model/ImageSimlarityFinally (3).ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "T4"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "OAWN28MEA1Zo"
      },
      "outputs": [],
      "source": [
        "import pickle\n",
        "import random\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "import os\n",
        "import matplotlib.pyplot as plt\n",
        "import random\n",
        "from PIL import Image\n",
        "import  cv2\n",
        "from imblearn.over_sampling import SMOTE\n",
        "from sklearn.metrics import confusion_matrix\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras.preprocessing import image\n",
        "from tensorflow.keras.applications.resnet50 import preprocess_input\n",
        "import tensorflow as tf\n",
        "from sklearn.metrics.pairwise import cosine_similarity\n",
        "import pandas as pd\n",
        "from numpy import trapz\n",
        "from tensorflow.keras.applications.resnet50 import ResNet50\n",
        "\n",
        "from tensorflow.keras.applications.resnet import ResNet101\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.layers import Dense, GlobalAveragePooling2D\n",
        "from keras.optimizers import Adam\n",
        "import keras\n",
        "from keras.layers import Input,Conv2D,Activation,Dense,Lambda,Flatten,Embedding,PReLU,BatchNormalization,Dropout\n",
        "from keras.models import Model\n",
        "import keras.backend as K\n",
        "from keras.utils import to_categorical\n",
        "import numpy as np\n",
        "from keras.callbacks import EarlyStopping\n",
        "\n",
        "from keras.applications.densenet import DenseNet201\n",
        "\n",
        "from tensorflow.keras import layers, losses, optimizers, metrics"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "<h2>Model</h2>"
      ],
      "metadata": {
        "id": "mdO3eYE6BAVe"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class Models():\n",
        "  def __init__(self,shape,features_dims,num_classes):\n",
        "      self.shape = shape\n",
        "      self.features_dims = features_dims\n",
        "      self.num_classes = num_classes\n",
        "  def Resnet50(self):\n",
        "    resnet_50 = ResNet50(weights='imagenet', include_top=False)\n",
        "    resnet_50.trainable=False\n",
        "    inputs = Input(self.shape)\n",
        "    x = resnet_50(inputs)\n",
        "    x = GlobalAveragePooling2D()(x)\n",
        "    # x = Dropout(0.3)(x)\n",
        "    features = Dense(self.features_dims, activation=\"relu\", name=\"features\")(x)\n",
        "    outputs =Dense(self.num_classes, activation=\"softmax\")(features)\n",
        "    return Model(inputs, [features, outputs])\n",
        "\n",
        "  def Resnet101(self):\n",
        "      resnet_101 = ResNet101(weights='imagenet', include_top=False)\n",
        "      resnet_101.trainable=False\n",
        "      inputs = Input(self.shape)\n",
        "      x = resnet_101(inputs)\n",
        "      x = GlobalAveragePooling2D()(x)\n",
        "      features = Dense(self.features_dims, activation=\"relu\", name=\"features\")(x)\n",
        "      outputs =Dense(self.num_classes, activation=\"softmax\")(features)\n",
        "      return Model(inputs, [features, outputs])\n",
        "\n",
        "  def DenseNet201(self):\n",
        "      densenet_201 =  DenseNet201(weights='imagenet', include_top=False)\n",
        "      densenet_201.trainable=False\n",
        "      inputs = Input(self.shape)\n",
        "      x = densenet_201(inputs)\n",
        "      x = GlobalAveragePooling2D()(x)\n",
        "      # x = Dropout(0.3)(x)\n",
        "      features = Dense(self.features_dims, activation=\"relu\", name=\"features\")(x)\n",
        "      outputs =Dense(num_classes, activation=\"softmax\")(features)\n",
        "      return Model(inputs, [features, outputs])\n"
      ],
      "metadata": {
        "id": "P6f8P3E-BD4R"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "<h2>Prepare data</h2>"
      ],
      "metadata": {
        "id": "wVCbFR4oBZQr"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pqCsrvhqN7Pd",
        "outputId": "f720b4fe-b790-4d51-e77d-4e2a889b0c27"
      },
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "with open('/content/drive/MyDrive/resized_train_data.pickle', 'rb') as f:\n",
        "    x_train, y_train = pickle.load(f)"
      ],
      "metadata": {
        "id": "CDztdWtdLqLF"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "with open('/content/drive/MyDrive/resized_valid_data.pickle', 'rb') as f:\n",
        "    x_valid, y_valid = pickle.load(f)"
      ],
      "metadata": {
        "id": "neEQvnE9Lyhj"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "y_train_one_hot=to_categorical(y_train)\n",
        "y_valid_one_hot=to_categorical(y_valid)\n",
        "\n",
        "print(x_train.shape)\n",
        "print(y_train.shape)\n",
        "print(y_valid_one_hot.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cFnsbAmeL4rI",
        "outputId": "201962f9-5f25-4dd9-cd4c-9cc3daf43c65"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(14569, 72, 72, 3)\n",
            "(14569,)\n",
            "(5440, 80)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def shufflerandom(x_,y_,y_one_hot):\n",
        "  indices = np.arange(len(y_))\n",
        "  np.random.shuffle(indices)\n",
        "  x_ = x_[indices]\n",
        "  y_one_hot = y_one_hot[indices]\n",
        "  y_ = y_[indices]"
      ],
      "metadata": {
        "id": "BEGH63rzL69j"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "shufflerandom(x_train,y_train,y_train_one_hot)\n",
        "shufflerandom(x_valid,y_valid,y_valid_one_hot)"
      ],
      "metadata": {
        "id": "nrChrM2hMTkR"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "x_train = x_train.astype('float32')\n",
        "x_valid = x_valid.astype('float32')\n",
        "x_train /= 255\n",
        "x_valid /= 255"
      ],
      "metadata": {
        "id": "RbJkehUEM6YO"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "<h2>Center Loss</h2>"
      ],
      "metadata": {
        "id": "Ph3iN0-7Lkp0"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class CenterLoss(losses.Loss):\n",
        "    def __init__(self, num_classes, feature_dims, alpha=0.5, reduction=losses.Reduction.AUTO, name='center_loss'):\n",
        "        super(CenterLoss, self).__init__(reduction=reduction, name=name)\n",
        "\n",
        "        self.num_classes = num_classes\n",
        "        self.feature_dims = feature_dims\n",
        "\n",
        "        # hyper parameter for updating the center point\n",
        "        self.alpha = alpha\n",
        "\n",
        "        self.centers = tf.Variable(tf.zeros(shape = (self.num_classes, self.feature_dims)))\n",
        "\n",
        "\n",
        "    def call(self, y_true, y_pred):\n",
        "        \"\"\"\n",
        "            y_ture : same teacher signal as for classification (1-hot vector)\n",
        "                     shape = (batch_size, num_classes)\n",
        "            y_pred : output of features in the middle layer of the model\n",
        "                     shape = (batch_size, feature_dims)\n",
        "        \"\"\"\n",
        "\n",
        "        labels = tf.argmax(y_true, axis=-1)\n",
        "\n",
        "        centers_batch = tf.gather(self.centers, labels)\n",
        "\n",
        "        diff = centers_batch - y_pred\n",
        "\n",
        "        loss = tf.reduce_mean(tf.square(diff))\n",
        "\n",
        "\n",
        "        unique_label, unique_idx, unique_count = tf.unique_with_counts(labels)\n",
        "        appear_times = tf.gather(unique_count, unique_idx)\n",
        "        appear_times = tf.reshape(appear_times, [-1, 1])\n",
        "\n",
        "        diff = diff / tf.cast((1 + appear_times), tf.float32)\n",
        "\n",
        "        diff = self.alpha * diff\n",
        "\n",
        "        self.centers = tf.compat.v1.scatter_sub(self.centers, labels, diff)\n",
        "\n",
        "        return loss"
      ],
      "metadata": {
        "id": "-uuD2grpLpTW"
      },
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "num_classes = 80\n",
        "feature_dims = 2048\n",
        "\n",
        "center_loss_obj = CenterLoss(num_classes, feature_dims)"
      ],
      "metadata": {
        "id": "DU5IgbWoNRqt"
      },
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def each_img_distance(y_true, y_pred):\n",
        "    \"\"\"\n",
        "        y_true: true labels (1-hot vector)\n",
        "        y_pred: predicted labels (output of features in the middle layer)\n",
        "\n",
        "    \"\"\"\n",
        "\n",
        "    centers = center_loss_obj.centers  # Get the centers from center_loss_obj\n",
        "    labels = tf.argmax(y_true, axis=-1)  # Retrieve the predicted labels by taking the index of the maximum value in y_true\n",
        "    centers_batch = tf.gather(centers, labels)  # Gather the corresponding centers based on the predicted labels\n",
        "    distances = tf.norm(y_pred - centers_batch, axis=1)  # Calculate the L2 (Euclidean) distance between y_pred and centers_batch\n",
        "\n",
        "    return distances"
      ],
      "metadata": {
        "id": "PGN2GRw1NExV"
      },
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def average_center_distance(y_true, y_pred):\n",
        "    \"\"\"\n",
        "        y_true: true labels (1-hot vector)\n",
        "        y_pred: predicted labels (output of features in the middle layer)\n",
        "\n",
        "    \"\"\"\n",
        "\n",
        "    centers = center_loss_obj.centers  # Get the centers from the center_loss_obj\n",
        "    num_centers = tf.shape(centers)[0]  # Get the number of centers\n",
        "\n",
        "    # Calculate pairwise Euclidean distances between centers\n",
        "    tiled_centers = tf.tile(tf.expand_dims(centers, axis=1), [1, num_centers, 1])  # Tile the centers to create pairs\n",
        "    pairwise_distances = tf.norm(tiled_centers - tf.expand_dims(centers, axis=0), axis=-1)  # Calculate Euclidean distances\n",
        "\n",
        "    # Exclude self-distances and calculate average distance\n",
        "    mask = tf.ones([num_centers, num_centers]) - tf.eye(num_centers)  # Create a mask to exclude self-distances\n",
        "    valid_distances = tf.boolean_mask(pairwise_distances, tf.cast(mask, tf.bool))  # Apply mask to exclude self-distances\n",
        "    avg_distance = tf.reduce_mean(valid_distances)  # Calculate the average distance\n",
        "\n",
        "    return avg_distance"
      ],
      "metadata": {
        "id": "wAu0rVLnNLxL"
      },
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "shape = (72, 72,3)\n",
        "features_dims = 2048\n",
        "num_classes = 80\n",
        "\n",
        "model_builder = Models(shape, features_dims, num_classes)\n",
        "\n",
        "model = model_builder.Resnet50()\n",
        "\n",
        "\n",
        "\n",
        "model.compile(optimizer=\"adam\",\n",
        "              loss=[center_loss_obj, losses.CategoricalCrossentropy()],\n",
        "              metrics=[[each_img_distance, average_center_distance, metrics.CategoricalAccuracy()], [metrics.CategoricalAccuracy()]])\n",
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vXYsYMQtNdtV",
        "outputId": "87cfd515-7ca4-43c2-b1be-4917fe7e693d"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_2\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_6 (InputLayer)        [(None, 72, 72, 3)]       0         \n",
            "                                                                 \n",
            " resnet50 (Functional)       (None, None, None, 2048)  23587712  \n",
            "                                                                 \n",
            " global_average_pooling2d_2   (None, 2048)             0         \n",
            " (GlobalAveragePooling2D)                                        \n",
            "                                                                 \n",
            " features (Dense)            (None, 2048)              4196352   \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 80)                163920    \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 27,947,984\n",
            "Trainable params: 4,360,272\n",
            "Non-trainable params: 23,587,712\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(x_train.shape,y_train_one_hot.shape,)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xKMi79wuP9pT",
        "outputId": "c11bf818-ec55-401a-8698-fefe232f78ec"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(14569, 72, 72, 3) (14569, 80)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "history_with_centerloss1 = model.fit(\n",
        "    x_train,\n",
        "    y_train_one_hot,\n",
        "    validation_data=(x_valid,y_valid_one_hot),\n",
        "    epochs=5,\n",
        "    batch_size=32,\n",
        ")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-PPaOWspOUtq",
        "outputId": "df180be0-4b45-4ad1-fee0-3c6508783169"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/5\n",
            "456/456 [==============================] - 29s 54ms/step - loss: 4.2313 - features_loss: 0.0070 - dense_2_loss: 4.2242 - features_each_img_distance: 2.3080 - features_average_center_distance: 2.4707 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.0355 - val_loss: 4.0678 - val_features_loss: 0.0047 - val_dense_2_loss: 4.0632 - val_features_each_img_distance: 2.5138 - val_features_average_center_distance: 2.6991 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.0664\n",
            "Epoch 2/5\n",
            "456/456 [==============================] - 15s 33ms/step - loss: 3.9552 - features_loss: 0.0072 - dense_2_loss: 3.9481 - features_each_img_distance: 2.7071 - features_average_center_distance: 3.1467 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.0793 - val_loss: 3.8607 - val_features_loss: 0.0073 - val_dense_2_loss: 3.8535 - val_features_each_img_distance: 3.2928 - val_features_average_center_distance: 3.3247 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.0921\n",
            "Epoch 3/5\n",
            "456/456 [==============================] - 15s 34ms/step - loss: 3.7438 - features_loss: 0.0087 - dense_2_loss: 3.7350 - features_each_img_distance: 3.0386 - features_average_center_distance: 3.6050 - features_categorical_accuracy: 6.8639e-05 - dense_2_categorical_accuracy: 0.1138 - val_loss: 3.6915 - val_features_loss: 0.0081 - val_dense_2_loss: 3.6834 - val_features_each_img_distance: 3.5416 - val_features_average_center_distance: 3.3898 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.1274\n",
            "Epoch 4/5\n",
            "456/456 [==============================] - 14s 31ms/step - loss: 3.5733 - features_loss: 0.0096 - dense_2_loss: 3.5637 - features_each_img_distance: 3.2171 - features_average_center_distance: 3.8224 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.1457 - val_loss: 3.5737 - val_features_loss: 0.0087 - val_dense_2_loss: 3.5650 - val_features_each_img_distance: 3.7290 - val_features_average_center_distance: 3.6181 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.1450\n",
            "Epoch 5/5\n",
            "456/456 [==============================] - 14s 31ms/step - loss: 3.4443 - features_loss: 0.0102 - dense_2_loss: 3.4341 - features_each_img_distance: 3.3278 - features_average_center_distance: 3.9397 - features_categorical_accuracy: 6.8639e-05 - dense_2_categorical_accuracy: 0.1725 - val_loss: 3.4437 - val_features_loss: 0.0090 - val_dense_2_loss: 3.4347 - val_features_each_img_distance: 3.7988 - val_features_average_center_distance: 3.6858 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.1767\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.trainable = True\n",
        "\n",
        "model.compile(keras.optimizers.Adam(1e-5),\n",
        "              loss=[center_loss_obj, losses.CategoricalCrossentropy()],\n",
        "              metrics=[[each_img_distance, average_center_distance, metrics.CategoricalAccuracy()], [metrics.CategoricalAccuracy()]])\n",
        "\n",
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "x8P5ir-Fby0T",
        "outputId": "99edb735-ec16-49bc-8373-03de230baa10"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_2\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_6 (InputLayer)        [(None, 72, 72, 3)]       0         \n",
            "                                                                 \n",
            " resnet50 (Functional)       (None, None, None, 2048)  23587712  \n",
            "                                                                 \n",
            " global_average_pooling2d_2   (None, 2048)             0         \n",
            " (GlobalAveragePooling2D)                                        \n",
            "                                                                 \n",
            " features (Dense)            (None, 2048)              4196352   \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 80)                163920    \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 27,947,984\n",
            "Trainable params: 27,894,864\n",
            "Non-trainable params: 53,120\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "history_with_centerloss1 = model.fit(\n",
        "    x_train,\n",
        "    y_train_one_hot,\n",
        "    validation_data=(x_valid,y_valid_one_hot),\n",
        "    epochs=50,\n",
        "    batch_size=32,\n",
        ")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Z3I1DVW1eT4d",
        "outputId": "95b25483-2890-458a-c04c-06b46584a2df"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/50\n",
            "456/456 [==============================] - 44s 84ms/step - loss: 8.5462 - features_loss: 0.3146 - dense_2_loss: 8.2315 - features_each_img_distance: 16.0188 - features_average_center_distance: 17.3962 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.0771 - val_loss: 136.1408 - val_features_loss: 51.5723 - val_dense_2_loss: 84.5686 - val_features_each_img_distance: 190.7948 - val_features_average_center_distance: 149.3380 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.0136\n",
            "Epoch 2/50\n",
            "456/456 [==============================] - 36s 80ms/step - loss: 4.1492 - features_loss: 1.0250 - dense_2_loss: 3.1242 - features_each_img_distance: 14.6134 - features_average_center_distance: 15.1431 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.2363 - val_loss: 4.9074 - val_features_loss: 0.5849 - val_dense_2_loss: 4.3226 - val_features_each_img_distance: 19.3844 - val_features_average_center_distance: 15.6775 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.1590\n",
            "Epoch 3/50\n",
