0.9671 - MCI_Precision: 0.9720 - AD_Recall: 0.9467 - CN_Recall: 0.9637 - MCI_Recall: 0.9748 - val_loss: 1.0135 - val_Accuracy: 0.8324 - val_AD_Precision: 0.9851 - val_CN_Precision: 0.8800 - val_MCI_Precision: 0.7917 - val_AD_Recall: 0.5197 - val_CN_Recall: 0.7817 - val_MCI_Recall: 0.9652\n",
            "Epoch 16/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0988 - Accuracy: 0.9713 - AD_Precision: 0.9624 - CN_Precision: 0.9707 - MCI_Precision: 0.9748 - AD_Recall: 0.9586 - CN_Recall: 0.9684 - MCI_Recall: 0.9762 - val_loss: 1.4593 - val_Accuracy: 0.7851 - val_AD_Precision: 0.9811 - val_CN_Precision: 0.8742 - val_MCI_Precision: 0.7378 - val_AD_Recall: 0.4094 - val_CN_Recall: 0.6701 - val_MCI_Recall: 0.9706\n",
            "Epoch 17/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.1113 - Accuracy: 0.9677 - AD_Precision: 0.9562 - CN_Precision: 0.9627 - MCI_Precision: 0.9755 - AD_Recall: 0.9467 - CN_Recall: 0.9673 - MCI_Recall: 0.9748 - val_loss: 0.8316 - val_Accuracy: 0.8352 - val_AD_Precision: 0.8866 - val_CN_Precision: 0.7233 - val_MCI_Precision: 0.9041 - val_AD_Recall: 0.6772 - val_CN_Recall: 0.9289 - val_MCI_Recall: 0.8316\n",
            "Epoch 18/50\n",
            "698/698 [==============================] - 101s 145ms/step - loss: 0.0907 - Accuracy: 0.9728 - AD_Precision: 0.9647 - CN_Precision: 0.9672 - MCI_Precision: 0.9789 - AD_Recall: 0.9704 - CN_Recall: 0.9661 - MCI_Recall: 0.9762 - val_loss: 0.7325 - val_Accuracy: 0.8639 - val_AD_Precision: 0.8716 - val_CN_Precision: 0.8785 - val_MCI_Precision: 0.8575 - val_AD_Recall: 0.7480 - val_CN_Recall: 0.8071 - val_MCI_Recall: 0.9332\n",
            "Epoch 19/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.0848 - Accuracy: 0.9760 - AD_Precision: 0.9724 - CN_Precision: 0.9742 - MCI_Precision: 0.9790 - AD_Recall: 0.9724 - CN_Recall: 0.9719 - MCI_Recall: 0.9783 - val_loss: 0.7856 - val_Accuracy: 0.8567 - val_AD_Precision: 0.8807 - val_CN_Precision: 0.7573 - val_MCI_Precision: 0.9164 - val_AD_Recall: 0.7559 - val_CN_Recall: 0.9188 - val_MCI_Recall: 0.8503\n",
            "Epoch 20/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0538 - Accuracy: 0.9821 - AD_Precision: 0.9762 - CN_Precision: 0.9836 - MCI_Precision: 0.9853 - AD_Recall: 0.9704 - CN_Recall: 0.9813 - MCI_Recall: 0.9853 - val_loss: 0.9564 - val_Accuracy: 0.8410 - val_AD_Precision: 0.9767 - val_CN_Precision: 0.9038 - val_MCI_Precision: 0.7934 - val_AD_Recall: 0.6614 - val_CN_Recall: 0.7157 - val_MCI_Recall: 0.9652\n",
            "Epoch 21/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0721 - Accuracy: 0.9778 - AD_Precision: 0.9821 - CN_Precision: 0.9730 - MCI_Precision: 0.9790 - AD_Recall: 0.9763 - CN_Recall: 0.9708 - MCI_Recall: 0.9818 - val_loss: 0.7703 - val_Accuracy: 0.8610 - val_AD_Precision: 0.9457 - val_CN_Precision: 0.8131 - val_MCI_Precision: 0.8718 - val_AD_Recall: 0.6850 - val_CN_Recall: 0.8832 - val_MCI_Recall: 0.9091\n",
            "Epoch 22/50\n",
