archs_seg_mn.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "name": "dbunet_BN_segmentation_median_nerve.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "WfhZ4xmMPKCe",
        "0Lwyml4qW7ts",
        "Yh36iM9VW7uC"
      ],
      "toc_visible": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "i_yB1dFDXxBU",
        "outputId": "c523e27d-a048-456a-cfd1-6cd9580f13aa"
      },
      "source": [
        "\n",
        "from google.colab import drive\n",
        "drive.mount('/content/drive')\n",
        "\n",
        "root_path = '/content/drive/MyDrive/CVDL'\n",
        "dumps_path =  f\"{root_path}/dumps/\"\n",
        "dataset  = f\"{root_path}/dataset/\"\n",
        "\n",
        "!cp \"$root_path/modules/lib.py\" .\n",
        "!cp \"$root_path/modules/plots.py\" .\n",
        "!cp \"$root_path/modules/metrics.py\" .\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WfhZ4xmMPKCe"
      },
      "source": [
        "# Preprocessing"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0Lwyml4qW7ts"
      },
      "source": [
        "## Imports"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_DKfoow7sh3u"
      },
      "source": [
        "import cv2\n",
        "import pickle\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "import cv2 as cv\n",
        "import os\n",
        "import tensorflow as tf\n",
        "import importlib\n",
        "from datetime import datetime\n",
        "\n",
        "from keras.initializers import glorot_uniform\n",
        "\n",
        "import lib\n",
        "import plots\n",
        "import metrics\n",
        "\n",
        "tf.keras.backend.clear_session()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Yh36iM9VW7uC"
      },
      "source": [
        "## Parameters"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IYOYkQy_aNSD"
      },
      "source": [
        "# Deafault Res: 606x468 (ar 101 : 78)\n",
        "\n",
        "# Tested Res's:\n",
        "# 224x160 (ar 7 : 5)\n",
        "# 384x288\n",
        "# 576x432\n",
        "# 512x384\n",
        "input_res = (384, 288)\n",
        "\n",
        "seed = 42\n",
        "\n",
        "params = {\n",
        "          'x':input_res[0],                                \n",
        "          'y':input_res[1],                                \n",
        "          'batch_size': 8,                        \n",
        "          'n_channels_mask': 1,                            \n",
        "          'n_channels': 1,                        \n",
        "          'shuffle': True,              \n",
        "          'learningRate': 1e-3,                      \n",
        "          'momentum': 0.98,                        \n",
        "          'nEpoches': 100,                        \n",
        "          }\n",
        "\n",
        "ratio_test = 0.85\n",
        "ratio_val  = 0.8\n",
        "\n",
        "lib.create_dump([params, seed], dumps_path + \"params_du\")\n",
        "\n",
        "type_ds  = ['Images','Masks']\n",
        "modality = ['train','validation','test']\n",
        "\n",
        "# get path Img and Mask\n",
        "path_img  = os.path.join(dataset, type_ds[0])\n",
        "path_msk  = os.path.join(dataset, type_ds[1])\n",
        "path_     = [path_img, path_msk]\n",
        "\n",
        "# get length imgs and proof that is equal with masks\n",
        "length_imgs  = len(os.listdir(os.path.join(dataset,type_ds[0] + \"/\"))) \n",
        "length_masks = len(os.listdir(os.path.join(dataset,type_ds[1] + \"/\")))\n",
        "params['length_data'] = length_imgs,\n",
        "\n",
        "# get name of file imgs\n",
        "imgs_list_path = os.listdir(path_img)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VKlQcWFDW7uE"
      },
      "source": [
        "## Data loading "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Vmt_YNsFPssY",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c53e824e-8759-4a9b-c0c5-0d1442bcbc59"
      },
      "source": [
        "# LOAD & SPLIT DATA ============================================================\n",
        "'''\n",
        "mode = 1  # mode = 0: Random Split\n",
        "          # mode = 1: Split for Patients\n",
        "          \n",
        "dataset_list = lib.get_dataset_list(\n",
        "                                mode, \n",
        "                                path_, \n",
        "                                imgs_list_path, \n",
        "                                params, \n",
        "                                ratio_test, \n",
        "                                ratio_val, \n",
        "                                seed, \n",
        "                                modality)\n",
        "\n",
        "# Dumping splitted data\n",
        "lib.create_dump(dataset_list, f\"{dumps_path}data_rt{ratio_test}_rv{ratio_val}_{input_res[0]}x{input_res[1]}\")\n",
        "\n",
        "'''"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'\\nmode = 1  #\\xa0mode = 0: Random Split\\n          #\\xa0mode = 1: Split for Patients\\n          \\ndataset_list = lib.get_dataset_list(\\n                                mode, \\n                                path_, \\n                                imgs_list_path, \\n                                params, \\n                                ratio_test, \\n                                ratio_val, \\n                                seed, \\n                                modality)\\n\\n# Dumping splitted data\\nlib.create_dump(dataset_list, f\"{dumps_path}data_rt{ratio_test}_rv{ratio_val}_{input_res[0]}x{input_res[1]}\")\\n\\n'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9VS9PNzw6Pp2"
      },
      "source": [
        "##Loading Data from Pickle"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OF4QPe3Rv57S"
      },
      "source": [