            "456/456 [==============================] - 35s 76ms/step - loss: 2.4229 - features_loss: 0.0934 - dense_2_loss: 2.3294 - features_each_img_distance: 9.6011 - features_average_center_distance: 14.3568 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.4173 - val_loss: 2.3235 - val_features_loss: 0.1667 - val_dense_2_loss: 2.1568 - val_features_each_img_distance: 13.8715 - val_features_average_center_distance: 16.1873 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.4921\n",
            "Epoch 4/50\n",
            "456/456 [==============================] - 36s 80ms/step - loss: 1.7296 - features_loss: 0.1094 - dense_2_loss: 1.6202 - features_each_img_distance: 10.9336 - features_average_center_distance: 17.5111 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.5889 - val_loss: 1.6011 - val_features_loss: 0.1541 - val_dense_2_loss: 1.4470 - val_features_each_img_distance: 15.2360 - val_features_average_center_distance: 19.3099 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.6511\n",
            "Epoch 5/50\n",
            "456/456 [==============================] - 35s 76ms/step - loss: 1.2004 - features_loss: 0.1301 - dense_2_loss: 1.0703 - features_each_img_distance: 12.2152 - features_average_center_distance: 20.5929 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.7346 - val_loss: 1.2384 - val_features_loss: 0.1642 - val_dense_2_loss: 1.0742 - val_features_each_img_distance: 16.4876 - val_features_average_center_distance: 21.4352 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.7531\n",
            "Epoch 6/50\n",
            "456/456 [==============================] - 35s 77ms/step - loss: 0.8443 - features_loss: 0.1427 - dense_2_loss: 0.7016 - features_each_img_distance: 12.9092 - features_average_center_distance: 22.5753 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.8230 - val_loss: 0.9481 - val_features_loss: 0.1733 - val_dense_2_loss: 0.7748 - val_features_each_img_distance: 17.1142 - val_features_average_center_distance: 22.5880 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.8074\n",
            "Epoch 7/50\n",
            "456/456 [==============================] - 36s 78ms/step - loss: 0.6202 - features_loss: 0.1468 - dense_2_loss: 0.4733 - features_each_img_distance: 13.1883 - features_average_center_distance: 23.8435 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.8823 - val_loss: 0.7850 - val_features_loss: 0.1719 - val_dense_2_loss: 0.6131 - val_features_each_img_distance: 17.2538 - val_features_average_center_distance: 23.9338 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.8428\n",
            "Epoch 8/50\n",
            "456/456 [==============================] - 37s 81ms/step - loss: 0.4686 - features_loss: 0.1495 - dense_2_loss: 0.3191 - features_each_img_distance: 13.3060 - features_average_center_distance: 24.8020 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9251 - val_loss: 0.6614 - val_features_loss: 0.1701 - val_dense_2_loss: 0.4913 - val_features_each_img_distance: 17.3396 - val_features_average_center_distance: 25.0731 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.8673\n",
            "Epoch 9/50\n",
            "456/456 [==============================] - 37s 82ms/step - loss: 0.3611 - features_loss: 0.1455 - dense_2_loss: 0.2156 - features_each_img_distance: 13.1465 - features_average_center_distance: 25.2807 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9562 - val_loss: 0.6139 - val_features_loss: 0.1634 - val_dense_2_loss: 0.4505 - val_features_each_img_distance: 17.0015 - val_features_average_center_distance: 25.0601 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.8814\n",
            "Epoch 10/50\n",
            "456/456 [==============================] - 35s 77ms/step - loss: 0.2981 - features_loss: 0.1416 - dense_2_loss: 0.1565 - features_each_img_distance: 12.9801 - features_average_center_distance: 25.4478 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9734 - val_loss: 0.5365 - val_features_loss: 0.1574 - val_dense_2_loss: 0.3791 - val_features_each_img_distance: 16.6480 - val_features_average_center_distance: 24.8164 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9011\n",
            "Epoch 11/50\n",
            "456/456 [==============================] - 35s 77ms/step - loss: 0.2458 - features_loss: 0.1336 - dense_2_loss: 0.1121 - features_each_img_distance: 12.6010 - features_average_center_distance: 25.4928 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9842 - val_loss: 0.4780 - val_features_loss: 0.1473 - val_dense_2_loss: 0.3307 - val_features_each_img_distance: 16.2270 - val_features_average_center_distance: 24.9631 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9134\n",
            "Epoch 12/50\n",
            "456/456 [==============================] - 39s 85ms/step - loss: 0.2110 - features_loss: 0.1249 - dense_2_loss: 0.0861 - features_each_img_distance: 12.1931 - features_average_center_distance: 25.2201 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9890 - val_loss: 0.5007 - val_features_loss: 0.1420 - val_dense_2_loss: 0.3588 - val_features_each_img_distance: 15.8034 - val_features_average_center_distance: 24.9442 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9068\n",
            "Epoch 13/50\n",
            "456/456 [==============================] - 36s 79ms/step - loss: 0.1848 - features_loss: 0.1173 - dense_2_loss: 0.0675 - features_each_img_distance: 11.8175 - features_average_center_distance: 25.0693 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9932 - val_loss: 0.4027 - val_features_loss: 0.1273 - val_dense_2_loss: 0.2754 - val_features_each_img_distance: 15.1166 - val_features_average_center_distance: 24.3345 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9239\n",
            "Epoch 14/50\n",
            "456/456 [==============================] - 35s 77ms/step - loss: 0.1600 - features_loss: 0.1080 - dense_2_loss: 0.0520 - features_each_img_distance: 11.3387 - features_average_center_distance: 24.7204 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9967 - val_loss: 0.3773 - val_features_loss: 0.1186 - val_dense_2_loss: 0.2588 - val_features_each_img_distance: 14.5941 - val_features_average_center_distance: 24.1645 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9281\n",
            "Epoch 15/50\n",
            "456/456 [==============================] - 35s 77ms/step - loss: 0.1480 - features_loss: 0.1017 - dense_2_loss: 0.0462 - features_each_img_distance: 11.0070 - features_average_center_distance: 24.3812 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9965 - val_loss: 0.3542 - val_features_loss: 0.1107 - val_dense_2_loss: 0.2435 - val_features_each_img_distance: 14.0989 - val_features_average_center_distance: 23.6244 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9325\n",
            "Epoch 16/50\n",
            "456/456 [==============================] - 37s 82ms/step - loss: 0.1330 - features_loss: 0.0936 - dense_2_loss: 0.0394 - features_each_img_distance: 10.5428 - features_average_center_distance: 24.0265 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9971 - val_loss: 0.3354 - val_features_loss: 0.1034 - val_dense_2_loss: 0.2320 - val_features_each_img_distance: 13.6291 - val_features_average_center_distance: 23.5259 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9360\n",
            "Epoch 17/50\n",
            "456/456 [==============================] - 37s 82ms/step - loss: 0.1217 - features_loss: 0.0867 - dense_2_loss: 0.0350 - features_each_img_distance: 10.1515 - features_average_center_distance: 23.6949 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9974 - val_loss: 0.3143 - val_features_loss: 0.0948 - val_dense_2_loss: 0.2195 - val_features_each_img_distance: 13.0218 - val_features_average_center_distance: 23.1346 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9430\n",
            "Epoch 18/50\n",
            "456/456 [==============================] - 37s 81ms/step - loss: 0.1062 - features_loss: 0.0780 - dense_2_loss: 0.0282 - features_each_img_distance: 9.6418 - features_average_center_distance: 23.2116 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9984 - val_loss: 0.2990 - val_features_loss: 0.0868 - val_dense_2_loss: 0.2122 - val_features_each_img_distance: 12.4875 - val_features_average_center_distance: 22.3791 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9410\n",
            "Epoch 19/50\n",
            "456/456 [==============================] - 35s 78ms/step - loss: 0.0936 - features_loss: 0.0714 - dense_2_loss: 0.0221 - features_each_img_distance: 9.2044 - features_average_center_distance: 22.8939 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9993 - val_loss: 0.2764 - val_features_loss: 0.0785 - val_dense_2_loss: 0.1979 - val_features_each_img_distance: 11.8698 - val_features_average_center_distance: 22.0473 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9474\n",
            "Epoch 20/50\n",
            "456/456 [==============================] - 35s 78ms/step - loss: 0.0874 - features_loss: 0.0659 - dense_2_loss: 0.0215 - features_each_img_distance: 8.8625 - features_average_center_distance: 22.5542 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9992 - val_loss: 0.2625 - val_features_loss: 0.0721 - val_dense_2_loss: 0.1904 - val_features_each_img_distance: 11.3894 - val_features_average_center_distance: 21.9602 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9480\n",
            "Epoch 21/50\n",
            "456/456 [==============================] - 37s 82ms/step - loss: 0.0793 - features_loss: 0.0600 - dense_2_loss: 0.0193 - features_each_img_distance: 8.4368 - features_average_center_distance: 22.2137 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9990 - val_loss: 0.2538 - val_features_loss: 0.0671 - val_dense_2_loss: 0.1867 - val_features_each_img_distance: 10.9728 - val_features_average_center_distance: 21.5352 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9496\n",
            "Epoch 22/50\n",
            "456/456 [==============================] - 36s 78ms/step - loss: 0.0731 - features_loss: 0.0564 - dense_2_loss: 0.0167 - features_each_img_distance: 8.1786 - features_average_center_distance: 22.2213 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9992 - val_loss: 0.2386 - val_features_loss: 0.0635 - val_dense_2_loss: 0.1751 - val_features_each_img_distance: 10.6610 - val_features_average_center_distance: 21.3811 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9542\n",
            "Epoch 23/50\n",
            "456/456 [==============================] - 35s 78ms/step - loss: 0.0656 - features_loss: 0.0515 - dense_2_loss: 0.0141 - features_each_img_distance: 7.8014 - features_average_center_distance: 22.0082 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9998 - val_loss: 0.2345 - val_features_loss: 0.0579 - val_dense_2_loss: 0.1767 - val_features_each_img_distance: 10.1458 - val_features_average_center_distance: 21.2593 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9533\n",
            "Epoch 24/50\n",
            "456/456 [==============================] - 36s 78ms/step - loss: 0.0607 - features_loss: 0.0476 - dense_2_loss: 0.0131 - features_each_img_distance: 7.5057 - features_average_center_distance: 21.6062 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9995 - val_loss: 0.2285 - val_features_loss: 0.0548 - val_dense_2_loss: 0.1737 - val_features_each_img_distance: 9.8786 - val_features_average_center_distance: 20.9448 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9518\n",
            "Epoch 25/50\n",
            "456/456 [==============================] - 37s 82ms/step - loss: 0.0530 - features_loss: 0.0424 - dense_2_loss: 0.0106 - features_each_img_distance: 7.0865 - features_average_center_distance: 21.2946 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9997 - val_loss: 0.2122 - val_features_loss: 0.0488 - val_dense_2_loss: 0.1634 - val_features_each_img_distance: 9.3115 - val_features_average_center_distance: 20.4742 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9517\n",
            "Epoch 26/50\n",
            "456/456 [==============================] - 36s 78ms/step - loss: 0.0490 - features_loss: 0.0393 - dense_2_loss: 0.0098 - features_each_img_distance: 6.8212 - features_average_center_distance: 21.1552 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9999 - val_loss: 0.1991 - val_features_loss: 0.0454 - val_dense_2_loss: 0.1537 - val_features_each_img_distance: 8.9801 - val_features_average_center_distance: 20.4531 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9561\n",
            "Epoch 27/50\n",
            "456/456 [==============================] - 35s 78ms/step - loss: 0.0460 - features_loss: 0.0363 - dense_2_loss: 0.0097 - features_each_img_distance: 6.5414 - features_average_center_distance: 20.8829 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9995 - val_loss: 0.1900 - val_features_loss: 0.0418 - val_dense_2_loss: 0.1482 - val_features_each_img_distance: 8.5970 - val_features_average_center_distance: 20.3190 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9579\n",
            "Epoch 28/50\n",
            "456/456 [==============================] - 35s 78ms/step - loss: 0.0395 - features_loss: 0.0322 - dense_2_loss: 0.0073 - features_each_img_distance: 6.1576 - features_average_center_distance: 20.6334 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9999 - val_loss: 0.1802 - val_features_loss: 0.0383 - val_dense_2_loss: 0.1418 - val_features_each_img_distance: 8.2501 - val_features_average_center_distance: 20.1066 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9586\n",
            "Epoch 29/50\n",
            "456/456 [==============================] - 37s 82ms/step - loss: 0.0422 - features_loss: 0.0328 - dense_2_loss: 0.0094 - features_each_img_distance: 6.1892 - features_average_center_distance: 20.9481 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9992 - val_loss: 0.1886 - val_features_loss: 0.0410 - val_dense_2_loss: 0.1476 - val_features_each_img_distance: 8.4858 - val_features_average_center_distance: 20.5057 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9581\n",
            "Epoch 30/50\n",
            "456/456 [==============================] - 37s 82ms/step - loss: 0.0377 - features_loss: 0.0305 - dense_2_loss: 0.0071 - features_each_img_distance: 5.9893 - features_average_center_distance: 20.6578 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9996 - val_loss: 0.1769 - val_features_loss: 0.0351 - val_dense_2_loss: 0.1419 - val_features_each_img_distance: 7.8649 - val_features_average_center_distance: 19.9354 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9603\n",
            "Epoch 31/50\n",
            "456/456 [==============================] - 37s 82ms/step - loss: 0.0327 - features_loss: 0.0268 - dense_2_loss: 0.0059 - features_each_img_distance: 5.6132 - features_average_center_distance: 20.2981 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9999 - val_loss: 0.1715 - val_features_loss: 0.0324 - val_dense_2_loss: 0.1391 - val_features_each_img_distance: 7.5415 - val_features_average_center_distance: 19.6902 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9599\n",
            "Epoch 32/50\n",
            "456/456 [==============================] - 38s 83ms/step - loss: 0.0295 - features_loss: 0.0240 - dense_2_loss: 0.0055 - features_each_img_distance: 5.3158 - features_average_center_distance: 20.0389 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9998 - val_loss: 0.1655 - val_features_loss: 0.0294 - val_dense_2_loss: 0.1361 - val_features_each_img_distance: 7.1926 - val_features_average_center_distance: 19.2478 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9601\n",
            "Epoch 33/50\n",
            "456/456 [==============================] - 36s 79ms/step - loss: 0.0282 - features_loss: 0.0233 - dense_2_loss: 0.0049 - features_each_img_distance: 5.2254 - features_average_center_distance: 20.0587 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9999 - val_loss: 0.1632 - val_features_loss: 0.0286 - val_dense_2_loss: 0.1346 - val_features_each_img_distance: 7.0754 - val_features_average_center_distance: 19.2111 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9640\n",
            "Epoch 34/50\n",
            "456/456 [==============================] - 36s 78ms/step - loss: 0.0259 - features_loss: 0.0215 - dense_2_loss: 0.0043 - features_each_img_distance: 5.0186 - features_average_center_distance: 19.8696 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 1.0000 - val_loss: 0.1576 - val_features_loss: 0.0265 - val_dense_2_loss: 0.1311 - val_features_each_img_distance: 6.7813 - val_features_average_center_distance: 19.3204 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9623\n",
            "Epoch 35/50\n",
            "456/456 [==============================] - 36s 78ms/step - loss: 0.0246 - features_loss: 0.0202 - dense_2_loss: 0.0044 - features_each_img_distance: 4.8643 - features_average_center_distance: 19.8413 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9999 - val_loss: 0.1516 - val_features_loss: 0.0254 - val_dense_2_loss: 0.1262 - val_features_each_img_distance: 6.6247 - val_features_average_center_distance: 19.2684 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9642\n",
            "Epoch 36/50\n",
            "456/456 [==============================] - 36s 78ms/step - loss: 0.0247 - features_loss: 0.0203 - dense_2_loss: 0.0044 - features_each_img_distance: 4.8578 - features_average_center_distance: 19.9517 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9996 - val_loss: 0.1582 - val_features_loss: 0.0247 - val_dense_2_loss: 0.1334 - val_features_each_img_distance: 6.5342 - val_features_average_center_distance: 19.0810 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9627\n",
            "Epoch 37/50\n",
            "456/456 [==============================] - 36s 78ms/step - loss: 0.0200 - features_loss: 0.0170 - dense_2_loss: 0.0030 - features_each_img_distance: 4.4595 - features_average_center_distance: 19.4316 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 1.0000 - val_loss: 0.1498 - val_features_loss: 0.0220 - val_dense_2_loss: 0.1278 - val_features_each_img_distance: 6.1477 - val_features_average_center_distance: 18.7485 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9653\n",
            "Epoch 38/50\n",
            "456/456 [==============================] - 36s 78ms/step - loss: 0.0190 - features_loss: 0.0158 - dense_2_loss: 0.0031 - features_each_img_distance: 4.2989 - features_average_center_distance: 19.3342 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9999 - val_loss: 0.1524 - val_features_loss: 0.0210 - val_dense_2_loss: 0.1314 - val_features_each_img_distance: 6.0005 - val_features_average_center_distance: 18.7724 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9638\n",
            "Epoch 39/50\n",