            "698/698 [==============================] - 101s 144ms/step - loss: 0.0703 - Accuracy: 0.9806 - AD_Precision: 0.9802 - CN_Precision: 0.9789 - MCI_Precision: 0.9825 - AD_Recall: 0.9783 - CN_Recall: 0.9754 - MCI_Recall: 0.9839 - val_loss: 0.6892 - val_Accuracy: 0.8725 - val_AD_Precision: 0.9099 - val_CN_Precision: 0.9012 - val_MCI_Precision: 0.8506 - val_AD_Recall: 0.7953 - val_CN_Recall: 0.7868 - val_MCI_Recall: 0.9439\n",
            "Epoch 23/50\n",
            "698/698 [==============================] - 97s 140ms/step - loss: 0.0453 - Accuracy: 0.9857 - AD_Precision: 0.9822 - CN_Precision: 0.9860 - MCI_Precision: 0.9867 - AD_Recall: 0.9783 - CN_Recall: 0.9871 - MCI_Recall: 0.9874 - val_loss: 0.7692 - val_Accuracy: 0.8582 - val_AD_Precision: 0.8015 - val_CN_Precision: 0.8283 - val_MCI_Precision: 0.8940 - val_AD_Recall: 0.8268 - val_CN_Recall: 0.8325 - val_MCI_Recall: 0.8797\n",
            "Epoch 24/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0587 - Accuracy: 0.9803 - AD_Precision: 0.9782 - CN_Precision: 0.9790 - MCI_Precision: 0.9825 - AD_Recall: 0.9724 - CN_Recall: 0.9801 - MCI_Recall: 0.9825 - val_loss: 1.3519 - val_Accuracy: 0.8223 - val_AD_Precision: 0.9870 - val_CN_Precision: 0.9489 - val_MCI_Precision: 0.7603 - val_AD_Recall: 0.5984 - val_CN_Recall: 0.6599 - val_MCI_Recall: 0.9840\n",
            "Epoch 25/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0458 - Accuracy: 0.9871 - AD_Precision: 0.9822 - CN_Precision: 0.9860 - MCI_Precision: 0.9895 - AD_Recall: 0.9822 - CN_Recall: 0.9860 - MCI_Recall: 0.9895 - val_loss: 0.8727 - val_Accuracy: 0.8639 - val_AD_Precision: 0.9216 - val_CN_Precision: 0.9152 - val_MCI_Precision: 0.8306 - val_AD_Recall: 0.7402 - val_CN_Recall: 0.7665 - val_MCI_Recall: 0.9572\n",
            "Epoch 26/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0511 - Accuracy: 0.9839 - AD_Precision: 0.9763 - CN_Precision: 0.9836 - MCI_Precision: 0.9867 - AD_Recall: 0.9763 - CN_Recall: 0.9848 - MCI_Recall: 0.9860 - val_loss: 1.0946 - val_Accuracy: 0.8539 - val_AD_Precision: 0.9881 - val_CN_Precision: 0.9096 - val_MCI_Precision: 0.8080 - val_AD_Recall: 0.6535 - val_CN_Recall: 0.7665 - val_MCI_Recall: 0.9679\n",
            "Epoch 27/50\n",
            "698/698 [==============================] - 97s 140ms/step - loss: 0.0417 - Accuracy: 0.9875 - AD_Precision: 0.9822 - CN_Precision: 0.9871 - MCI_Precision: 0.9902 - AD_Recall: 0.9803 - CN_Recall: 0.9848 - MCI_Recall: 0.9916 - val_loss: 0.8004 - val_Accuracy: 0.8625 - val_AD_Precision: 0.8911 - val_CN_Precision: 0.8840 - val_MCI_Precision: 0.8478 - val_AD_Recall: 0.7087 - val_CN_Recall: 0.8122 - val_MCI_Recall: 0.9385\n",
            "Epoch 28/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0560 - Accuracy: 0.9853 - AD_Precision: 0.9725 - CN_Precision: 0.9906 - MCI_Precision: 0.9874 - AD_Recall: 0.9763 - CN_Recall: 0.9871 - MCI_Recall: 0.9860 - val_loss: 1.4875 - val_Accuracy: 0.8037 - val_AD_Precision: 0.9815 - val_CN_Precision: 0.9221 - val_MCI_Precision: 0.7485 - val_AD_Recall: 0.4173 - val_CN_Recall: 0.7208 - val_MCI_Recall: 0.9786\n",
            "Epoch 29/50\n",