        "dataset_list = pickle.load(open( f\"{dumps_path}data_rt{ratio_test}_rv{ratio_val}_{input_res[0]}x{input_res[1]}\", \"rb\" ))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TDi2fSk96V14"
      },
      "source": [
        "##Data Augmentation"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_u3XQo-vW7uJ"
      },
      "source": [
        "# DATA AUGMENTATION ============================================================\n",
        "train_generator, val_generator = lib.data_augmentation(seed, dataset_list, params)\n",
        "\n",
        "params[\"length_training\"]   = len(dataset_list[0]['train'])\n",
        "params['length_validation'] = len(dataset_list[0]['validation'])\n",
        "\n",
        "augmented_batch_plot = plots.plot_augmented_batch(train_generator)\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eZMPcQNURjPd"
      },
      "source": [
        "print(params[\"length_training\"])\n",
        "print(params['length_validation'])\n",
        "print(len(dataset_list[0]['test']))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AJRSxu-3W7uP"
      },
      "source": [
        "# Model definition & training"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qaDOvDLTYOTH"
      },
      "source": [
        "## Training"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pfFpxzd269xL"
      },
      "source": [
        "params['learningRate'] = 1e-3\n",
        "params['nEpoches'] = 100\n",
        "\n",
        "\n",
        "models_name = ['unet', 'lightweight_unet', 'dbunet', 'dbunet_BNDrop', 'dbunet_BN']\n",
        "arch_name = models_name[4]\n",
        "model_weights = f\"{root_path}/dataset/results/models/{arch_name}_Glorot_e{params['nEpoches']}_lr{params['learningRate']}\"\n",
        "log_path = f\"{model_weights}/logs/fit/\" + datetime.now().strftime(\"%Y%m%d-%H%M%S\")\n",
        "!rm -rf $log_path\n",
        "!cp \"$root_path/modules/models/{arch_name}.py\" .\n",
        "arch = importlib.import_module(arch_name)\n",
        "'''if 'dbunet' in arch_name:\n",
        "  model = arch.build_model(input_size)\n",
        "else:'''\n",
        "\n",
        "\n",
        "input_size = (input_res[1], input_res[0], 1)\n",
        "initializer = tf.keras.initializers.GlorotUniform()\n",
        "model = arch.build_model(initializer, input_size)\n",
        "#model.summary()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WfKA5g_BtmcE",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "733e6128-5a38-4d67-ec32-3543f77e3529"
      },
      "source": [
        "# Model savepath\n",
        "#now = datetime.now()\n",
        "#current_time = now.strftime(\"%H:%M:%S\")\n",
        "\n",
        "# Default loss, metric\n",
        "#loss = 'binary_crossentropy'\n",
        "#metric = 'accuracy'\n",
        "loss      = metrics.bce_dice_loss\n",
        "recall    = tf.keras.metrics.Recall()\n",
        "precision = tf.keras.metrics.Precision()\n",
        "metric    = [metrics.dice, metrics.iou, recall, precision]\n",
        "\n",
        "print(f'Model wil be saved in {model_weights}')\n",
        "\n",
        "train_history, model = lib.train(model_weights, log_path, params, model, loss, metric, train_generator, val_generator)\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model wil be saved in /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/utils/generic_utils.py:497: CustomMaskWarning: Custom mask layers require a config and must override get_config. When loading, the custom mask layer must be passed to the custom_objects argument.\n",
            "  category=CustomMaskWarning)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/100\n",
            "21/21 [==============================] - 36s 1s/step - loss: 1.4699 - dice: 0.0233 - iou: 0.0106 - recall: 0.0479 - precision: 0.0311 - val_loss: 1.6561 - val_dice: 0.0222 - val_iou: 8.0712e-09 - val_recall: 0.0000e+00 - val_precision: 0.0000e+00\n",
            "\n",
            "Epoch 00001: val_loss improved from inf to 1.65615, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 2/100\n",
            "21/21 [==============================] - 12s 589ms/step - loss: 1.2469 - dice: 0.0342 - iou: 5.2360e-06 - recall: 4.9391e-06 - precision: 0.0090 - val_loss: 1.4860 - val_dice: 0.0203 - val_iou: 9.3318e-09 - val_recall: 0.0000e+00 - val_precision: 0.0000e+00\n",
            "\n",
            "Epoch 00002: val_loss improved from 1.65615 to 1.48605, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 3/100\n",
            "21/21 [==============================] - 12s 559ms/step - loss: 1.0461 - dice: 0.0760 - iou: 1.0244e-08 - recall: 0.0000e+00 - precision: 0.0000e+00 - val_loss: 1.2212 - val_dice: 0.0405 - val_iou: 9.4372e-09 - val_recall: 0.0000e+00 - val_precision: 0.0000e+00\n",
            "\n",
            "Epoch 00003: val_loss improved from 1.48605 to 1.22118, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 4/100\n",
            "21/21 [==============================] - 12s 561ms/step - loss: 0.8627 - dice: 0.1847 - iou: 1.0299e-08 - recall: 0.0000e+00 - precision: 0.0000e+00 - val_loss: 1.1157 - val_dice: 0.0303 - val_iou: 8.6567e-09 - val_recall: 0.0000e+00 - val_precision: 0.0000e+00\n",
            "\n",
            "Epoch 00004: val_loss improved from 1.22118 to 1.11567, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 5/100\n",
            "21/21 [==============================] - 12s 560ms/step - loss: 0.6331 - dice: 0.3898 - iou: 0.3997 - recall: 0.4283 - precision: 0.7526 - val_loss: 1.0577 - val_dice: 0.0161 - val_iou: 0.0024 - val_recall: 0.0025 - val_precision: 0.2719\n",
            "\n",
            "Epoch 00005: val_loss improved from 1.11567 to 1.05768, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 6/100\n",
            "21/21 [==============================] - 12s 591ms/step - loss: 0.4083 - dice: 0.6108 - iou: 0.6297 - recall: 0.6952 - precision: 0.7832 - val_loss: 1.0342 - val_dice: 0.0621 - val_iou: 0.0576 - val_recall: 0.0382 - val_precision: 0.6557\n",