            "456/456 [==============================] - 37s 82ms/step - loss: 0.0172 - features_loss: 0.0144 - dense_2_loss: 0.0029 - features_each_img_distance: 4.0902 - features_average_center_distance: 19.1316 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9999 - val_loss: 0.1483 - val_features_loss: 0.0198 - val_dense_2_loss: 0.1286 - val_features_each_img_distance: 5.8038 - val_features_average_center_distance: 18.6411 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9642\n",
            "Epoch 40/50\n",
            "456/456 [==============================] - 37s 82ms/step - loss: 0.0160 - features_loss: 0.0135 - dense_2_loss: 0.0026 - features_each_img_distance: 3.9639 - features_average_center_distance: 19.0508 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9999 - val_loss: 0.1427 - val_features_loss: 0.0180 - val_dense_2_loss: 0.1247 - val_features_each_img_distance: 5.5192 - val_features_average_center_distance: 18.5838 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9651\n",
            "Epoch 41/50\n",
            "456/456 [==============================] - 36s 78ms/step - loss: 0.0145 - features_loss: 0.0124 - dense_2_loss: 0.0021 - features_each_img_distance: 3.7904 - features_average_center_distance: 18.8952 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 1.0000 - val_loss: 0.1520 - val_features_loss: 0.0170 - val_dense_2_loss: 0.1350 - val_features_each_img_distance: 5.3374 - val_features_average_center_distance: 18.4019 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9634\n",
            "Epoch 42/50\n",
            "456/456 [==============================] - 37s 82ms/step - loss: 0.0135 - features_loss: 0.0115 - dense_2_loss: 0.0020 - features_each_img_distance: 3.6455 - features_average_center_distance: 18.8872 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9999 - val_loss: 0.1434 - val_features_loss: 0.0163 - val_dense_2_loss: 0.1271 - val_features_each_img_distance: 5.2226 - val_features_average_center_distance: 18.4750 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9658\n",
            "Epoch 43/50\n",
            "456/456 [==============================] - 36s 78ms/step - loss: 0.0135 - features_loss: 0.0108 - dense_2_loss: 0.0026 - features_each_img_distance: 3.5389 - features_average_center_distance: 18.7032 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9998 - val_loss: 0.1452 - val_features_loss: 0.0155 - val_dense_2_loss: 0.1296 - val_features_each_img_distance: 5.0865 - val_features_average_center_distance: 18.1402 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9642\n",
            "Epoch 44/50\n",
            "456/456 [==============================] - 37s 82ms/step - loss: 0.0117 - features_loss: 0.0099 - dense_2_loss: 0.0018 - features_each_img_distance: 3.3823 - features_average_center_distance: 18.6453 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9999 - val_loss: 0.1446 - val_features_loss: 0.0146 - val_dense_2_loss: 0.1300 - val_features_each_img_distance: 4.9136 - val_features_average_center_distance: 18.0359 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9676\n",
            "Epoch 45/50\n",
            "456/456 [==============================] - 36s 78ms/step - loss: 0.0167 - features_loss: 0.0132 - dense_2_loss: 0.0035 - features_each_img_distance: 3.8454 - features_average_center_distance: 19.5182 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9995 - val_loss: 0.1576 - val_features_loss: 0.0184 - val_dense_2_loss: 0.1391 - val_features_each_img_distance: 5.5506 - val_features_average_center_distance: 18.8265 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9621\n",
            "Epoch 46/50\n",
            "456/456 [==============================] - 36s 78ms/step - loss: 0.0122 - features_loss: 0.0104 - dense_2_loss: 0.0017 - features_each_img_distance: 3.4619 - features_average_center_distance: 18.7013 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9999 - val_loss: 0.1466 - val_features_loss: 0.0144 - val_dense_2_loss: 0.1322 - val_features_each_img_distance: 4.8847 - val_features_average_center_distance: 18.1059 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9619\n",
            "Epoch 47/50\n",
            "456/456 [==============================] - 35s 78ms/step - loss: 0.0105 - features_loss: 0.0089 - dense_2_loss: 0.0016 - features_each_img_distance: 3.1996 - features_average_center_distance: 18.3663 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9999 - val_loss: 0.1385 - val_features_loss: 0.0135 - val_dense_2_loss: 0.1249 - val_features_each_img_distance: 4.7060 - val_features_average_center_distance: 17.9895 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9642\n",
            "Epoch 48/50\n",
            "456/456 [==============================] - 35s 77ms/step - loss: 0.0111 - features_loss: 0.0091 - dense_2_loss: 0.0020 - features_each_img_distance: 3.2330 - features_average_center_distance: 18.4907 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9999 - val_loss: 0.1381 - val_features_loss: 0.0130 - val_dense_2_loss: 0.1251 - val_features_each_img_distance: 4.6060 - val_features_average_center_distance: 17.8036 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9642\n",
            "Epoch 49/50\n",
            "456/456 [==============================] - 37s 82ms/step - loss: 0.0101 - features_loss: 0.0085 - dense_2_loss: 0.0016 - features_each_img_distance: 3.1213 - features_average_center_distance: 18.3154 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9998 - val_loss: 0.1397 - val_features_loss: 0.0124 - val_dense_2_loss: 0.1272 - val_features_each_img_distance: 4.4808 - val_features_average_center_distance: 17.6727 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9647\n",
            "Epoch 50/50\n",
            "456/456 [==============================] - 35s 78ms/step - loss: 0.0104 - features_loss: 0.0083 - dense_2_loss: 0.0021 - features_each_img_distance: 3.0844 - features_average_center_distance: 18.1852 - features_categorical_accuracy: 0.0000e+00 - dense_2_categorical_accuracy: 0.9997 - val_loss: 0.1427 - val_features_loss: 0.0129 - val_dense_2_loss: 0.1298 - val_features_each_img_distance: 4.5620 - val_features_average_center_distance: 17.8285 - val_features_categorical_accuracy: 0.0000e+00 - val_dense_2_categorical_accuracy: 0.9653\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.save_weights('/content/drive/MyDrive/centerloss.h5')"
      ],
      "metadata": {
        "id": "TS5r4Lune8eK"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "fig = plt.figure(figsize = (15, 5))\n",
        "ax = fig.add_subplot(1, 2, 1)\n",
        "ax.plot(history_with_centerloss1.history[\"features_loss\"], label = \"center loss on feature layer - train\")\n",
        "ax.plot(history_with_centerloss1.history[\"val_features_loss\"], label = \"center loss on feature layer - test\")\n",
        "ax.set_title(\"Loss\")\n",
        "ax.legend()\n",
        "\n",
        "ax = fig.add_subplot(1, 2, 2)\n",
        "ax.plot(history_with_centerloss1.history[\"features_each_img_distance\"], label = \"distance of each image from its center - train\")\n",
        "ax.plot(history_with_centerloss1.history[\"val_features_each_img_distance\"], label = \"distance of each image from its center - test\")\n",
        "ax.plot(history_with_centerloss1.history[\"features_average_center_distance\"], label = \"average distance between all the centers - train\")\n",
        "ax.plot(history_with_centerloss1.history[\"val_features_average_center_distance\"], label = \"average distance between all the centers - test\")\n",
        "ax.set_title(\"Metrics\")\n",
        "ax.legend()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 466
        },
        "id": "qTCalrt-pVyw",
        "outputId": "ac17ffcd-4072-4668-9208-9ebdce2732ff"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7a1980b95330>"
            ]
          },
          "metadata": {},
          "execution_count": 46
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1500x500 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# استخراج مقادیر متریک‌ها از تاریخچه\n",
        "each_img_distance_train = history_with_centerloss1.history['features_each_img_distance']\n",
        "average_center_distance_train = history_with_centerloss1.history['features_average_center_distance']\n",
        "loss_train = history_with_centerloss1.history['loss']\n",
        "accuracy_train = history_with_centerloss1.history['dense_2_categorical_accuracy']\n",
        "\n",
        "each_img_distance_val = history_with_centerloss1.history['val_features_each_img_distance']\n",
        "average_center_distance_val = history_with_centerloss1.history['val_features_average_center_distance']\n",
        "loss_val = history_with_centerloss1.history['val_loss']\n",
        "accuracy_val = history_with_centerloss1.history['val_dense_2_categorical_accuracy']\n",
        "\n",
        "\n",
        "fig, axes = plt.subplots(2, 2, figsize=(12, 8))\n",
        "\n",
        "\n",
        "axes[0, 0].plot(each_img_distance_train, label='Train')\n",
        "axes[0, 0].plot(each_img_distance_val, label='Validation')\n",
        "axes[0, 0].set_title('each_img_distance')\n",
        "axes[0, 0].set_xlabel('Epochs')\n",
        "axes[0, 0].set_ylabel('Value')\n",
        "axes[0, 0].legend()\n",
        "\n",
        "\n",
        "axes[0, 1].plot(average_center_distance_train, label='Train')\n",
        "axes[0, 1].plot(average_center_distance_val, label='Validation')\n",
        "axes[0, 1].set_title('average_center_distance')\n",
        "axes[0, 1].set_xlabel('Epochs')\n",
        "axes[0, 1].set_ylabel('Value')\n",
        "axes[0, 1].legend()\n",
        "\n",
        "\n",
        "axes[1, 0].plot(loss_train, label='Train')\n",
        "axes[1, 0].plot(loss_val, label='Validation')\n",
        "axes[1, 0].set_title('Loss')\n",
        "axes[1, 0].set_xlabel('Epochs')\n",
        "axes[1, 0].set_ylabel('Value')\n",
        "axes[1, 0].legend()\n",
        "\n",
        "\n",
        "axes[1, 1].plot(accuracy_train, label='Train')\n",
        "axes[1, 1].plot(accuracy_val, label='Validation')\n",
        "axes[1, 1].set_title('Accuracy')\n",
        "axes[1, 1].set_xlabel('Epochs')\n",
        "axes[1, 1].set_ylabel('Value')\n",
        "axes[1, 1].legend()\n",
        "\n",
        "\n",
        "plt.tight_layout()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 787
        },
        "id": "d6wEa2FHobSh",
        "outputId": "d29cf2cd-d067-4757-8872-dbdb9ed85bab"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1200x800 with 4 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "<h2>Load Model</h2>"
      ],
      "metadata": {
        "id": "Tr3URSS8tD6N"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model_builder1 = Models(shape, features_dims, num_classes)\n",
        "\n",
        "model = model_builder1.Resnet50()\n",
        "model.trainable=True\n",
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bWeXSOrqtImj",
        "outputId": "45050ff9-e1ea-4614-9b24-446b98e28040"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_3\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_8 (InputLayer)        [(None, 72, 72, 3)]       0         \n",
            "                                                                 \n",
            " resnet50 (Functional)       (None, None, None, 2048)  23587712  \n",
            "                                                                 \n",
            " global_average_pooling2d_3   (None, 2048)             0         \n",
            " (GlobalAveragePooling2D)                                        \n",
            "                                                                 \n",
            " features (Dense)            (None, 2048)              4196352   \n",
            "                                                                 \n",
            " dense_3 (Dense)             (None, 80)                163920    \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 27,947,984\n",
            "Trainable params: 27,894,864\n",
            "Non-trainable params: 53,120\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.load_weights('/content/drive/MyDrive/centerloss.h5')"
      ],
      "metadata": {
        "id": "UI3c-0FeuKDi"
      },
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "feature_extractor = tf.keras.models.Model(inputs=model.input, outputs=model.get_layer('features').output)"
      ],
      "metadata": {
        "id": "1V_PrwCfDreQ"
      },
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Load the feature extractor model\n",
        "feature_extractor = tf.keras.models.Model(inputs=model.input, outputs=model.get_layer('features').output)\n",
        "# feature_extractor1 = tf.keras.models.Model(inputs=model.input, outputs=model.get_layer('features').output)\n",
        "features=feature_extractor.predict(x_valid)\n",
        "# feature_extractor1.predict(x_valid)\n",
        "# are_equal = np.allclose(feature_extractor, feature_extractor1)\n",
        "# print(are_equal)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZHX0zN4cu5Hu",
        "outputId": "f31da50c-9933-4c7d-ccc1-cab563f9ec4b"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "170/170 [==============================] - 8s 20ms/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "features.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Ghi_e6p2wmTY",
        "outputId": "d3739fab-99bd-4c7e-fe3f-556bdf822393"
      },
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(5440, 2048)"
            ]
          },
          "metadata": {},
          "execution_count": 41
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.manifold import TSNE\n",
        "\n",
        "tsne = TSNE(n_components=2, random_state=42)\n",
        "tsne_vectors = tsne.fit_transform(features)\n",
        "# tsne_centers=  tsne.fit_transform(center_loss_obj.centers)\n"
      ],
      "metadata": {
        "id": "lmUq3-quwyR1"
      },
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(center_loss_obj.centers[0])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hRvdnbHDxsnP",
        "outputId": "85911604-a4a1-43c1-945c-1ce0086147a8"
      },
      "execution_count": 73,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "tf.Tensor(\n",
            "[2.5278434e-02 2.3338150e-03 8.9301565e-04 ... 1.3507958e-01 2.3335249e+00\n",
            " 3.0531641e-04], shape=(2048,), dtype=float32)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "plt.scatter(tsne_vectors[:, 0], tsne_vectors[:, 1], c=y_valid, cmap='tab20')\n",
        "# plt.scatter(tsne_centers[:, 0], tsne_centers[:, 1], c='red', marker='x', label='Centers')\n",
        "plt.colorbar()\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 467
        },
        "id": "0NKMfgkWxFAc",
        "outputId": "2d6a9c4a-ae8b-45a0-f916-c0317618c05f"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:matplotlib.legend:No artists with labels found to put in legend.  Note that artists whose label start with an underscore are ignored when legend() is called with no argument.\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "with open('/content/drive/MyDrive/resized_test_data.pickle', 'rb') as f:\n",
        "    x_test, y_test = pickle.load(f)"
      ],
      "metadata": {
        "id": "p15heI-_y185"
      },
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# imgs=x_test.copy()\n",
        "# labels=y_test.copy()\n",
        "\n",
        "indices_1 =np.random.choice(np.where(y_test == 1)[0],size=500)\n",
        "indices_0 =np.random.choice( np.where(y_test == 0)[0],size=500)\n",
        "inc=np.concatenate([indices_1,indices_0])\n",
        "# print(inc)\n",
        "\n",
        "imgs=x_test[inc]\n",
        "labels=y_test[inc]\n",
        "\n",
        "print(imgs.shape,labels.shape)\n",
        "print(labels)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "w4YUA8d4y_ED",
        "outputId": "e27db10d-66ef-4d38-95ee-20222a85250a"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(1000, 2, 72, 72, 3) (1000,)\n",
            "[1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
            " 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
            " 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
            " 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
            " 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
            " 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
            " 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
            " 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
            " 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
            " 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
            " 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
            " 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
            " 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
            " 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            " 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            " 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            " 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            " 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            " 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            " 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            " 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            " 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            " 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            " 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            " 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            " 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            " 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
            " 0]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "similarities = np.zeros(len(imgs))\n",
        "for i in range(len(imgs)):\n",
        "    print(imgs[i][0].shape)\n",
        "    image1 = np.expand_dims(imgs[i][0], axis=0)\n",
        "    image2 = np.expand_dims(imgs[i][1], axis=0)\n",
        "    image1=image1.astype('float32')\n",
        "    image2=image2.astype('float32')\n",
        "\n",
        "    image1/=255\n",
        "    image2/=255\n",
        "    print(image1.shape)\n",
        "    features1 = feature_extractor.predict(image1)\n",
        "    features2 = feature_extractor.predict(image2)\n",
        "    print(features1.shape,features2.shape)\n",
        "    similarity = cosine_similarity(features1, features2)\n",
        "    similarities[i] = similarity[0, 0]\n",
        "    print(i)\n",
        "print(similarities)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OEaC9CKyyqH4",
        "outputId": "c798284c-7bc5-43ab-8f30-bdf99c7867d9"
      },