            "698/698 [==============================] - 101s 145ms/step - loss: 0.0526 - Accuracy: 0.9849 - AD_Precision: 0.9880 - CN_Precision: 0.9802 - MCI_Precision: 0.9867 - AD_Recall: 0.9783 - CN_Recall: 0.9836 - MCI_Recall: 0.9881 - val_loss: 0.8781 - val_Accuracy: 0.8768 - val_AD_Precision: 0.9468 - val_CN_Precision: 0.8842 - val_MCI_Precision: 0.8596 - val_AD_Recall: 0.7008 - val_CN_Recall: 0.8528 - val_MCI_Recall: 0.9492\n",
            "Epoch 30/50\n",
            "698/698 [==============================] - 101s 145ms/step - loss: 0.0352 - Accuracy: 0.9896 - AD_Precision: 0.9881 - CN_Precision: 0.9894 - MCI_Precision: 0.9902 - AD_Recall: 0.9822 - CN_Recall: 0.9871 - MCI_Recall: 0.9937 - val_loss: 0.7519 - val_Accuracy: 0.8782 - val_AD_Precision: 0.8772 - val_CN_Precision: 0.9176 - val_MCI_Precision: 0.8623 - val_AD_Recall: 0.7874 - val_CN_Recall: 0.7919 - val_MCI_Recall: 0.9545\n",
            "Epoch 31/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.0529 - Accuracy: 0.9839 - AD_Precision: 0.9843 - CN_Precision: 0.9812 - MCI_Precision: 0.9866 - AD_Recall: 0.9921 - CN_Recall: 0.9789 - MCI_Recall: 0.9839 - val_loss: 1.4169 - val_Accuracy: 0.8109 - val_AD_Precision: 0.9425 - val_CN_Precision: 0.9590 - val_MCI_Precision: 0.7536 - val_AD_Recall: 0.6457 - val_CN_Recall: 0.5939 - val_MCI_Recall: 0.9813\n",
            "Epoch 32/50\n",
            "698/698 [==============================] - 102s 146ms/step - loss: 0.0309 - Accuracy: 0.9918 - AD_Precision: 0.9823 - CN_Precision: 0.9977 - MCI_Precision: 0.9916 - AD_Recall: 0.9842 - CN_Recall: 0.9942 - MCI_Recall: 0.9930 - val_loss: 0.7477 - val_Accuracy: 0.8825 - val_AD_Precision: 0.8413 - val_CN_Precision: 0.8737 - val_MCI_Precision: 0.9029 - val_AD_Recall: 0.8346 - val_CN_Recall: 0.8426 - val_MCI_Recall: 0.9198\n",
            "Epoch 33/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.0428 - Accuracy: 0.9860 - AD_Precision: 0.9920 - CN_Precision: 0.9824 - MCI_Precision: 0.9867 - AD_Recall: 0.9842 - CN_Recall: 0.9801 - MCI_Recall: 0.9902 - val_loss: 1.0164 - val_Accuracy: 0.8481 - val_AD_Precision: 0.9425 - val_CN_Precision: 0.9146 - val_MCI_Precision: 0.8072 - val_AD_Recall: 0.6457 - val_CN_Recall: 0.7614 - val_MCI_Recall: 0.9626\n",
            "Epoch 34/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0418 - Accuracy: 0.9878 - AD_Precision: 0.9784 - CN_Precision: 0.9906 - MCI_Precision: 0.9902 - AD_Recall: 0.9842 - CN_Recall: 0.9895 - MCI_Recall: 0.9881 - val_loss: 1.2930 - val_Accuracy: 0.8309 - val_AD_Precision: 0.9839 - val_CN_Precision: 0.8833 - val_MCI_Precision: 0.7912 - val_AD_Recall: 0.4803 - val_CN_Recall: 0.8071 - val_MCI_Recall: 0.9626\n",
            "Epoch 35/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0419 - Accuracy: 0.9889 - AD_Precision: 0.9862 - CN_Precision: 0.9883 - MCI_Precision: 0.9902 - AD_Recall: 0.9842 - CN_Recall: 0.9895 - MCI_Recall: 0.9902 - val_loss: 0.8929 - val_Accuracy: 0.8739 - val_AD_Precision: 0.9223 - val_CN_Precision: 0.8983 - val_MCI_Precision: 0.8513 - val_AD_Recall: 0.7480 - val_CN_Recall: 0.8071 - val_MCI_Recall: 0.9492\n",