            "\n",
            "Epoch 00006: val_loss improved from 1.05768 to 1.03424, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 7/100\n",
            "21/21 [==============================] - 12s 588ms/step - loss: 0.2538 - dice: 0.7609 - iou: 0.7004 - recall: 0.7333 - precision: 0.8561 - val_loss: 1.0545 - val_dice: 0.0292 - val_iou: 0.0130 - val_recall: 0.0094 - val_precision: 0.2117\n",
            "\n",
            "Epoch 00007: val_loss did not improve from 1.03424\n",
            "Epoch 8/100\n",
            "21/21 [==============================] - 12s 585ms/step - loss: 0.2277 - dice: 0.7867 - iou: 0.7011 - recall: 0.7367 - precision: 0.8655 - val_loss: 0.9778 - val_dice: 0.0986 - val_iou: 0.0522 - val_recall: 0.0492 - val_precision: 0.5874\n",
            "\n",
            "Epoch 00008: val_loss improved from 1.03424 to 0.97781, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 9/100\n",
            "21/21 [==============================] - 12s 587ms/step - loss: 0.2052 - dice: 0.8113 - iou: 0.7273 - recall: 0.7484 - precision: 0.8792 - val_loss: 1.1041 - val_dice: 3.5860e-04 - val_iou: 9.5102e-09 - val_recall: 0.0000e+00 - val_precision: 0.0000e+00\n",
            "\n",
            "Epoch 00009: val_loss did not improve from 0.97781\n",
            "Epoch 10/100\n",
            "21/21 [==============================] - 12s 589ms/step - loss: 0.2044 - dice: 0.8085 - iou: 0.7159 - recall: 0.7485 - precision: 0.8840 - val_loss: 0.6491 - val_dice: 0.4507 - val_iou: 0.2943 - val_recall: 0.6931 - val_precision: 0.3518\n",
            "\n",
            "Epoch 00010: val_loss improved from 0.97781 to 0.64915, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 11/100\n",
            "21/21 [==============================] - 12s 565ms/step - loss: 0.1698 - dice: 0.8417 - iou: 0.7579 - recall: 0.7704 - precision: 0.9130 - val_loss: 0.4322 - val_dice: 0.6164 - val_iou: 0.4812 - val_recall: 0.6347 - val_precision: 0.6159\n",
            "\n",
            "Epoch 00011: val_loss improved from 0.64915 to 0.43223, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 12/100\n",
            "21/21 [==============================] - 12s 565ms/step - loss: 0.1746 - dice: 0.8357 - iou: 0.7536 - recall: 0.7779 - precision: 0.8961 - val_loss: 0.2081 - val_dice: 0.8052 - val_iou: 0.7604 - val_recall: 0.7598 - val_precision: 0.9063\n",
            "\n",
            "Epoch 00012: val_loss improved from 0.43223 to 0.20810, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 13/100\n",
            "21/21 [==============================] - 12s 564ms/step - loss: 0.1734 - dice: 0.8383 - iou: 0.7530 - recall: 0.7811 - precision: 0.8864 - val_loss: 0.2125 - val_dice: 0.8099 - val_iou: 0.7492 - val_recall: 0.7452 - val_precision: 0.9115\n",
            "\n",
            "Epoch 00013: val_loss did not improve from 0.20810\n",
            "Epoch 14/100\n",
            "21/21 [==============================] - 12s 562ms/step - loss: 0.1818 - dice: 0.8319 - iou: 0.7462 - recall: 0.7710 - precision: 0.8967 - val_loss: 0.2882 - val_dice: 0.7361 - val_iou: 0.6489 - val_recall: 0.6704 - val_precision: 0.8557\n",
            "\n",
            "Epoch 00014: val_loss did not improve from 0.20810\n",
            "Epoch 15/100\n",
            "21/21 [==============================] - 12s 560ms/step - loss: 0.1607 - dice: 0.8522 - iou: 0.7684 - recall: 0.7843 - precision: 0.9064 - val_loss: 0.2420 - val_dice: 0.7850 - val_iou: 0.7163 - val_recall: 0.6997 - val_precision: 0.9287\n",
            "\n",
            "Epoch 00015: val_loss did not improve from 0.20810\n",
            "Epoch 16/100\n",
            "21/21 [==============================] - 12s 561ms/step - loss: 0.1469 - dice: 0.8650 - iou: 0.7841 - recall: 0.8110 - precision: 0.9154 - val_loss: 0.5450 - val_dice: 0.5106 - val_iou: 0.4081 - val_recall: 0.3699 - val_precision: 0.9565\n",
            "\n",
            "Epoch 00016: val_loss did not improve from 0.20810\n",
            "Epoch 17/100\n",
            "21/21 [==============================] - 12s 557ms/step - loss: 0.1510 - dice: 0.8617 - iou: 0.7837 - recall: 0.7842 - precision: 0.9240 - val_loss: 0.2352 - val_dice: 0.8006 - val_iou: 0.7288 - val_recall: 0.6941 - val_precision: 0.9359\n",
            "\n",
            "Epoch 00017: val_loss did not improve from 0.20810\n",
            "Epoch 18/100\n",
            "21/21 [==============================] - 12s 560ms/step - loss: 0.1451 - dice: 0.8645 - iou: 0.7828 - recall: 0.8062 - precision: 0.9221 - val_loss: 0.2033 - val_dice: 0.8195 - val_iou: 0.7535 - val_recall: 0.7501 - val_precision: 0.9040\n",
            "\n",
            "Epoch 00018: val_loss improved from 0.20810 to 0.20330, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 19/100\n",
            "21/21 [==============================] - 12s 590ms/step - loss: 0.1207 - dice: 0.8886 - iou: 0.8215 - recall: 0.8165 - precision: 0.9472 - val_loss: 0.1884 - val_dice: 0.8327 - val_iou: 0.7502 - val_recall: 0.7306 - val_precision: 0.9459\n",
            "\n",
            "Epoch 00019: val_loss improved from 0.20330 to 0.18836, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 20/100\n",
            "21/21 [==============================] - 12s 564ms/step - loss: 0.1262 - dice: 0.8841 - iou: 0.8114 - recall: 0.8048 - precision: 0.9467 - val_loss: 0.2009 - val_dice: 0.8262 - val_iou: 0.7560 - val_recall: 0.7322 - val_precision: 0.9324\n",
            "\n",
            "Epoch 00020: val_loss did not improve from 0.18836\n",
            "Epoch 21/100\n",
            "21/21 [==============================] - 12s 559ms/step - loss: 0.1342 - dice: 0.8770 - iou: 0.8019 - recall: 0.8172 - precision: 0.9247 - val_loss: 0.1630 - val_dice: 0.8503 - val_iou: 0.7638 - val_recall: 0.8330 - val_precision: 0.8815\n",