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1;30;43mStreaming output truncated to the last 5000 lines.\u001b[0m\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "194\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "195\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "196\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "197\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "198\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "199\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "200\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "201\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "202\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "203\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "204\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "205\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 58ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "206\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "207\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 52ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "208\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "209\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "210\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "211\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "212\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "213\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "214\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 49ms/step\n",
            "1/1 [==============================] - 0s 50ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "215\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "216\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "217\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "218\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "219\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "220\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "221\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "222\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "223\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 45ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "224\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "225\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "226\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "227\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "228\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "229\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "230\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "231\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "232\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 32ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "233\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "234\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "235\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "236\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "237\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "238\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "239\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "240\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "241\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "242\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "243\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "244\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "245\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "246\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "247\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "248\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "249\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "250\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "251\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "252\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "253\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "254\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "255\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "256\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "257\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "258\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "259\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "260\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "261\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "262\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "263\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "264\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "265\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "266\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "267\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "268\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "269\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "270\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "271\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "272\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "273\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "274\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "275\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "276\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "277\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "278\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "279\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "280\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "281\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "282\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "283\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "284\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "285\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "1/1 [==============================] - 0s 50ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "286\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "287\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "288\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "289\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "290\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "1/1 [==============================] - 0s 52ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "291\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "292\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "293\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "294\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "295\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "296\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "297\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "298\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "299\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 49ms/step\n",
            "1/1 [==============================] - 0s 53ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "300\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "301\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "302\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "303\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "304\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "305\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 32ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "306\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "307\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "308\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "309\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "310\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "311\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "312\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "313\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "314\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "315\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "316\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "317\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "318\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "319\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "320\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "321\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "322\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "323\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "324\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "325\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 32ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "326\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "327\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "328\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "329\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "330\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "331\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "332\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "333\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "334\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "335\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "336\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "337\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "338\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "339\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "340\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "341\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "342\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "343\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "344\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 32ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "345\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "346\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "347\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "348\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "349\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "350\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 32ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "351\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "352\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "353\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "354\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "355\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "356\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 32ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "357\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "358\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "359\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "360\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "361\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "362\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 34ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "363\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "364\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "365\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "366\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "367\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "368\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "369\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "370\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "371\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "372\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "373\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "374\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 47ms/step\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "375\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 55ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "376\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "377\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "378\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 50ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "379\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "380\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "381\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "382\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "383\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "384\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 53ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "385\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "386\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 56ms/step\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "387\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "388\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "389\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 34ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "390\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "391\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "392\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "393\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "394\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "395\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "396\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "397\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "398\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "399\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "400\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "401\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "402\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "403\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 32ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "404\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "405\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "406\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "407\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "408\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 31ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "409\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "410\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "411\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "412\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "413\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "414\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "415\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "416\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "417\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "418\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "419\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "420\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "421\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "422\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "423\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "424\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "425\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "426\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "427\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "428\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "429\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "430\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "431\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "432\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "433\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "434\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 31ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "435\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "436\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "437\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "438\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "439\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "440\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "441\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "442\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "443\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "444\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "445\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "446\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 31ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "447\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "448\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "449\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "450\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "451\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "1/1 [==============================] - 0s 34ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "452\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "1/1 [==============================] - 0s 51ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "453\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 64ms/step\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "454\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "455\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 45ms/step\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "456\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 47ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "457\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 51ms/step\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "458\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "459\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "460\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "1/1 [==============================] - 0s 51ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "461\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "462\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "463\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "1/1 [==============================] - 0s 45ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "464\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "465\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 60ms/step\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "466\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 52ms/step\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "467\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "468\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 45ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "469\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "470\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "471\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "472\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "473\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "474\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "475\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "476\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "477\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "478\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "479\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "480\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "481\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "482\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 32ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "483\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "484\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "485\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "486\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "487\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "488\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "489\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 31ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "490\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "491\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "492\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "493\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "494\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "495\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 32ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "496\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "497\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "498\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "499\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "500\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "501\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "502\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "503\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "504\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "505\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "506\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "507\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "508\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "509\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "510\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "511\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "512\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "513\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "514\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "515\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "516\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "517\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "518\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "519\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "520\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 34ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "521\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "522\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "523\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "524\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "525\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "526\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "527\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "528\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "529\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "530\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 50ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "531\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 51ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "532\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "533\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "534\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "535\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "536\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "537\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "538\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "539\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "540\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 52ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "541\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "1/1 [==============================] - 0s 57ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "542\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 50ms/step\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "543\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "544\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "545\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "546\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 53ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "547\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 51ms/step\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "548\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "549\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 49ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "550\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 63ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "551\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "552\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "553\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "554\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "555\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "556\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "557\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "558\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "559\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "560\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "561\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "562\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "563\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "564\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "565\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "566\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "567\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "568\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "569\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "570\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "571\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "572\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "573\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "574\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "575\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "576\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "577\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "578\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "579\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "580\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "581\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "582\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "583\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "584\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "585\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "586\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "587\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "588\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "589\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "590\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "591\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 32ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "592\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "593\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "594\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "595\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "596\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "597\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "598\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "599\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "600\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "601\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "602\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "603\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "604\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "605\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "606\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "607\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "608\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "609\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "610\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "611\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "612\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 45ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "613\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "1/1 [==============================] - 0s 50ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "614\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 53ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "615\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 64ms/step\n",
            "1/1 [==============================] - 0s 76ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "616\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "1/1 [==============================] - 0s 57ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "617\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "618\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "619\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "620\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "621\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "1/1 [==============================] - 0s 49ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "622\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "623\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "624\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 53ms/step\n",
            "1/1 [==============================] - 0s 47ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "625\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 47ms/step\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "626\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 61ms/step\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "627\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "628\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "629\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 53ms/step\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "630\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 50ms/step\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "631\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "1/1 [==============================] - 0s 54ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "632\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 53ms/step\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "633\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "634\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "635\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "636\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "637\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "638\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "639\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "640\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "641\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 31ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "642\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "643\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "644\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "645\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "646\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "647\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "648\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "649\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "650\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "651\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "652\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "653\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "654\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "655\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "656\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "657\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "658\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 45ms/step\n",
            "1/1 [==============================] - 0s 31ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "659\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 31ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "660\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "661\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "662\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 32ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "663\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "664\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "665\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "666\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "667\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "668\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 32ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "669\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "670\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "671\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "1/1 [==============================] - 0s 31ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "672\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "673\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "674\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 31ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "675\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "676\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "677\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "678\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 32ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "679\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "680\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 34ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "681\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 31ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "682\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "683\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "684\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "685\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 31ms/step\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "686\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "687\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 32ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "688\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "689\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 51ms/step\n",
            "1/1 [==============================] - 0s 47ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "690\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 52ms/step\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "691\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 57ms/step\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "692\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 54ms/step\n",
            "1/1 [==============================] - 0s 51ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "693\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 45ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "694\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "1/1 [==============================] - 0s 54ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "695\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 50ms/step\n",
            "1/1 [==============================] - 0s 60ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "696\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "697\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "698\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "1/1 [==============================] - 0s 49ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "699\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 63ms/step\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "700\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "701\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "702\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 51ms/step\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "703\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 53ms/step\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "704\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "705\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "706\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "707\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "1/1 [==============================] - 0s 52ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "708\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "709\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 51ms/step\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "710\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "711\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "712\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "713\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "714\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "715\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "716\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "717\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "718\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "719\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 31ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "720\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "721\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "722\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "723\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "724\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "725\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "726\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "727\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "728\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "729\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "730\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "731\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "732\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "733\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "734\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "735\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "736\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "737\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "738\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "739\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "740\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "741\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "742\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "743\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "744\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "745\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "746\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "747\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "748\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "749\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "750\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "751\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "752\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "753\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "754\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "755\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "756\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "757\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "758\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "759\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "760\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "761\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "762\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "763\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "764\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "765\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "766\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "767\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "768\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "769\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "770\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 53ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "771\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 45ms/step\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "772\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "773\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "774\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 47ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "775\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "776\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 34ms/step\n",