            "Epoch 36/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0289 - Accuracy: 0.9900 - AD_Precision: 0.9862 - CN_Precision: 0.9906 - MCI_Precision: 0.9909 - AD_Recall: 0.9882 - CN_Recall: 0.9860 - MCI_Recall: 0.9923 - val_loss: 0.9501 - val_Accuracy: 0.8596 - val_AD_Precision: 0.9263 - val_CN_Precision: 0.7551 - val_MCI_Precision: 0.9157 - val_AD_Recall: 0.6929 - val_CN_Recall: 0.9391 - val_MCI_Recall: 0.8717\n",
            "Epoch 37/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0369 - Accuracy: 0.9903 - AD_Precision: 0.9881 - CN_Precision: 0.9872 - MCI_Precision: 0.9930 - AD_Recall: 0.9862 - CN_Recall: 0.9918 - MCI_Recall: 0.9909 - val_loss: 0.9830 - val_Accuracy: 0.8653 - val_AD_Precision: 0.9767 - val_CN_Precision: 0.8594 - val_MCI_Precision: 0.8452 - val_AD_Recall: 0.6614 - val_CN_Recall: 0.8376 - val_MCI_Recall: 0.9492\n",
            "Epoch 38/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0286 - Accuracy: 0.9903 - AD_Precision: 0.9921 - CN_Precision: 0.9871 - MCI_Precision: 0.9916 - AD_Recall: 0.9862 - CN_Recall: 0.9883 - MCI_Recall: 0.9930 - val_loss: 1.1133 - val_Accuracy: 0.8596 - val_AD_Precision: 0.8783 - val_CN_Precision: 0.9388 - val_MCI_Precision: 0.8276 - val_AD_Recall: 0.7953 - val_CN_Recall: 0.7005 - val_MCI_Recall: 0.9626\n",
            "Epoch 39/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0471 - Accuracy: 0.9875 - AD_Precision: 0.9802 - CN_Precision: 0.9883 - MCI_Precision: 0.9895 - AD_Recall: 0.9763 - CN_Recall: 0.9895 - MCI_Recall: 0.9895 - val_loss: 0.7940 - val_Accuracy: 0.8725 - val_AD_Precision: 0.9091 - val_CN_Precision: 0.8279 - val_MCI_Precision: 0.8901 - val_AD_Recall: 0.7087 - val_CN_Recall: 0.9036 - val_MCI_Recall: 0.9091\n",
            "Epoch 40/50\n",
            "698/698 [==============================] - 97s 140ms/step - loss: 0.0435 - Accuracy: 0.9871 - AD_Precision: 0.9782 - CN_Precision: 0.9872 - MCI_Precision: 0.9909 - AD_Recall: 0.9744 - CN_Recall: 0.9895 - MCI_Recall: 0.9902 - val_loss: 0.8140 - val_Accuracy: 0.8811 - val_AD_Precision: 0.8417 - val_CN_Precision: 0.9130 - val_MCI_Precision: 0.8804 - val_AD_Recall: 0.7953 - val_CN_Recall: 0.8528 - val_MCI_Recall: 0.9251\n",
            "Epoch 41/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.0222 - Accuracy: 0.9932 - AD_Precision: 0.9940 - CN_Precision: 0.9941 - MCI_Precision: 0.9923 - AD_Recall: 0.9882 - CN_Recall: 0.9918 - MCI_Recall: 0.9951 - val_loss: 0.8992 - val_Accuracy: 0.8653 - val_AD_Precision: 0.9200 - val_CN_Precision: 0.7860 - val_MCI_Precision: 0.9022 - val_AD_Recall: 0.7244 - val_CN_Recall: 0.9137 - val_MCI_Recall: 0.8877\n",
            "Epoch 42/50\n",
            "698/698 [==============================] - 97s 140ms/step - loss: 0.0320 - Accuracy: 0.9907 - AD_Precision: 0.9921 - CN_Precision: 0.9906 - MCI_Precision: 0.9902 - AD_Recall: 0.9882 - CN_Recall: 0.9895 - MCI_Recall: 0.9923 - val_loss: 0.9674 - val_Accuracy: 0.8524 - val_AD_Precision: 0.8168 - val_CN_Precision: 0.7531 - val_MCI_Precision: 0.9441 - val_AD_Recall: 0.8425 - val_CN_Recall: 0.9289 - val_MCI_Recall: 0.8128\n",
            "Epoch 43/50\n",