            "\n",
            "Epoch 00021: val_loss improved from 0.18836 to 0.16298, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 22/100\n",
            "21/21 [==============================] - 12s 559ms/step - loss: 0.1145 - dice: 0.8938 - iou: 0.8225 - recall: 0.8251 - precision: 0.9483 - val_loss: 0.1552 - val_dice: 0.8640 - val_iou: 0.7961 - val_recall: 0.7957 - val_precision: 0.9347\n",
            "\n",
            "Epoch 00022: val_loss improved from 0.16298 to 0.15524, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 23/100\n",
            "21/21 [==============================] - 12s 589ms/step - loss: 0.1222 - dice: 0.8879 - iou: 0.8212 - recall: 0.8198 - precision: 0.9406 - val_loss: 0.1491 - val_dice: 0.8656 - val_iou: 0.7764 - val_recall: 0.9010 - val_precision: 0.8419\n",
            "\n",
            "Epoch 00023: val_loss improved from 0.15524 to 0.14911, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 24/100\n",
            "21/21 [==============================] - 12s 564ms/step - loss: 0.1199 - dice: 0.8904 - iou: 0.8232 - recall: 0.8253 - precision: 0.9397 - val_loss: 0.2442 - val_dice: 0.7877 - val_iou: 0.7236 - val_recall: 0.6985 - val_precision: 0.9055\n",
            "\n",
            "Epoch 00024: val_loss did not improve from 0.14911\n",
            "Epoch 25/100\n",
            "21/21 [==============================] - 12s 558ms/step - loss: 0.1253 - dice: 0.8848 - iou: 0.8133 - recall: 0.8172 - precision: 0.9340 - val_loss: 0.1509 - val_dice: 0.8666 - val_iou: 0.7922 - val_recall: 0.8622 - val_precision: 0.8939\n",
            "\n",
            "Epoch 00025: val_loss did not improve from 0.14911\n",
            "Epoch 26/100\n",
            "21/21 [==============================] - 12s 558ms/step - loss: 0.1076 - dice: 0.9006 - iou: 0.8376 - recall: 0.8329 - precision: 0.9525 - val_loss: 0.2517 - val_dice: 0.7843 - val_iou: 0.7070 - val_recall: 0.6429 - val_precision: 0.9752\n",
            "\n",
            "Epoch 00026: val_loss did not improve from 0.14911\n",
            "Epoch 27/100\n",
            "21/21 [==============================] - 12s 587ms/step - loss: 0.1146 - dice: 0.8949 - iou: 0.8284 - recall: 0.8222 - precision: 0.9526 - val_loss: 0.1234 - val_dice: 0.8823 - val_iou: 0.8156 - val_recall: 0.8244 - val_precision: 0.9575\n",
            "\n",
            "Epoch 00027: val_loss improved from 0.14911 to 0.12336, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 28/100\n",
            "21/21 [==============================] - 12s 563ms/step - loss: 0.1085 - dice: 0.8985 - iou: 0.8328 - recall: 0.8307 - precision: 0.9511 - val_loss: 0.1832 - val_dice: 0.8417 - val_iou: 0.7876 - val_recall: 0.7380 - val_precision: 0.9543\n",
            "\n",
            "Epoch 00028: val_loss did not improve from 0.12336\n",
            "Epoch 29/100\n",
            "21/21 [==============================] - 12s 587ms/step - loss: 0.1218 - dice: 0.8873 - iou: 0.8189 - recall: 0.8223 - precision: 0.9393 - val_loss: 0.1334 - val_dice: 0.8838 - val_iou: 0.8149 - val_recall: 0.8310 - val_precision: 0.9368\n",
            "\n",
            "Epoch 00029: val_loss did not improve from 0.12336\n",
            "Epoch 30/100\n",
            "21/21 [==============================] - 12s 588ms/step - loss: 0.1031 - dice: 0.9044 - iou: 0.8464 - recall: 0.8340 - precision: 0.9583 - val_loss: 0.1213 - val_dice: 0.8912 - val_iou: 0.8299 - val_recall: 0.8696 - val_precision: 0.9146\n",
            "\n",
            "Epoch 00030: val_loss improved from 0.12336 to 0.12132, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 31/100\n",
            "21/21 [==============================] - 12s 564ms/step - loss: 0.1142 - dice: 0.8953 - iou: 0.8270 - recall: 0.8306 - precision: 0.9417 - val_loss: 0.1750 - val_dice: 0.8546 - val_iou: 0.7962 - val_recall: 0.7805 - val_precision: 0.9220\n",
            "\n",
            "Epoch 00031: val_loss did not improve from 0.12132\n",
            "Epoch 32/100\n",
            "21/21 [==============================] - 12s 558ms/step - loss: 0.1176 - dice: 0.8927 - iou: 0.8294 - recall: 0.8226 - precision: 0.9438 - val_loss: 0.5992 - val_dice: 0.4936 - val_iou: 0.3947 - val_recall: 0.3303 - val_precision: 0.9639\n",
            "\n",
            "Epoch 00032: val_loss did not improve from 0.12132\n",
            "Epoch 33/100\n",
            "21/21 [==============================] - 12s 557ms/step - loss: 0.1253 - dice: 0.8852 - iou: 0.8157 - recall: 0.8191 - precision: 0.9393 - val_loss: 0.1523 - val_dice: 0.8651 - val_iou: 0.7836 - val_recall: 0.7984 - val_precision: 0.9335\n",
            "\n",
            "Epoch 00033: val_loss did not improve from 0.12132\n",
            "Epoch 34/100\n",
            "21/21 [==============================] - 12s 561ms/step - loss: 0.1058 - dice: 0.8992 - iou: 0.8348 - recall: 0.8266 - precision: 0.9564 - val_loss: 0.2090 - val_dice: 0.8216 - val_iou: 0.7746 - val_recall: 0.7534 - val_precision: 0.8962\n",
            "\n",
            "Epoch 00034: val_loss did not improve from 0.12132\n",
            "Epoch 35/100\n",
            "21/21 [==============================] - 12s 588ms/step - loss: 0.1205 - dice: 0.8875 - iou: 0.8179 - recall: 0.8170 - precision: 0.9468 - val_loss: 0.1739 - val_dice: 0.8503 - val_iou: 0.7995 - val_recall: 0.7925 - val_precision: 0.9194\n",
            "\n",
            "Epoch 00035: val_loss did not improve from 0.12132\n",
            "Epoch 36/100\n",
            "21/21 [==============================] - 12s 559ms/step - loss: 0.1193 - dice: 0.8909 - iou: 0.8214 - recall: 0.8186 - precision: 0.9400 - val_loss: 0.1531 - val_dice: 0.8644 - val_iou: 0.7883 - val_recall: 0.8761 - val_precision: 0.8631\n",
            "\n",
            "Epoch 00036: val_loss did not improve from 0.12132\n",
            "Epoch 37/100\n",