            "1/1 [==============================] - 0s 34ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "777\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 60ms/step\n",
            "1/1 [==============================] - 0s 71ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "778\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 80ms/step\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "779\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 52ms/step\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "780\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 67ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "781\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "1/1 [==============================] - 0s 62ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "782\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 63ms/step\n",
            "1/1 [==============================] - 0s 61ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "783\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "784\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 45ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "785\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "1/1 [==============================] - 0s 52ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "786\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 49ms/step\n",
            "1/1 [==============================] - 0s 45ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "787\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "788\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "789\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "790\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 73ms/step\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "791\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "792\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "793\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "794\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "795\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "796\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "797\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "798\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "799\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "800\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "801\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "802\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "803\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "804\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "805\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "806\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "807\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "808\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "809\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "810\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "811\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "812\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "813\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "814\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "815\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "816\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "817\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "818\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "819\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "820\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "821\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "822\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "823\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 32ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "824\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "825\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "826\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "827\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "828\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "829\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "830\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "831\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "832\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "833\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "834\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "835\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "836\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "837\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "838\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "839\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "840\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "841\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "842\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "843\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "844\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "845\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "846\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "847\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "848\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "849\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "850\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "1/1 [==============================] - 0s 45ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "851\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 45ms/step\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "852\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "853\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 58ms/step\n",
            "1/1 [==============================] - 0s 47ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "854\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 50ms/step\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "855\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 34ms/step\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "856\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "1/1 [==============================] - 0s 50ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "857\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "858\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "859\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "860\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "861\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "862\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "1/1 [==============================] - 0s 59ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "863\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "864\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "865\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 53ms/step\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "866\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "1/1 [==============================] - 0s 51ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "867\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 56ms/step\n",
            "1/1 [==============================] - 0s 77ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "868\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "869\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "870\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "871\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "872\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "873\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 59ms/step\n",
            "1/1 [==============================] - 0s 58ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "874\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "1/1 [==============================] - 0s 45ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "875\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 49ms/step\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "876\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "877\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "878\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "879\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "880\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "881\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "882\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 32ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "883\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "884\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "885\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "886\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "887\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "888\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "889\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "890\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "891\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "892\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "893\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 31ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "894\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "895\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "896\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "897\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "898\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "899\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "900\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "901\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "902\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "903\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "904\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "905\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 34ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "906\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "907\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "908\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "909\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "910\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "911\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 34ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "912\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "913\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "914\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "915\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "916\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "917\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "918\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "919\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "920\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "921\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "922\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "923\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "924\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "925\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "926\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "927\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "928\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "929\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "930\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "931\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "932\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "933\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "934\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 59ms/step\n",
            "1/1 [==============================] - 0s 55ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "935\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "1/1 [==============================] - 0s 57ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "936\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 54ms/step\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "937\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 51ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "938\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "1/1 [==============================] - 0s 34ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "939\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 55ms/step\n",
            "1/1 [==============================] - 0s 49ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "940\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "941\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 34ms/step\n",
            "1/1 [==============================] - 0s 34ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "942\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "943\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 53ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "944\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 54ms/step\n",
            "1/1 [==============================] - 0s 63ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "945\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 73ms/step\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "946\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "1/1 [==============================] - 0s 34ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "947\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 34ms/step\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "948\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "1/1 [==============================] - 0s 37ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "949\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 55ms/step\n",
            "1/1 [==============================] - 0s 51ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "950\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "951\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 79ms/step\n",
            "1/1 [==============================] - 0s 57ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "952\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 63ms/step\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "953\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 59ms/step\n",
            "1/1 [==============================] - 0s 49ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "954\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 57ms/step\n",
            "1/1 [==============================] - 0s 51ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "955\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "956\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 48ms/step\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "957\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 53ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "958\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "959\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "960\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "961\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "962\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "963\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "964\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "965\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "966\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 31ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "967\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "968\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "969\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "970\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "971\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "972\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "973\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "974\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "975\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "976\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "977\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "978\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "979\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "980\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "981\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "982\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "983\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "984\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "985\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "986\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "987\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "988\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "989\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "990\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "991\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "992\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "993\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "994\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "995\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "996\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "997\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "998\n",