            "698/698 [==============================] - 102s 145ms/step - loss: 0.0442 - Accuracy: 0.9885 - AD_Precision: 0.9862 - CN_Precision: 0.9883 - MCI_Precision: 0.9895 - AD_Recall: 0.9842 - CN_Recall: 0.9883 - MCI_Recall: 0.9902 - val_loss: 0.7805 - val_Accuracy: 0.8840 - val_AD_Precision: 0.8583 - val_CN_Precision: 0.8700 - val_MCI_Precision: 0.9019 - val_AD_Recall: 0.8110 - val_CN_Recall: 0.8832 - val_MCI_Recall: 0.9091\n",
            "Epoch 44/50\n",
            "698/698 [==============================] - 98s 141ms/step - loss: 0.0366 - Accuracy: 0.9896 - AD_Precision: 0.9823 - CN_Precision: 0.9918 - MCI_Precision: 0.9909 - AD_Recall: 0.9842 - CN_Recall: 0.9918 - MCI_Recall: 0.9902 - val_loss: 1.0666 - val_Accuracy: 0.8639 - val_AD_Precision: 0.9625 - val_CN_Precision: 0.8373 - val_MCI_Precision: 0.8624 - val_AD_Recall: 0.6063 - val_CN_Recall: 0.8883 - val_MCI_Recall: 0.9385\n",
            "Epoch 45/50\n",
            "698/698 [==============================] - 102s 146ms/step - loss: 0.0386 - Accuracy: 0.9892 - AD_Precision: 0.9901 - CN_Precision: 0.9860 - MCI_Precision: 0.9909 - AD_Recall: 0.9842 - CN_Recall: 0.9895 - MCI_Recall: 0.9909 - val_loss: 0.7824 - val_Accuracy: 0.8911 - val_AD_Precision: 0.8359 - val_CN_Precision: 0.9171 - val_MCI_Precision: 0.8969 - val_AD_Recall: 0.8425 - val_CN_Recall: 0.8426 - val_MCI_Recall: 0.9305\n",
            "Epoch 46/50\n",
            "698/698 [==============================] - 99s 141ms/step - loss: 0.0277 - Accuracy: 0.9918 - AD_Precision: 0.9862 - CN_Precision: 0.9895 - MCI_Precision: 0.9951 - AD_Recall: 0.9862 - CN_Recall: 0.9906 - MCI_Recall: 0.9944 - val_loss: 0.8693 - val_Accuracy: 0.8854 - val_AD_Precision: 0.8346 - val_CN_Precision: 0.8978 - val_MCI_Precision: 0.8984 - val_AD_Recall: 0.8346 - val_CN_Recall: 0.8477 - val_MCI_Recall: 0.9225\n",
            "Epoch 47/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0220 - Accuracy: 0.9935 - AD_Precision: 0.9882 - CN_Precision: 0.9941 - MCI_Precision: 0.9951 - AD_Recall: 0.9941 - CN_Recall: 0.9918 - MCI_Recall: 0.9944 - val_loss: 0.9061 - val_Accuracy: 0.8897 - val_AD_Precision: 0.8952 - val_CN_Precision: 0.8974 - val_MCI_Precision: 0.8844 - val_AD_Recall: 0.7402 - val_CN_Recall: 0.8883 - val_MCI_Recall: 0.9412\n",
            "Epoch 48/50\n",
            "698/698 [==============================] - 97s 139ms/step - loss: 0.0411 - Accuracy: 0.9878 - AD_Precision: 0.9863 - CN_Precision: 0.9848 - MCI_Precision: 0.9902 - AD_Recall: 0.9921 - CN_Recall: 0.9825 - MCI_Recall: 0.9895 - val_loss: 1.2445 - val_Accuracy: 0.8596 - val_AD_Precision: 0.9231 - val_CN_Precision: 0.9329 - val_MCI_Precision: 0.8194 - val_AD_Recall: 0.6614 - val_CN_Recall: 0.7766 - val_MCI_Recall: 0.9706\n",
            "Epoch 49/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.0141 - Accuracy: 0.9961 - AD_Precision: 0.9941 - CN_Precision: 0.9953 - MCI_Precision: 0.9972 - AD_Recall: 0.9921 - CN_Recall: 0.9977 - MCI_Recall: 0.9958 - val_loss: 0.8776 - val_Accuracy: 0.8782 - val_AD_Precision: 0.8374 - val_CN_Precision: 0.9138 - val_MCI_Precision: 0.8797 - val_AD_Recall: 0.8110 - val_CN_Recall: 0.8071 - val_MCI_Recall: 0.9385\n",
            "Epoch 50/50\n",