            "21/21 [==============================] - 12s 562ms/step - loss: 0.1162 - dice: 0.8905 - iou: 0.8159 - recall: 0.8217 - precision: 0.9502 - val_loss: 0.1595 - val_dice: 0.8502 - val_iou: 0.7917 - val_recall: 0.7954 - val_precision: 0.9350\n",
            "\n",
            "Epoch 00037: val_loss did not improve from 0.12132\n",
            "Epoch 38/100\n",
            "21/21 [==============================] - 12s 588ms/step - loss: 0.1068 - dice: 0.9002 - iou: 0.8364 - recall: 0.8338 - precision: 0.9533 - val_loss: 0.1436 - val_dice: 0.8637 - val_iou: 0.7949 - val_recall: 0.8537 - val_precision: 0.8906\n",
            "\n",
            "Epoch 00038: val_loss did not improve from 0.12132\n",
            "Epoch 39/100\n",
            "21/21 [==============================] - 12s 585ms/step - loss: 0.0989 - dice: 0.9093 - iou: 0.8481 - recall: 0.8434 - precision: 0.9585 - val_loss: 0.1493 - val_dice: 0.8663 - val_iou: 0.8177 - val_recall: 0.8088 - val_precision: 0.9322\n",
            "\n",
            "Epoch 00039: val_loss did not improve from 0.12132\n",
            "Epoch 40/100\n",
            "21/21 [==============================] - 12s 560ms/step - loss: 0.1015 - dice: 0.9054 - iou: 0.8401 - recall: 0.8355 - precision: 0.9590 - val_loss: 0.1291 - val_dice: 0.8859 - val_iou: 0.8193 - val_recall: 0.8542 - val_precision: 0.9072\n",
            "\n",
            "Epoch 00040: val_loss did not improve from 0.12132\n",
            "\n",
            "Epoch 00040: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
            "Epoch 41/100\n",
            "21/21 [==============================] - 12s 560ms/step - loss: 0.1020 - dice: 0.9066 - iou: 0.8481 - recall: 0.8247 - precision: 0.9626 - val_loss: 0.1374 - val_dice: 0.8811 - val_iou: 0.8230 - val_recall: 0.8247 - val_precision: 0.9267\n",
            "\n",
            "Epoch 00041: val_loss did not improve from 0.12132\n",
            "Epoch 42/100\n",
            "21/21 [==============================] - 12s 562ms/step - loss: 0.0971 - dice: 0.9103 - iou: 0.8523 - recall: 0.8433 - precision: 0.9599 - val_loss: 0.1814 - val_dice: 0.8362 - val_iou: 0.7988 - val_recall: 0.7603 - val_precision: 0.9301\n",
            "\n",
            "Epoch 00042: val_loss did not improve from 0.12132\n",
            "Epoch 43/100\n",
            "21/21 [==============================] - 12s 587ms/step - loss: 0.0849 - dice: 0.9210 - iou: 0.8688 - recall: 0.8467 - precision: 0.9736 - val_loss: 0.1180 - val_dice: 0.8960 - val_iou: 0.8307 - val_recall: 0.8459 - val_precision: 0.9470\n",
            "\n",
            "Epoch 00043: val_loss improved from 0.12132 to 0.11798, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 44/100\n",
            "21/21 [==============================] - 12s 590ms/step - loss: 0.0904 - dice: 0.9166 - iou: 0.8616 - recall: 0.8455 - precision: 0.9669 - val_loss: 0.1494 - val_dice: 0.8621 - val_iou: 0.8078 - val_recall: 0.8130 - val_precision: 0.9188\n",
            "\n",
            "Epoch 00044: val_loss did not improve from 0.11798\n",
            "Epoch 45/100\n",
            "21/21 [==============================] - 12s 586ms/step - loss: 0.0910 - dice: 0.9165 - iou: 0.8609 - recall: 0.8470 - precision: 0.9637 - val_loss: 0.1409 - val_dice: 0.8750 - val_iou: 0.8220 - val_recall: 0.8074 - val_precision: 0.9421\n",
            "\n",
            "Epoch 00045: val_loss did not improve from 0.11798\n",
            "Epoch 46/100\n",
            "21/21 [==============================] - 12s 588ms/step - loss: 0.0889 - dice: 0.9181 - iou: 0.8646 - recall: 0.8459 - precision: 0.9673 - val_loss: 0.1353 - val_dice: 0.8819 - val_iou: 0.8270 - val_recall: 0.8149 - val_precision: 0.9501\n",
            "\n",
            "Epoch 00046: val_loss did not improve from 0.11798\n",
            "Epoch 47/100\n",
            "21/21 [==============================] - 12s 586ms/step - loss: 0.0893 - dice: 0.9183 - iou: 0.8643 - recall: 0.8430 - precision: 0.9692 - val_loss: 0.1096 - val_dice: 0.9023 - val_iou: 0.8468 - val_recall: 0.8499 - val_precision: 0.9565\n",
            "\n",
            "Epoch 00047: val_loss improved from 0.11798 to 0.10962, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 48/100\n",
            "21/21 [==============================] - 12s 566ms/step - loss: 0.0911 - dice: 0.9165 - iou: 0.8580 - recall: 0.8463 - precision: 0.9650 - val_loss: 0.1835 - val_dice: 0.8395 - val_iou: 0.7873 - val_recall: 0.7417 - val_precision: 0.9487\n",
            "\n",
            "Epoch 00048: val_loss did not improve from 0.10962\n",
            "Epoch 49/100\n",
            "21/21 [==============================] - 12s 559ms/step - loss: 0.0867 - dice: 0.9200 - iou: 0.8677 - recall: 0.8489 - precision: 0.9704 - val_loss: 0.1555 - val_dice: 0.8658 - val_iou: 0.8130 - val_recall: 0.7817 - val_precision: 0.9593\n",
            "\n",
            "Epoch 00049: val_loss did not improve from 0.10962\n",
            "Epoch 50/100\n",
            "21/21 [==============================] - 12s 588ms/step - loss: 0.0846 - dice: 0.9217 - iou: 0.8693 - recall: 0.8526 - precision: 0.9716 - val_loss: 0.1556 - val_dice: 0.8633 - val_iou: 0.8047 - val_recall: 0.7872 - val_precision: 0.9464\n",
            "\n",
            "Epoch 00050: val_loss did not improve from 0.10962\n",
            "Epoch 51/100\n",
            "21/21 [==============================] - 12s 561ms/step - loss: 0.0840 - dice: 0.9228 - iou: 0.8723 - recall: 0.8515 - precision: 0.9726 - val_loss: 0.1458 - val_dice: 0.8733 - val_iou: 0.8190 - val_recall: 0.7943 - val_precision: 0.9526\n",
            "\n",
            "Epoch 00051: val_loss did not improve from 0.10962\n",
            "Epoch 52/100\n",
            "21/21 [==============================] - 12s 587ms/step - loss: 0.0874 - dice: 0.9195 - iou: 0.8626 - recall: 0.8419 - precision: 0.9723 - val_loss: 0.1265 - val_dice: 0.8882 - val_iou: 0.8329 - val_recall: 0.8353 - val_precision: 0.9362\n",