            "(72, 72, 3)\n",
            "(1, 72, 72, 3)\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "(1, 2048) (1, 2048)\n",
            "999\n",
            "[0.41267806 0.65190417 0.66185153 0.77581793 0.5942077  0.55570567\n",
            " 0.64649403 0.7559557  0.61947471 0.88887548 0.67392576 0.53543472\n",
            " 0.466791   0.74675632 0.57445139 0.33392924 0.52639258 0.69661617\n",
            " 0.56452656 0.62361878 0.59650695 0.8392241  0.70550191 0.77508563\n",
            " 0.77321953 0.62446964 0.65680617 0.72598809 0.63045108 0.70261335\n",
            " 0.51773268 0.75438923 0.65712845 0.61487257 0.65484178 0.52809596\n",
            " 0.71210432 0.78909826 0.46487874 0.60568053 0.83518672 0.69834119\n",
            " 0.69375539 0.75062704 0.49856436 0.62604976 0.68338454 0.54733914\n",
            " 0.58096904 0.60962439 0.47327265 0.74188018 0.70276713 0.86363947\n",
            " 0.73827934 0.78100771 0.64934814 0.56239879 0.64373034 0.77324861\n",
            " 0.48729438 0.64716929 0.73576593 0.76588118 0.69716769 0.46151724\n",
            " 0.71697116 0.58929765 0.71197712 0.66675413 0.54677176 0.6445449\n",
            " 0.62119353 0.6807968  0.694848   0.85305727 0.63932049 0.82016724\n",
            " 0.69298249 0.43298179 0.80018139 0.75298166 0.54512709 0.65460443\n",
            " 0.84666485 0.66674268 0.77644235 0.5962671  0.71917808 0.75238103\n",
            " 0.51386225 0.91043562 0.76499867 0.90128827 0.46751255 0.64592659\n",
            " 0.70693988 0.81137127 0.81255054 0.4979251  0.67650008 0.47142124\n",
            " 0.72329378 0.67415857 0.67873585 0.44543442 0.63464367 0.69580066\n",
            " 0.76721036 0.58720279 0.5739612  0.79808468 0.47042626 0.58481693\n",
            " 0.63697159 0.60722351 0.54020959 0.83084291 0.70821583 0.70695174\n",
            " 0.67172992 0.77137804 0.72546446 0.64797819 0.84224862 0.79640687\n",
            " 0.48729438 0.73389852 0.78909832 0.61244333 0.52993715 0.81806827\n",
            " 0.72986573 0.68973535 0.72063857 0.62616569 0.53177774 0.64683425\n",
            " 0.7830739  0.65046859 0.61649358 0.54280806 0.46600056 0.75103927\n",
            " 0.5515337  0.79751998 0.45686224 0.64346993 0.63556111 0.7291795\n",
            " 0.61562228 0.79251719 0.73623073 0.69449139 0.90234613 0.73827934\n",
            " 0.72662449 0.67286146 0.70314044 0.74414718 0.55533636 0.61662763\n",
            " 0.52845454 0.59373367 0.63044494 0.6785934  0.73220146 0.63242316\n",
            " 0.68237376 0.5938648  0.62523615 0.83490914 0.79591626 0.74830222\n",
            " 0.56079382 0.74554509 0.47724047 0.61598426 0.91941059 0.82175905\n",
            " 0.68247002 0.74871433 0.58830553 0.55132079 0.71991742 0.54748797\n",
            " 0.7291795  0.7651763  0.57982856 0.65286994 0.72416365 0.69276458\n",
            " 0.80874407 0.82089019 0.59075916 0.92607975 0.85688221 0.62837493\n",
            " 0.43519169 0.58591169 0.74310863 0.48978931 0.49503842 0.64828616\n",
            " 0.78048587 0.69364744 0.75191563 0.63944221 0.51919562 0.77243054\n",
            " 0.6469813  0.49512559 0.75816023 0.87556118 0.51222169 0.61696589\n",
            " 0.82933009 0.39516205 0.54585284 0.88263595 0.70759803 0.63996172\n",
            " 0.63549286 0.58916742 0.53129125 0.68066406 0.72446626 0.67455339\n",
            " 0.70067096 0.56102884 0.80759907 0.59835064 0.6633721  0.69493872\n",
            " 0.6571604  0.43107259 0.72895694 0.66272438 0.72352004 0.52570379\n",
            " 0.71114492 0.86285233 0.70268786 0.50225246 0.81847066 0.6641646\n",
            " 0.72141242 0.80603528 0.72241426 0.68191886 0.78831911 0.61341119\n",
            " 0.49053839 0.64180505 0.75948197 0.48885286 0.65480173 0.63516033\n",
            " 0.68316525 0.49147588 0.70697373 0.75238103 0.80245876 0.795726\n",
            " 0.71124184 0.86186278 0.75521678 0.84044272 0.53952324 0.61296791\n",
            " 0.69079411 0.69903111 0.72659671 0.70322537 0.4685728  0.67392576\n",
            " 0.62935042 0.80085498 0.6862092  0.50091815 0.61562824 0.65680617\n",
            " 0.67501628 0.82749993 0.6644097  0.66432726 0.73279977 0.69115639\n",
            " 0.56079382 0.8347975  0.57971799 0.59156066 0.45190904 0.63582033\n",
            " 0.7948752  0.39644721 0.47404292 0.69590098 0.65470457 0.80214196\n",
            " 0.69993126 0.57665354 0.30243567 0.55415356 0.72141325 0.9368633\n",
            " 0.7339468  0.79566163 0.6505084  0.63231742 0.47009516 0.44493473\n",
            " 0.54309297 0.40878102 0.63168848 0.85435086 0.57982856 0.5882684\n",
            " 0.74532092 0.68123645 0.61454356 0.71507645 0.6314286  0.52730805\n",
            " 0.63752961 0.49503842 0.75207949 0.66044086 0.55077356 0.75521678\n",
            " 0.54894936 0.54287714 0.75845242 0.6974594  0.49552619 0.43290752\n",
            " 0.76743317 0.72446626 0.52829754 0.65764993 0.70261335 0.68237376\n",
            " 0.66667897 0.39604098 0.50707543 0.58387065 0.62508488 0.6838367\n",
            " 0.6784246  0.70314044 0.74165535 0.47618705 0.83213967 0.641554\n",
            " 0.70067096 0.6641646  0.72941941 0.80274022 0.62030673 0.5868485\n",
            " 0.63322425 0.64449829 0.57565057 0.69381142 0.50725859 0.66356921\n",
            " 0.6019969  0.87311983 0.69620466 0.60426164 0.66888565 0.53231204\n",
            " 0.81850398 0.77644235 0.70808107 0.41902503 0.60471183 0.51456499\n",
            " 0.52669603 0.50706553 0.80889505 0.71825522 0.52218187 0.63261294\n",
            " 0.6256032  0.64959538 0.73133087 0.5164417  0.82839984 0.52966553\n",
            " 0.85799962 0.64684284 0.62683249 0.5517996  0.52730805 0.55744886\n",
            " 0.884206   0.68781072 0.48124409 0.55242884 0.68631971 0.31170711\n",
            " 0.70695174 0.46617436 0.75816023 0.66518503 0.63466078 0.82254934\n",
            " 0.57665354 0.60850579 0.76712978 0.83402437 0.84817171 0.71103001\n",
            " 0.77306652 0.68066406 0.58328795 0.69718194 0.46176332 0.83326066\n",
            " 0.69973928 0.60263687 0.53970307 0.57144588 0.51222169 0.89590645\n",
            " 0.59165454 0.75042415 0.65109062 0.76029253 0.56518799 0.61837852\n",
            " 0.72804719 0.55714452 0.60501325 0.66431743 0.79323888 0.64590102\n",
            " 0.69254988 0.59365368 0.65738046 0.77433872 0.81186414 0.80018532\n",
            " 0.63861436 0.6291523  0.63515645 0.6186462  0.76155794 0.58784348\n",
            " 0.42958939 0.83326066 0.73127675 0.74731016 0.44436216 0.35735422\n",
            " 0.59835064 0.63996172 0.65014309 0.7884748  0.79175162 0.75549638\n",
            " 0.53688884 0.66190022 0.63564962 0.58045584 0.66961116 0.65088773\n",
            " 0.70808107 0.6516453  0.59624374 0.65312183 0.5663327  0.62524498\n",
            " 0.58506918 0.83987999 0.86531627 0.47735837 0.9045344  0.71114492\n",
            " 0.66248345 0.87849283 0.72912157 0.70836836 0.66543412 0.47218305\n",
            " 0.72804129 0.52262139 0.55492699 0.68908644 0.79251719 0.72837621\n",
            " 0.70710742 0.47184783 0.74717224 0.5922969  0.65140104 0.69222558\n",
            " 0.70620602 0.71534646 0.43569311 0.78579384 0.58490932 0.45833969\n",
            " 0.58294666 0.54966021 0.66972315 0.62138605 0.60773516 0.47433466\n",
            " 0.21761684 0.58312708 0.57422405 0.31880611 0.75101221 0.63825035\n",
            " 0.5523873  0.47596604 0.61696064 0.63400954 0.68299574 0.46531391\n",
            " 0.48002392 0.28945109 0.73334062 0.53267413 0.46544221 0.6176911\n",
            " 0.52336848 0.55286789 0.48377743 0.51309204 0.31880611 0.45747793\n",
            " 0.27378398 0.27306724 0.47418743 0.57248056 0.4424969  0.46643275\n",
            " 0.61970061 0.65609729 0.66207844 0.39632857 0.54219466 0.48621842\n",
            " 0.38394612 0.65067542 0.39613256 0.20587635 0.54905528 0.61157149\n",
            " 0.45754921 0.45209652 0.70514619 0.6331206  0.62151444 0.44961405\n",
            " 0.47059271 0.44399151 0.46213728 0.61906385 0.45685241 0.46269304\n",
            " 0.54212123 0.31711131 0.40302616 0.57442367 0.42239451 0.58052319\n",
            " 0.74619526 0.52426636 0.8257587  0.60703117 0.5910027  0.56473929\n",
            " 0.43477574 0.58034521 0.58200407 0.63202679 0.35730889 0.64030492\n",
            " 0.58820212 0.64513063 0.61293721 0.24278252 0.63434881 0.5377295\n",
            " 0.51988918 0.70158482 0.46367633 0.60589415 0.40941137 0.61860764\n",
            " 0.69706237 0.55007315 0.54212123 0.46831644 0.60612738 0.60496294\n",
            " 0.71522516 0.47747508 0.61075062 0.23725258 0.4733265  0.7903623\n",
            " 0.51284724 0.59194642 0.52288866 0.50771666 0.50707799 0.62415993\n",
            " 0.26954767 0.57255888 0.49191594 0.63265699 0.52617222 0.54296434\n",
            " 0.54914129 0.57672268 0.54136252 0.60713458 0.31448957 0.52724797\n",
            " 0.5860213  0.37020719 0.43507901 0.3748205  0.53366339 0.46236417\n",
            " 0.51411515 0.41216385 0.32756609 0.53225464 0.30567387 0.32737568\n",
            " 0.40457022 0.36400723 0.63207936 0.34279591 0.28612572 0.43566167\n",
            " 0.53090119 0.39070228 0.5523873  0.55482137 0.60428119 0.7020818\n",
            " 0.48874682 0.56420004 0.58790314 0.31065685 0.49968627 0.6000284\n",
            " 0.65506339 0.62412435 0.43385202 0.47673672 0.56255275 0.42034459\n",
            " 0.45754921 0.57133758 0.58696032 0.33738762 0.47418743 0.65335763\n",
            " 0.6640417  0.24509022 0.64645392 0.60690308 0.34587181 0.55262744\n",
            " 0.58671391 0.5121237  0.50382167 0.46442062 0.40668207 0.60304952\n",
            " 0.61382926 0.55446786 0.48031792 0.59170455 0.5657028  0.57674879\n",
            " 0.67510438 0.28326011 0.56658769 0.72263491 0.33166465 0.45295438\n",
            " 0.49186778 0.52054298 0.38708556 0.33848614 0.50315094 0.63051873\n",
            " 0.61799896 0.63750023 0.549968   0.30086905 0.61696064 0.60852355\n",
            " 0.44458047 0.7541092  0.28133529 0.53729856 0.40229195 0.77674341\n",
            " 0.50831234 0.17126779 0.66197491 0.74910069 0.44391793 0.37521109\n",
            " 0.48939526 0.34703001 0.40833777 0.65295798 0.4791781  0.45489848\n",
            " 0.58925176 0.47193176 0.78375554 0.46136689 0.5356853  0.58105123\n",
            " 0.49530569 0.49715063 0.68066132 0.66737926 0.31233513 0.32936025\n",
            " 0.51988918 0.57877553 0.59737194 0.56441456 0.773453   0.43180269\n",
            " 0.45735243 0.40011209 0.62914771 0.61825573 0.60713458 0.46467328\n",
            " 0.54051805 0.41381255 0.34609818 0.53391534 0.35352722 0.27819821\n",
            " 0.53346992 0.50301647 0.39843854 0.54615366 0.60562396 0.38752091\n",
            " 0.65435123 0.61744159 0.64231199 0.44887596 0.73832536 0.44895184\n",
            " 0.38261622 0.46275279 0.34426367 0.61812001 0.36492366 0.43792129\n",
            " 0.60887241 0.64551592 0.48533034 0.51256347 0.55559707 0.46442062\n",
            " 0.47677898 0.68355954 0.4566268  0.57250583 0.55007315 0.44765741\n",
            " 0.51800406 0.58079052 0.76695812 0.63730586 0.38602185 0.5676617\n",
            " 0.44464299 0.46675563 0.71789777 0.69727051 0.57089251 0.47746086\n",
            " 0.65742528 0.29573834 0.3465457  0.58196974 0.49511582 0.48030153\n",
            " 0.52655596 0.6117717  0.61906385 0.69888955 0.6570344  0.61753821\n",
            " 0.73173833 0.63928646 0.34568062 0.38752091 0.58529496 0.34128904\n",
            " 0.32982939 0.39240777 0.49432641 0.69187623 0.58451504 0.76149786\n",
            " 0.71897113 0.3551026  0.44152588 0.61688256 0.6576274  0.32671028\n",
            " 0.580477   0.47433466 0.43720609 0.73415744 0.73791474 0.58792198\n",
            " 0.51605368 0.53399724 0.44745314 0.38996732 0.56645042 0.42578262\n",
            " 0.52729523 0.58597648 0.39070228 0.43261695 0.35304865 0.66649163\n",
            " 0.64641464 0.5238474  0.31648529 0.61799896 0.45443982 0.6185149\n",
            " 0.49499139 0.62024975 0.39910114 0.47517264 0.45078072 0.51782656\n",
            " 0.58129168 0.59477448 0.40806431 0.40105671 0.42226315 0.53609365\n",
            " 0.53391534 0.74669164 0.47222966 0.43550634 0.66385293 0.37483907\n",
            " 0.5588578  0.5302397  0.58382785 0.62914771 0.59981942 0.62727177\n",
            " 0.54272127 0.56257105 0.65682113 0.6862846  0.56215119 0.53539538\n",
            " 0.66638219 0.59456801 0.29882464 0.72836542 0.48496625 0.38596606\n",
            " 0.58868885 0.71227944 0.40302616 0.57187748 0.4467774  0.55514991\n",
            " 0.52902174 0.40169775 0.67066681 0.51230645 0.46888459 0.591438\n",
            " 0.5057649  0.62555921 0.43957514 0.53322309 0.44399151 0.5837971\n",
            " 0.40031523 0.77307892 0.79925388 0.43721479 0.55469048 0.14982629\n",
            " 0.47788402 0.6142748  0.4467774  0.60608196 0.32893771 0.45153898\n",
            " 0.65128613 0.69666481 0.50727266 0.54315662 0.36059266 0.28133529\n",
            " 0.43598494 0.49577883 0.63220286 0.57314456 0.67640823 0.65058291\n",
            " 0.43026781 0.42903385 0.58833838 0.53097892 0.35476106 0.58542705\n",
            " 0.60891044 0.56446457 0.64603752 0.59509587 0.42226315 0.6570344\n",
            " 0.44399151 0.66388541 0.43617702 0.580477   0.5128926  0.63977528\n",
            " 0.23017299 0.58542705 0.64048946 0.47021884 0.66094387 0.525594\n",
            " 0.28159726 0.55258822 0.53932959 0.23017299 0.24278252 0.56130981\n",
            " 0.44496036 0.47036594 0.56557792 0.35807586 0.63537335 0.55540645\n",
            " 0.61508632 0.76499581 0.65281498 0.72936916 0.46831644 0.78673095\n",
            " 0.43307167 0.55681705 0.65905678 0.79228693 0.50464821 0.59392595\n",
            " 0.40536547 0.37410533 0.79925388 0.47667557 0.19723308 0.54568708\n",
            " 0.46903133 0.57672268 0.64749146 0.55422443 0.49037924 0.55439365\n",
            " 0.51202643 0.31603727 0.301864   0.58920276 0.62757468 0.54004371\n",
            " 0.58409607 0.44185543 0.60083437 0.46236417 0.59034556 0.33159187\n",
            " 0.67859483 0.71741176 0.32991922 0.47831276]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(similarities[similarities<0.95])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8CBxLjJALSHe",
        "outputId": "d11f7cb6-5af9-4b16-dd78-b672221ed349"
      },
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[0.41267806 0.65190417 0.66185153 0.77581793 0.5942077  0.55570567\n",
            " 0.64649403 0.7559557  0.61947471 0.88887548 0.67392576 0.53543472\n",
            " 0.466791   0.74675632 0.57445139 0.33392924 0.52639258 0.69661617\n",
            " 0.56452656 0.62361878 0.59650695 0.8392241  0.70550191 0.77508563\n",
            " 0.77321953 0.62446964 0.65680617 0.72598809 0.63045108 0.70261335\n",
            " 0.51773268 0.75438923 0.65712845 0.61487257 0.65484178 0.52809596\n",
            " 0.71210432 0.78909826 0.46487874 0.60568053 0.83518672 0.69834119\n",
            " 0.69375539 0.75062704 0.49856436 0.62604976 0.68338454 0.54733914\n",
            " 0.58096904 0.60962439 0.47327265 0.74188018 0.70276713 0.86363947\n",
            " 0.73827934 0.78100771 0.64934814 0.56239879 0.64373034 0.77324861\n",
            " 0.48729438 0.64716929 0.73576593 0.76588118 0.69716769 0.46151724\n",
            " 0.71697116 0.58929765 0.71197712 0.66675413 0.54677176 0.6445449\n",
            " 0.62119353 0.6807968  0.694848   0.85305727 0.63932049 0.82016724\n",
            " 0.69298249 0.43298179 0.80018139 0.75298166 0.54512709 0.65460443\n",
            " 0.84666485 0.66674268 0.77644235 0.5962671  0.71917808 0.75238103\n",
            " 0.51386225 0.91043562 0.76499867 0.90128827 0.46751255 0.64592659\n",
            " 0.70693988 0.81137127 0.81255054 0.4979251  0.67650008 0.47142124\n",
            " 0.72329378 0.67415857 0.67873585 0.44543442 0.63464367 0.69580066\n",
            " 0.76721036 0.58720279 0.5739612  0.79808468 0.47042626 0.58481693\n",
            " 0.63697159 0.60722351 0.54020959 0.83084291 0.70821583 0.70695174\n",
            " 0.67172992 0.77137804 0.72546446 0.64797819 0.84224862 0.79640687\n",
            " 0.48729438 0.73389852 0.78909832 0.61244333 0.52993715 0.81806827\n",
            " 0.72986573 0.68973535 0.72063857 0.62616569 0.53177774 0.64683425\n",
            " 0.7830739  0.65046859 0.61649358 0.54280806 0.46600056 0.75103927\n",
            " 0.5515337  0.79751998 0.45686224 0.64346993 0.63556111 0.7291795\n",
            " 0.61562228 0.79251719 0.73623073 0.69449139 0.90234613 0.73827934\n",
            " 0.72662449 0.67286146 0.70314044 0.74414718 0.55533636 0.61662763\n",
            " 0.52845454 0.59373367 0.63044494 0.6785934  0.73220146 0.63242316\n",
            " 0.68237376 0.5938648  0.62523615 0.83490914 0.79591626 0.74830222\n",
            " 0.56079382 0.74554509 0.47724047 0.61598426 0.91941059 0.82175905\n",
            " 0.68247002 0.74871433 0.58830553 0.55132079 0.71991742 0.54748797\n",
            " 0.7291795  0.7651763  0.57982856 0.65286994 0.72416365 0.69276458\n",
            " 0.80874407 0.82089019 0.59075916 0.92607975 0.85688221 0.62837493\n",
            " 0.43519169 0.58591169 0.74310863 0.48978931 0.49503842 0.64828616\n",
            " 0.78048587 0.69364744 0.75191563 0.63944221 0.51919562 0.77243054\n",
            " 0.6469813  0.49512559 0.75816023 0.87556118 0.51222169 0.61696589\n",
            " 0.82933009 0.39516205 0.54585284 0.88263595 0.70759803 0.63996172\n",
            " 0.63549286 0.58916742 0.53129125 0.68066406 0.72446626 0.67455339\n",
            " 0.70067096 0.56102884 0.80759907 0.59835064 0.6633721  0.69493872\n",
            " 0.6571604  0.43107259 0.72895694 0.66272438 0.72352004 0.52570379\n",
            " 0.71114492 0.86285233 0.70268786 0.50225246 0.81847066 0.6641646\n",
            " 0.72141242 0.80603528 0.72241426 0.68191886 0.78831911 0.61341119\n",
            " 0.49053839 0.64180505 0.75948197 0.48885286 0.65480173 0.63516033\n",
            " 0.68316525 0.49147588 0.70697373 0.75238103 0.80245876 0.795726\n",
            " 0.71124184 0.86186278 0.75521678 0.84044272 0.53952324 0.61296791\n",
            " 0.69079411 0.69903111 0.72659671 0.70322537 0.4685728  0.67392576\n",
            " 0.62935042 0.80085498 0.6862092  0.50091815 0.61562824 0.65680617\n",