            "698/698 [==============================] - 98s 140ms/step - loss: 0.0261 - Accuracy: 0.9907 - AD_Precision: 0.9882 - CN_Precision: 0.9906 - MCI_Precision: 0.9916 - AD_Recall: 0.9921 - CN_Recall: 0.9895 - MCI_Recall: 0.9909 - val_loss: 0.9036 - val_Accuracy: 0.8754 - val_AD_Precision: 0.8500 - val_CN_Precision: 0.8224 - val_MCI_Precision: 0.9148 - val_AD_Recall: 0.8031 - val_CN_Recall: 0.8934 - val_MCI_Recall: 0.8904\n"
          ]
        }
      ],
      "source": [
        "# learning_rate=10E-6\n",
        "savepath = '/content/drive/MyDrive/ADNI/resnet50_row_10E-6/'\n",
        "lr=0.000001\n",
        "if not os.path.isdir(savepath):\n",
        "      os.makedirs(savepath)\n",
        "csvloger_callback = tf.keras.callbacks.CSVLogger(f'{savepath}history.csv', separator=',', append=False)\n",
        "\n",
        "modelcheckpoint_callback = tf.keras.callbacks.ModelCheckpoint(savepath+'resnet50_row_10E-6_weights-epoch:{epoch:03d}-val_Accuracy:{val_Accuracy:.4f}.h5',monitor='val_Accuracy'\n",
        "    ,verbose=0,save_best_only=True,save_weights_only=True,mode='max',save_freq='epoch' , initial_value_threshold=0.85)\n",
        "\n",
        "OPT = tf.keras.optimizers.Adam(learning_rate=lr)\n",
        "model06 = resnet50_row(OPT, LOOS,METRICS,image_Shape)\n",
        "history06 = model06.fit(training_image_datagen, epochs=EPOCH, validation_data=testing_image_datagen,callbacks=[csvloger_callback,modelcheckpoint_callback])\n",
        "model06 = resnet50_row(OPT, LOOS,METRICS,image_Shape)\n",
        "model06.load_weights(glob.glob(f\"{savepath}resnet50_row_10E-6_weights-epoch:*-val_Accuracy:{max(history06.history['val_Accuracy']):.4f}.h5\")[0])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 477
        },
        "id": "4NDdptp1DchX",
        "outputId": "f2c1981b-6631-46cf-8162-59ed609f6846"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "          AD       0.84      0.84      0.84       127\n",
            "          CN       0.92      0.85      0.88       197\n",
            "         MCI       0.90      0.93      0.91       374\n",
            "\n",
            "    accuracy                           0.89       698\n",
            "   macro avg       0.88      0.87      0.88       698\n",
            "weighted avg       0.89      0.89      0.89       698\n",
            "\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "p = model06.predict(testing_image_datagen)\n",
        "predicted_lab = p.argmax(axis=-1)\n",
        "true_lab = testing_image_datagen.classes\n",
        "plot_confusion_matrix(true_lab, predicted_lab, testing_image_datagen.class_indices.keys(),savepath)\n",
        "report = classification_report(true_lab, predicted_lab, target_names=testing_image_datagen.class_indices.keys())\n",
        "print(report)\n",
        "dump_text(report,f'{savepath}classification_report.txt')"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "collapsed_sections": [],
      "name": "AD classification.ipynb",
      "provenance": []
    },
    "gpuClass": "standard",
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}