            "\n",
            "Epoch 00052: val_loss did not improve from 0.10962\n",
            "Epoch 53/100\n",
            "21/21 [==============================] - 12s 587ms/step - loss: 0.0840 - dice: 0.9221 - iou: 0.8711 - recall: 0.8555 - precision: 0.9692 - val_loss: 0.1288 - val_dice: 0.8852 - val_iou: 0.8272 - val_recall: 0.8315 - val_precision: 0.9391\n",
            "\n",
            "Epoch 00053: val_loss did not improve from 0.10962\n",
            "Epoch 54/100\n",
            "21/21 [==============================] - 12s 560ms/step - loss: 0.0855 - dice: 0.9211 - iou: 0.8696 - recall: 0.8427 - precision: 0.9741 - val_loss: 0.1226 - val_dice: 0.8922 - val_iou: 0.8390 - val_recall: 0.8294 - val_precision: 0.9622\n",
            "\n",
            "Epoch 00054: val_loss did not improve from 0.10962\n",
            "Epoch 55/100\n",
            "21/21 [==============================] - 12s 586ms/step - loss: 0.0844 - dice: 0.9227 - iou: 0.8721 - recall: 0.8455 - precision: 0.9741 - val_loss: 0.1433 - val_dice: 0.8719 - val_iou: 0.8182 - val_recall: 0.8069 - val_precision: 0.9445\n",
            "\n",
            "Epoch 00055: val_loss did not improve from 0.10962\n",
            "Epoch 56/100\n",
            "21/21 [==============================] - 12s 561ms/step - loss: 0.0833 - dice: 0.9230 - iou: 0.8716 - recall: 0.8507 - precision: 0.9734 - val_loss: 0.1389 - val_dice: 0.8798 - val_iou: 0.8261 - val_recall: 0.8060 - val_precision: 0.9534\n",
            "\n",
            "Epoch 00056: val_loss did not improve from 0.10962\n",
            "Epoch 57/100\n",
            "21/21 [==============================] - 12s 585ms/step - loss: 0.0847 - dice: 0.9217 - iou: 0.8688 - recall: 0.8495 - precision: 0.9721 - val_loss: 0.1547 - val_dice: 0.8596 - val_iou: 0.8116 - val_recall: 0.7963 - val_precision: 0.9480\n",
            "\n",
            "Epoch 00057: val_loss did not improve from 0.10962\n",
            "\n",
            "Epoch 00057: ReduceLROnPlateau reducing learning rate to 1.0000000474974514e-05.\n",
            "Epoch 58/100\n",
            "21/21 [==============================] - 12s 562ms/step - loss: 0.0824 - dice: 0.9242 - iou: 0.8738 - recall: 0.8539 - precision: 0.9729 - val_loss: 0.1223 - val_dice: 0.8925 - val_iou: 0.8371 - val_recall: 0.8283 - val_precision: 0.9516\n",
            "\n",
            "Epoch 00058: val_loss did not improve from 0.10962\n",
            "Epoch 59/100\n",
            "21/21 [==============================] - 12s 586ms/step - loss: 0.0831 - dice: 0.9227 - iou: 0.8718 - recall: 0.8520 - precision: 0.9710 - val_loss: 0.1363 - val_dice: 0.8812 - val_iou: 0.8228 - val_recall: 0.8126 - val_precision: 0.9500\n",
            "\n",
            "Epoch 00059: val_loss did not improve from 0.10962\n",
            "Epoch 60/100\n",
            "21/21 [==============================] - 12s 588ms/step - loss: 0.0844 - dice: 0.9216 - iou: 0.8721 - recall: 0.8520 - precision: 0.9708 - val_loss: 0.1336 - val_dice: 0.8837 - val_iou: 0.8241 - val_recall: 0.8150 - val_precision: 0.9454\n",
            "\n",
            "Epoch 00060: val_loss did not improve from 0.10962\n",
            "Epoch 61/100\n",
            "21/21 [==============================] - 12s 561ms/step - loss: 0.0844 - dice: 0.9223 - iou: 0.8682 - recall: 0.8438 - precision: 0.9747 - val_loss: 0.1379 - val_dice: 0.8737 - val_iou: 0.8223 - val_recall: 0.8136 - val_precision: 0.9482\n",
            "\n",
            "Epoch 00061: val_loss did not improve from 0.10962\n",
            "Epoch 62/100\n",
            "21/21 [==============================] - 12s 586ms/step - loss: 0.0833 - dice: 0.9225 - iou: 0.8717 - recall: 0.8545 - precision: 0.9714 - val_loss: 0.1335 - val_dice: 0.8764 - val_iou: 0.8193 - val_recall: 0.8100 - val_precision: 0.9435\n",
            "\n",
            "Epoch 00062: val_loss did not improve from 0.10962\n",
            "Epoch 63/100\n",
            "21/21 [==============================] - 12s 585ms/step - loss: 0.0845 - dice: 0.9218 - iou: 0.8675 - recall: 0.8497 - precision: 0.9712 - val_loss: 0.1533 - val_dice: 0.8665 - val_iou: 0.8120 - val_recall: 0.7945 - val_precision: 0.9369\n",
            "\n",
            "Epoch 00063: val_loss did not improve from 0.10962\n",
            "Epoch 64/100\n",
            "21/21 [==============================] - 12s 587ms/step - loss: 0.0853 - dice: 0.9208 - iou: 0.8679 - recall: 0.8491 - precision: 0.9716 - val_loss: 0.1464 - val_dice: 0.8723 - val_iou: 0.8178 - val_recall: 0.8061 - val_precision: 0.9453\n",
            "\n",
            "Epoch 00064: val_loss did not improve from 0.10962\n",
            "Epoch 65/100\n",
            "21/21 [==============================] - 12s 587ms/step - loss: 0.0824 - dice: 0.9240 - iou: 0.8725 - recall: 0.8503 - precision: 0.9734 - val_loss: 0.1321 - val_dice: 0.8865 - val_iou: 0.8278 - val_recall: 0.8070 - val_precision: 0.9538\n",
            "\n",
            "Epoch 00065: val_loss did not improve from 0.10962\n",
            "Epoch 66/100\n",
            "21/21 [==============================] - 12s 563ms/step - loss: 0.0839 - dice: 0.9221 - iou: 0.8684 - recall: 0.8481 - precision: 0.9726 - val_loss: 0.1363 - val_dice: 0.8754 - val_iou: 0.8230 - val_recall: 0.8133 - val_precision: 0.9492\n",
            "\n",
            "Epoch 00066: val_loss did not improve from 0.10962\n",
            "Epoch 67/100\n",
            "21/21 [==============================] - 12s 592ms/step - loss: 0.0843 - dice: 0.9212 - iou: 0.8700 - recall: 0.8530 - precision: 0.9699 - val_loss: 0.1753 - val_dice: 0.8519 - val_iou: 0.7966 - val_recall: 0.7615 - val_precision: 0.9411\n",
            "\n",
            "Epoch 00067: val_loss did not improve from 0.10962\n",
            "\n",