            " 0.67501628 0.82749993 0.6644097  0.66432726 0.73279977 0.69115639\n",
            " 0.56079382 0.8347975  0.57971799 0.59156066 0.45190904 0.63582033\n",
            " 0.7948752  0.39644721 0.47404292 0.69590098 0.65470457 0.80214196\n",
            " 0.69993126 0.57665354 0.30243567 0.55415356 0.72141325 0.9368633\n",
            " 0.7339468  0.79566163 0.6505084  0.63231742 0.47009516 0.44493473\n",
            " 0.54309297 0.40878102 0.63168848 0.85435086 0.57982856 0.5882684\n",
            " 0.74532092 0.68123645 0.61454356 0.71507645 0.6314286  0.52730805\n",
            " 0.63752961 0.49503842 0.75207949 0.66044086 0.55077356 0.75521678\n",
            " 0.54894936 0.54287714 0.75845242 0.6974594  0.49552619 0.43290752\n",
            " 0.76743317 0.72446626 0.52829754 0.65764993 0.70261335 0.68237376\n",
            " 0.66667897 0.39604098 0.50707543 0.58387065 0.62508488 0.6838367\n",
            " 0.6784246  0.70314044 0.74165535 0.47618705 0.83213967 0.641554\n",
            " 0.70067096 0.6641646  0.72941941 0.80274022 0.62030673 0.5868485\n",
            " 0.63322425 0.64449829 0.57565057 0.69381142 0.50725859 0.66356921\n",
            " 0.6019969  0.87311983 0.69620466 0.60426164 0.66888565 0.53231204\n",
            " 0.81850398 0.77644235 0.70808107 0.41902503 0.60471183 0.51456499\n",
            " 0.52669603 0.50706553 0.80889505 0.71825522 0.52218187 0.63261294\n",
            " 0.6256032  0.64959538 0.73133087 0.5164417  0.82839984 0.52966553\n",
            " 0.85799962 0.64684284 0.62683249 0.5517996  0.52730805 0.55744886\n",
            " 0.884206   0.68781072 0.48124409 0.55242884 0.68631971 0.31170711\n",
            " 0.70695174 0.46617436 0.75816023 0.66518503 0.63466078 0.82254934\n",
            " 0.57665354 0.60850579 0.76712978 0.83402437 0.84817171 0.71103001\n",
            " 0.77306652 0.68066406 0.58328795 0.69718194 0.46176332 0.83326066\n",
            " 0.69973928 0.60263687 0.53970307 0.57144588 0.51222169 0.89590645\n",
            " 0.59165454 0.75042415 0.65109062 0.76029253 0.56518799 0.61837852\n",
            " 0.72804719 0.55714452 0.60501325 0.66431743 0.79323888 0.64590102\n",
            " 0.69254988 0.59365368 0.65738046 0.77433872 0.81186414 0.80018532\n",
            " 0.63861436 0.6291523  0.63515645 0.6186462  0.76155794 0.58784348\n",
            " 0.42958939 0.83326066 0.73127675 0.74731016 0.44436216 0.35735422\n",
            " 0.59835064 0.63996172 0.65014309 0.7884748  0.79175162 0.75549638\n",
            " 0.53688884 0.66190022 0.63564962 0.58045584 0.66961116 0.65088773\n",
            " 0.70808107 0.6516453  0.59624374 0.65312183 0.5663327  0.62524498\n",
            " 0.58506918 0.83987999 0.86531627 0.47735837 0.9045344  0.71114492\n",
            " 0.66248345 0.87849283 0.72912157 0.70836836 0.66543412 0.47218305\n",
            " 0.72804129 0.52262139 0.55492699 0.68908644 0.79251719 0.72837621\n",
            " 0.70710742 0.47184783 0.74717224 0.5922969  0.65140104 0.69222558\n",
            " 0.70620602 0.71534646 0.43569311 0.78579384 0.58490932 0.45833969\n",
            " 0.58294666 0.54966021 0.66972315 0.62138605 0.60773516 0.47433466\n",
            " 0.21761684 0.58312708 0.57422405 0.31880611 0.75101221 0.63825035\n",
            " 0.5523873  0.47596604 0.61696064 0.63400954 0.68299574 0.46531391\n",
            " 0.48002392 0.28945109 0.73334062 0.53267413 0.46544221 0.6176911\n",
            " 0.52336848 0.55286789 0.48377743 0.51309204 0.31880611 0.45747793\n",
            " 0.27378398 0.27306724 0.47418743 0.57248056 0.4424969  0.46643275\n",
            " 0.61970061 0.65609729 0.66207844 0.39632857 0.54219466 0.48621842\n",
            " 0.38394612 0.65067542 0.39613256 0.20587635 0.54905528 0.61157149\n",
            " 0.45754921 0.45209652 0.70514619 0.6331206  0.62151444 0.44961405\n",
            " 0.47059271 0.44399151 0.46213728 0.61906385 0.45685241 0.46269304\n",
            " 0.54212123 0.31711131 0.40302616 0.57442367 0.42239451 0.58052319\n",
            " 0.74619526 0.52426636 0.8257587  0.60703117 0.5910027  0.56473929\n",
            " 0.43477574 0.58034521 0.58200407 0.63202679 0.35730889 0.64030492\n",
            " 0.58820212 0.64513063 0.61293721 0.24278252 0.63434881 0.5377295\n",
            " 0.51988918 0.70158482 0.46367633 0.60589415 0.40941137 0.61860764\n",
            " 0.69706237 0.55007315 0.54212123 0.46831644 0.60612738 0.60496294\n",
            " 0.71522516 0.47747508 0.61075062 0.23725258 0.4733265  0.7903623\n",
            " 0.51284724 0.59194642 0.52288866 0.50771666 0.50707799 0.62415993\n",
            " 0.26954767 0.57255888 0.49191594 0.63265699 0.52617222 0.54296434\n",
            " 0.54914129 0.57672268 0.54136252 0.60713458 0.31448957 0.52724797\n",
            " 0.5860213  0.37020719 0.43507901 0.3748205  0.53366339 0.46236417\n",
            " 0.51411515 0.41216385 0.32756609 0.53225464 0.30567387 0.32737568\n",
            " 0.40457022 0.36400723 0.63207936 0.34279591 0.28612572 0.43566167\n",
            " 0.53090119 0.39070228 0.5523873  0.55482137 0.60428119 0.7020818\n",
            " 0.48874682 0.56420004 0.58790314 0.31065685 0.49968627 0.6000284\n",
            " 0.65506339 0.62412435 0.43385202 0.47673672 0.56255275 0.42034459\n",
            " 0.45754921 0.57133758 0.58696032 0.33738762 0.47418743 0.65335763\n",
            " 0.6640417  0.24509022 0.64645392 0.60690308 0.34587181 0.55262744\n",
            " 0.58671391 0.5121237  0.50382167 0.46442062 0.40668207 0.60304952\n",
            " 0.61382926 0.55446786 0.48031792 0.59170455 0.5657028  0.57674879\n",
            " 0.67510438 0.28326011 0.56658769 0.72263491 0.33166465 0.45295438\n",
            " 0.49186778 0.52054298 0.38708556 0.33848614 0.50315094 0.63051873\n",
            " 0.61799896 0.63750023 0.549968   0.30086905 0.61696064 0.60852355\n",
            " 0.44458047 0.7541092  0.28133529 0.53729856 0.40229195 0.77674341\n",
            " 0.50831234 0.17126779 0.66197491 0.74910069 0.44391793 0.37521109\n",
            " 0.48939526 0.34703001 0.40833777 0.65295798 0.4791781  0.45489848\n",
            " 0.58925176 0.47193176 0.78375554 0.46136689 0.5356853  0.58105123\n",
            " 0.49530569 0.49715063 0.68066132 0.66737926 0.31233513 0.32936025\n",
            " 0.51988918 0.57877553 0.59737194 0.56441456 0.773453   0.43180269\n",
            " 0.45735243 0.40011209 0.62914771 0.61825573 0.60713458 0.46467328\n",
            " 0.54051805 0.41381255 0.34609818 0.53391534 0.35352722 0.27819821\n",
            " 0.53346992 0.50301647 0.39843854 0.54615366 0.60562396 0.38752091\n",
            " 0.65435123 0.61744159 0.64231199 0.44887596 0.73832536 0.44895184\n",
            " 0.38261622 0.46275279 0.34426367 0.61812001 0.36492366 0.43792129\n",
            " 0.60887241 0.64551592 0.48533034 0.51256347 0.55559707 0.46442062\n",
            " 0.47677898 0.68355954 0.4566268  0.57250583 0.55007315 0.44765741\n",
            " 0.51800406 0.58079052 0.76695812 0.63730586 0.38602185 0.5676617\n",
            " 0.44464299 0.46675563 0.71789777 0.69727051 0.57089251 0.47746086\n",
            " 0.65742528 0.29573834 0.3465457  0.58196974 0.49511582 0.48030153\n",
            " 0.52655596 0.6117717  0.61906385 0.69888955 0.6570344  0.61753821\n",
            " 0.73173833 0.63928646 0.34568062 0.38752091 0.58529496 0.34128904\n",
            " 0.32982939 0.39240777 0.49432641 0.69187623 0.58451504 0.76149786\n",
            " 0.71897113 0.3551026  0.44152588 0.61688256 0.6576274  0.32671028\n",
            " 0.580477   0.47433466 0.43720609 0.73415744 0.73791474 0.58792198\n",
            " 0.51605368 0.53399724 0.44745314 0.38996732 0.56645042 0.42578262\n",
            " 0.52729523 0.58597648 0.39070228 0.43261695 0.35304865 0.66649163\n",
            " 0.64641464 0.5238474  0.31648529 0.61799896 0.45443982 0.6185149\n",
            " 0.49499139 0.62024975 0.39910114 0.47517264 0.45078072 0.51782656\n",
            " 0.58129168 0.59477448 0.40806431 0.40105671 0.42226315 0.53609365\n",
            " 0.53391534 0.74669164 0.47222966 0.43550634 0.66385293 0.37483907\n",
            " 0.5588578  0.5302397  0.58382785 0.62914771 0.59981942 0.62727177\n",
            " 0.54272127 0.56257105 0.65682113 0.6862846  0.56215119 0.53539538\n",
            " 0.66638219 0.59456801 0.29882464 0.72836542 0.48496625 0.38596606\n",
            " 0.58868885 0.71227944 0.40302616 0.57187748 0.4467774  0.55514991\n",
            " 0.52902174 0.40169775 0.67066681 0.51230645 0.46888459 0.591438\n",
            " 0.5057649  0.62555921 0.43957514 0.53322309 0.44399151 0.5837971\n",
            " 0.40031523 0.77307892 0.79925388 0.43721479 0.55469048 0.14982629\n",
            " 0.47788402 0.6142748  0.4467774  0.60608196 0.32893771 0.45153898\n",
            " 0.65128613 0.69666481 0.50727266 0.54315662 0.36059266 0.28133529\n",
            " 0.43598494 0.49577883 0.63220286 0.57314456 0.67640823 0.65058291\n",
            " 0.43026781 0.42903385 0.58833838 0.53097892 0.35476106 0.58542705\n",
            " 0.60891044 0.56446457 0.64603752 0.59509587 0.42226315 0.6570344\n",
            " 0.44399151 0.66388541 0.43617702 0.580477   0.5128926  0.63977528\n",
            " 0.23017299 0.58542705 0.64048946 0.47021884 0.66094387 0.525594\n",
            " 0.28159726 0.55258822 0.53932959 0.23017299 0.24278252 0.56130981\n",
            " 0.44496036 0.47036594 0.56557792 0.35807586 0.63537335 0.55540645\n",
            " 0.61508632 0.76499581 0.65281498 0.72936916 0.46831644 0.78673095\n",
            " 0.43307167 0.55681705 0.65905678 0.79228693 0.50464821 0.59392595\n",
            " 0.40536547 0.37410533 0.79925388 0.47667557 0.19723308 0.54568708\n",
            " 0.46903133 0.57672268 0.64749146 0.55422443 0.49037924 0.55439365\n",
            " 0.51202643 0.31603727 0.301864   0.58920276 0.62757468 0.54004371\n",
            " 0.58409607 0.44185543 0.60083437 0.46236417 0.59034556 0.33159187\n",
            " 0.67859483 0.71741176 0.32991922 0.47831276]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "thresholds=np.array(list(map(lambda x: x / 100, range(0, 100, 1))))\n",
        "print(thresholds)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "h8GhECpX0AI9",
        "outputId": "bcf812a0-ac8a-4ec0-abdb-0d3bb407f0de"
      },
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[0.   0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1  0.11 0.12 0.13\n",
            " 0.14 0.15 0.16 0.17 0.18 0.19 0.2  0.21 0.22 0.23 0.24 0.25 0.26 0.27\n",
            " 0.28 0.29 0.3  0.31 0.32 0.33 0.34 0.35 0.36 0.37 0.38 0.39 0.4  0.41\n",
            " 0.42 0.43 0.44 0.45 0.46 0.47 0.48 0.49 0.5  0.51 0.52 0.53 0.54 0.55\n",
            " 0.56 0.57 0.58 0.59 0.6  0.61 0.62 0.63 0.64 0.65 0.66 0.67 0.68 0.69\n",
            " 0.7  0.71 0.72 0.73 0.74 0.75 0.76 0.77 0.78 0.79 0.8  0.81 0.82 0.83\n",
            " 0.84 0.85 0.86 0.87 0.88 0.89 0.9  0.91 0.92 0.93 0.94 0.95 0.96 0.97\n",
            " 0.98 0.99]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "roc_point=[]\n",
        "for threshold in thresholds:\n",
        "  tp=0;fp=0;fn=0;tn=0;\n",
        "  for i in range(len(similarities)):\n",
        "    actual=labels[i]\n",
        "    if similarities[i]>=threshold:\n",
        "      prediction_class=1\n",
        "    else:\n",
        "      prediction_class=0\n",
        "    if actual==1 and prediction_class==1:\n",
        "      tp+=1\n",
        "    elif actual==1 and prediction_class==0:\n",
        "      fn+=1\n",
        "    elif actual==0 and prediction_class==1:\n",
        "      fp+=1\n",
        "    elif actual==0 and prediction_class==0:\n",
        "      tn+=1\n",
        "  print(tp,fp,fn,tn)\n",
        "  tpr=tp/(tp+fn)\n",
        "  fpr=fp/(tn+fp)\n",
        "  print(tpr,fpr)\n",
        "  roc_point.append([tpr,fpr])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6_-caaDS0G1l",
        "outputId": "bcc62fe5-d5a2-453e-a9b8-c940d2ba9f84"
      },
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "500 500 0 0\n",
            "1.0 1.0\n",
            "500 500 0 0\n",
            "1.0 1.0\n",
            "500 500 0 0\n",
            "1.0 1.0\n",
            "500 500 0 0\n",
            "1.0 1.0\n",
            "500 500 0 0\n",
            "1.0 1.0\n",
            "500 500 0 0\n",
            "1.0 1.0\n",
            "500 500 0 0\n",
            "1.0 1.0\n",
            "500 500 0 0\n",
            "1.0 1.0\n",
            "500 500 0 0\n",
            "1.0 1.0\n",
            "500 500 0 0\n",
            "1.0 1.0\n",
            "500 500 0 0\n",
            "1.0 1.0\n",
            "500 500 0 0\n",
            "1.0 1.0\n",
            "500 500 0 0\n",
            "1.0 1.0\n",
            "500 500 0 0\n",
            "1.0 1.0\n",
            "500 500 0 0\n",
            "1.0 1.0\n",
            "500 499 0 1\n",
            "1.0 0.998\n",
            "500 499 0 1\n",
            "1.0 0.998\n",
            "500 499 0 1\n",
            "1.0 0.998\n",
            "500 498 0 2\n",
            "1.0 0.996\n",
            "500 498 0 2\n",
            "1.0 0.996\n",
            "500 497 0 3\n",
            "1.0 0.994\n",
            "500 496 0 4\n",
            "1.0 0.992\n",
            "500 495 0 5\n",
            "1.0 0.99\n",
            "500 495 0 5\n",
            "1.0 0.99\n",
            "500 492 0 8\n",
            "1.0 0.984\n",
            "500 489 0 11\n",
            "1.0 0.978\n",
            "500 489 0 11\n",
            "1.0 0.978\n",
            "500 488 0 12\n",
            "1.0 0.976\n",
            "500 485 0 15\n",
            "1.0 0.97\n",
            "500 479 0 21\n",
            "1.0 0.958\n",
            "500 477 0 23\n",
            "1.0 0.954\n",
            "499 474 1 26\n",
            "0.998 0.948\n",
            "498 466 2 34\n",
            "0.996 0.932\n",
            "498 459 2 41\n",
            "0.996 0.918\n",
            "497 455 3 45\n",
            "0.994 0.91\n",
            "497 447 3 53\n",
            "0.994 0.894\n",
            "496 441 4 59\n",
            "0.992 0.882\n",
            "496 438 4 62\n",
            "0.992 0.876\n",
            "496 433 4 67\n",
            "0.992 0.866\n",
            "496 425 4 75\n",
            "0.992 0.85\n",
            "493 418 7 82\n",
            "0.986 0.836\n",
            "492 405 8 95\n",
            "0.984 0.81\n",
            "490 403 10 97\n",
            "0.98 0.806\n",
            "489 397 11 103\n",
            "0.978 0.794\n",
            "485 381 15 119\n",
            "0.97 0.762\n",
            "482 364 18 136\n",
            "0.964 0.728\n",
            "480 351 20 149\n",
            "0.96 0.702\n",
            "472 333 28 167\n",
            "0.944 0.666\n",
            "462 313 38 187\n",
            "0.924 0.626\n",
            "457 304 43 196\n",
            "0.914 0.608\n",
            "449 294 51 206\n",
            "0.898 0.588\n",
            "444 285 56 215\n",
            "0.888 0.57\n",
            "437 272 63 228\n",
            "0.874 0.544\n",
            "425 261 75 239\n",
            "0.85 0.522\n",
            "418 244 82 256\n",
            "0.836 0.488\n",
            "408 229 92 271\n",
            "0.816 0.458\n",
            "397 212 103 288\n",
            "0.794 0.424\n",
            "390 199 110 301\n",
            "0.78 0.398\n",
            "381 186 119 314\n",
            "0.762 0.372\n",
            "367 156 133 344\n",
            "0.734 0.312\n",
            "354 145 146 355\n",
            "0.708 0.29\n",
            "345 128 155 372\n",
            "0.69 0.256\n",
            "331 106 169 394\n",
            "0.662 0.212\n",
            "317 96 183 404\n",
            "0.634 0.192\n",
            "294 82 206 418\n",
            "0.588 0.164\n",
            "277 73 223 427\n",
            "0.554 0.146\n",
            "257 58 243 442\n",
            "0.514 0.116\n",
            "238 48 262 452\n",
            "0.476 0.096\n",
            "227 44 273 456\n",
            "0.454 0.088\n",
            "211 40 289 460\n",
            "0.422 0.08\n",
            "188 35 312 465\n",
            "0.376 0.07\n",
            "167 32 333 468\n",
            "0.334 0.064\n",
            "155 27 345 473\n",
            "0.31 0.054\n",
            "133 24 367 476\n",
            "0.266 0.048\n",
            "123 19 377 481\n",
            "0.246 0.038\n",
            "112 16 388 484\n",
            "0.224 0.032\n",
            "95 14 405 486\n",
            "0.19 0.028\n",
            "87 11 413 489\n",
            "0.174 0.022\n",
            "77 8 423 492\n",
            "0.154 0.016\n",
            "70 5 430 495\n",
            "0.14 0.01\n",
            "59 1 441 499\n",
            "0.118 0.002\n",
            "49 1 451 499\n",
            "0.098 0.002\n",
            "43 1 457 499\n",
            "0.086 0.002\n",
            "36 0 464 500\n",
            "0.072 0.0\n",
            "26 0 474 500\n",
            "0.052 0.0\n",
            "22 0 478 500\n",
            "0.044 0.0\n",
            "18 0 482 500\n",
            "0.036 0.0\n",
            "14 0 486 500\n",
            "0.028 0.0\n",
            "11 0 489 500\n",
            "0.022 0.0\n",
            "8 0 492 500\n",
            "0.016 0.0\n",
            "7 0 493 500\n",
            "0.014 0.0\n",
            "4 0 496 500\n",
            "0.008 0.0\n",
            "2 0 498 500\n",
            "0.004 0.0\n",
            "1 0 499 500\n",
            "0.002 0.0\n",
            "0 0 500 500\n",
            "0.0 0.0\n",
            "0 0 500 500\n",
            "0.0 0.0\n",
            "0 0 500 500\n",
            "0.0 0.0\n",
            "0 0 500 500\n",
            "0.0 0.0\n",
            "0 0 500 500\n",
            "0.0 0.0\n",
            "0 0 500 500\n",
            "0.0 0.0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "pivot_test=pd.DataFrame(roc_point,columns=['x','y'])\n",
        "pivot_test['threshold']=thresholds"
      ],
      "metadata": {
        "id": "sztZ0pyJ0Hz-"
      },
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.scatter(pivot_test.y,pivot_test.x)\n",
        "plt.xlabel('false positive rate')\n",
        "plt.ylabel('true positive rate')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 466
        },
        "id": "yVHQLeQi0KlP",
        "outputId": "b5d234a9-4f2e-45c9-f222-edc95fc3db2f"
      },
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0, 0.5, 'true positive rate')"
            ]
          },
          "metadata": {},
          "execution_count": 28
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "AUC_testresnet50=np.abs(np.trapz(pivot_test.x,pivot_test.y))\n",
        "print(AUC_testresnet50)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qgDCCjfp0M0n",
        "outputId": "b418ab5c-87fc-477a-d848-1314cbf0cc28"
      },
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.786918\n"
          ]
        }
      ]
    }
  ]
}