            "Epoch 00067: ReduceLROnPlateau reducing learning rate to 1.0000000656873453e-06.\n",
            "Epoch 68/100\n",
            "21/21 [==============================] - 12s 587ms/step - loss: 0.0822 - dice: 0.9235 - iou: 0.8720 - recall: 0.8484 - precision: 0.9755 - val_loss: 0.1041 - val_dice: 0.9088 - val_iou: 0.8524 - val_recall: 0.8547 - val_precision: 0.9582\n",
            "\n",
            "Epoch 00068: val_loss improved from 0.10962 to 0.10411, saving model to /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n",
            "Epoch 69/100\n",
            "21/21 [==============================] - 12s 561ms/step - loss: 0.0822 - dice: 0.9244 - iou: 0.8731 - recall: 0.8503 - precision: 0.9742 - val_loss: 0.1406 - val_dice: 0.8755 - val_iou: 0.8213 - val_recall: 0.8169 - val_precision: 0.9404\n",
            "\n",
            "Epoch 00069: val_loss did not improve from 0.10411\n",
            "Epoch 70/100\n",
            "21/21 [==============================] - 12s 563ms/step - loss: 0.0842 - dice: 0.9217 - iou: 0.8700 - recall: 0.8490 - precision: 0.9724 - val_loss: 0.1407 - val_dice: 0.8777 - val_iou: 0.8253 - val_recall: 0.8097 - val_precision: 0.9490\n",
            "\n",
            "Epoch 00070: val_loss did not improve from 0.10411\n",
            "Epoch 71/100\n",
            "21/21 [==============================] - 12s 590ms/step - loss: 0.0844 - dice: 0.9222 - iou: 0.8709 - recall: 0.8512 - precision: 0.9712 - val_loss: 0.1374 - val_dice: 0.8796 - val_iou: 0.8273 - val_recall: 0.8150 - val_precision: 0.9482\n",
            "\n",
            "Epoch 00071: val_loss did not improve from 0.10411\n",
            "Epoch 72/100\n",
            "21/21 [==============================] - 12s 589ms/step - loss: 0.0860 - dice: 0.9207 - iou: 0.8667 - recall: 0.8456 - precision: 0.9707 - val_loss: 0.1424 - val_dice: 0.8763 - val_iou: 0.8292 - val_recall: 0.8102 - val_precision: 0.9479\n",
            "\n",
            "Epoch 00072: val_loss did not improve from 0.10411\n",
            "Epoch 73/100\n",
            "21/21 [==============================] - 12s 587ms/step - loss: 0.0837 - dice: 0.9224 - iou: 0.8705 - recall: 0.8499 - precision: 0.9730 - val_loss: 0.1152 - val_dice: 0.8903 - val_iou: 0.8335 - val_recall: 0.8372 - val_precision: 0.9502\n",
            "\n",
            "Epoch 00073: val_loss did not improve from 0.10411\n",
            "Epoch 74/100\n",
            "21/21 [==============================] - 12s 561ms/step - loss: 0.0812 - dice: 0.9252 - iou: 0.8761 - recall: 0.8523 - precision: 0.9745 - val_loss: 0.1354 - val_dice: 0.8783 - val_iou: 0.8239 - val_recall: 0.8171 - val_precision: 0.9458\n",
            "\n",
            "Epoch 00074: val_loss did not improve from 0.10411\n",
            "Epoch 75/100\n",
            "21/21 [==============================] - 12s 559ms/step - loss: 0.0879 - dice: 0.9186 - iou: 0.8597 - recall: 0.8456 - precision: 0.9700 - val_loss: 0.1328 - val_dice: 0.8850 - val_iou: 0.8277 - val_recall: 0.8218 - val_precision: 0.9455\n",
            "\n",
            "Epoch 00075: val_loss did not improve from 0.10411\n",
            "Epoch 76/100\n",
            "21/21 [==============================] - 12s 559ms/step - loss: 0.0833 - dice: 0.9223 - iou: 0.8721 - recall: 0.8507 - precision: 0.9733 - val_loss: 0.1364 - val_dice: 0.8781 - val_iou: 0.8263 - val_recall: 0.8163 - val_precision: 0.9480\n",
            "\n",
            "Epoch 00076: val_loss did not improve from 0.10411\n",
            "Epoch 77/100\n",
            "21/21 [==============================] - 12s 587ms/step - loss: 0.0843 - dice: 0.9226 - iou: 0.8720 - recall: 0.8472 - precision: 0.9743 - val_loss: 0.1314 - val_dice: 0.8840 - val_iou: 0.8312 - val_recall: 0.8231 - val_precision: 0.9455\n",
            "\n",
            "Epoch 00077: val_loss did not improve from 0.10411\n",
            "Epoch 78/100\n",
            "21/21 [==============================] - 12s 586ms/step - loss: 0.0847 - dice: 0.9222 - iou: 0.8713 - recall: 0.8486 - precision: 0.9729 - val_loss: 0.1379 - val_dice: 0.8765 - val_iou: 0.8181 - val_recall: 0.8151 - val_precision: 0.9437\n",
            "\n",
            "Epoch 00078: val_loss did not improve from 0.10411\n",
            "\n",
            "Epoch 00078: ReduceLROnPlateau reducing learning rate to 1.0000001111620805e-07.\n",
            "Epoch 79/100\n",
            "21/21 [==============================] - 12s 587ms/step - loss: 0.0839 - dice: 0.9223 - iou: 0.8710 - recall: 0.8506 - precision: 0.9726 - val_loss: 0.1659 - val_dice: 0.8531 - val_iou: 0.8044 - val_recall: 0.7717 - val_precision: 0.9496\n",
            "\n",
            "Epoch 00079: val_loss did not improve from 0.10411\n",
            "Epoch 80/100\n",
            "21/21 [==============================] - 12s 563ms/step - loss: 0.0825 - dice: 0.9240 - iou: 0.8743 - recall: 0.8504 - precision: 0.9740 - val_loss: 0.1310 - val_dice: 0.8848 - val_iou: 0.8318 - val_recall: 0.8193 - val_precision: 0.9460\n",
            "\n",
            "Epoch 00080: val_loss did not improve from 0.10411\n",
            "Epoch 81/100\n",
            "21/21 [==============================] - 12s 561ms/step - loss: 0.0862 - dice: 0.9204 - iou: 0.8693 - recall: 0.8483 - precision: 0.9707 - val_loss: 0.1315 - val_dice: 0.8861 - val_iou: 0.8287 - val_recall: 0.8167 - val_precision: 0.9477\n",
            "\n",
            "Epoch 00081: val_loss did not improve from 0.10411\n",
            "Epoch 82/100\n",
            "21/21 [==============================] - 12s 588ms/step - loss: 0.0834 - dice: 0.9230 - iou: 0.8721 - recall: 0.8500 - precision: 0.9719 - val_loss: 0.1613 - val_dice: 0.8631 - val_iou: 0.8105 - val_recall: 0.7757 - val_precision: 0.9461\n",
            "\n",
            "Epoch 00082: val_loss did not improve from 0.10411\n",
            "Epoch 83/100\n",
            "21/21 [==============================] - 12s 563ms/step - loss: 0.0820 - dice: 0.9241 - iou: 0.8736 - recall: 0.8511 - precision: 0.9737 - val_loss: 0.1213 - val_dice: 0.8901 - val_iou: 0.8285 - val_recall: 0.8352 - val_precision: 0.9516\n",
            "\n",
            "Epoch 00083: val_loss did not improve from 0.10411\n",
            "Epoch 84/100\n",
            "21/21 [==============================] - 12s 564ms/step - loss: 0.0849 - dice: 0.9218 - iou: 0.8708 - recall: 0.8486 - precision: 0.9726 - val_loss: 0.1434 - val_dice: 0.8710 - val_iou: 0.8161 - val_recall: 0.7918 - val_precision: 0.9431\n",
            "\n",
            "Epoch 00084: val_loss did not improve from 0.10411\n",
            "Epoch 85/100\n",
            "21/21 [==============================] - 12s 588ms/step - loss: 0.0834 - dice: 0.9232 - iou: 0.8731 - recall: 0.8511 - precision: 0.9736 - val_loss: 0.1375 - val_dice: 0.8791 - val_iou: 0.8233 - val_recall: 0.8175 - val_precision: 0.9440\n",
            "\n",
            "Epoch 00085: val_loss did not improve from 0.10411\n",
            "Epoch 86/100\n",
            "21/21 [==============================] - 12s 561ms/step - loss: 0.0812 - dice: 0.9253 - iou: 0.8758 - recall: 0.8507 - precision: 0.9746 - val_loss: 0.1319 - val_dice: 0.8843 - val_iou: 0.8271 - val_recall: 0.8169 - val_precision: 0.9479\n",
            "\n",
            "Epoch 00086: val_loss did not improve from 0.10411\n",
            "Epoch 87/100\n",
            "21/21 [==============================] - 12s 561ms/step - loss: 0.0843 - dice: 0.9222 - iou: 0.8706 - recall: 0.8506 - precision: 0.9699 - val_loss: 0.1337 - val_dice: 0.8828 - val_iou: 0.8333 - val_recall: 0.8227 - val_precision: 0.9470\n",
            "\n",
            "Epoch 00087: val_loss did not improve from 0.10411\n",
            "Epoch 88/100\n",
            "21/21 [==============================] - 12s 561ms/step - loss: 0.0852 - dice: 0.9222 - iou: 0.8674 - recall: 0.8475 - precision: 0.9729 - val_loss: 0.1234 - val_dice: 0.8923 - val_iou: 0.8409 - val_recall: 0.8325 - val_precision: 0.9459\n",
            "\n",
            "Epoch 00088: val_loss did not improve from 0.10411\n",
            "\n",
            "Epoch 00088: ReduceLROnPlateau reducing learning rate to 1.000000082740371e-08.\n",
            "Epoch 89/100\n",
            "21/21 [==============================] - 12s 587ms/step - loss: 0.0825 - dice: 0.9237 - iou: 0.8735 - recall: 0.8531 - precision: 0.9712 - val_loss: 0.1539 - val_dice: 0.8681 - val_iou: 0.8147 - val_recall: 0.7921 - val_precision: 0.9488\n",
            "\n",
            "Epoch 00089: val_loss did not improve from 0.10411\n",
            "Epoch 90/100\n",
            "21/21 [==============================] - 12s 561ms/step - loss: 0.0849 - dice: 0.9222 - iou: 0.8699 - recall: 0.8503 - precision: 0.9706 - val_loss: 0.1270 - val_dice: 0.8875 - val_iou: 0.8276 - val_recall: 0.8238 - val_precision: 0.9428\n",
            "\n",
            "Epoch 00090: val_loss did not improve from 0.10411\n",
            "Epoch 91/100\n",
            "21/21 [==============================] - 12s 589ms/step - loss: 0.0828 - dice: 0.9230 - iou: 0.8727 - recall: 0.8500 - precision: 0.9737 - val_loss: 0.1349 - val_dice: 0.8812 - val_iou: 0.8265 - val_recall: 0.8166 - val_precision: 0.9482\n",
            "\n",
            "Epoch 00091: val_loss did not improve from 0.10411\n",
            "Epoch 92/100\n",
            "21/21 [==============================] - 12s 588ms/step - loss: 0.0833 - dice: 0.9234 - iou: 0.8725 - recall: 0.8505 - precision: 0.9731 - val_loss: 0.1441 - val_dice: 0.8764 - val_iou: 0.8209 - val_recall: 0.7893 - val_precision: 0.9462\n",
            "\n",
            "Epoch 00092: val_loss did not improve from 0.10411\n",
            "Epoch 93/100\n",
            "21/21 [==============================] - 12s 589ms/step - loss: 0.0853 - dice: 0.9214 - iou: 0.8693 - recall: 0.8471 - precision: 0.9740 - val_loss: 0.1238 - val_dice: 0.8881 - val_iou: 0.8318 - val_recall: 0.8261 - val_precision: 0.9508\n",
            "\n",
            "Epoch 00093: val_loss did not improve from 0.10411\n",
            "Epoch 94/100\n",
            "21/21 [==============================] - 12s 563ms/step - loss: 0.0822 - dice: 0.9241 - iou: 0.8737 - recall: 0.8517 - precision: 0.9738 - val_loss: 0.1486 - val_dice: 0.8705 - val_iou: 0.8148 - val_recall: 0.8012 - val_precision: 0.9487\n",
            "\n",
            "Epoch 00094: val_loss did not improve from 0.10411\n",
            "Epoch 95/100\n",
            "21/21 [==============================] - 12s 590ms/step - loss: 0.0827 - dice: 0.9235 - iou: 0.8742 - recall: 0.8495 - precision: 0.9739 - val_loss: 0.1170 - val_dice: 0.8957 - val_iou: 0.8352 - val_recall: 0.8452 - val_precision: 0.9400\n",
            "\n",
            "Epoch 00095: val_loss did not improve from 0.10411\n",
            "Epoch 96/100\n",
            "21/21 [==============================] - 12s 588ms/step - loss: 0.0835 - dice: 0.9227 - iou: 0.8718 - recall: 0.8487 - precision: 0.9742 - val_loss: 0.1368 - val_dice: 0.8799 - val_iou: 0.8273 - val_recall: 0.8170 - val_precision: 0.9478\n",
            "\n",
            "Epoch 00096: val_loss did not improve from 0.10411\n",
            "Epoch 97/100\n",
            "21/21 [==============================] - 12s 558ms/step - loss: 0.0845 - dice: 0.9223 - iou: 0.8705 - recall: 0.8500 - precision: 0.9709 - val_loss: 0.1200 - val_dice: 0.8943 - val_iou: 0.8345 - val_recall: 0.8307 - val_precision: 0.9531\n",
            "\n",
            "Epoch 00097: val_loss did not improve from 0.10411\n",
            "Epoch 98/100\n",
            "21/21 [==============================] - 12s 560ms/step - loss: 0.0844 - dice: 0.9219 - iou: 0.8647 - recall: 0.8497 - precision: 0.9718 - val_loss: 0.1146 - val_dice: 0.8978 - val_iou: 0.8429 - val_recall: 0.8401 - val_precision: 0.9600\n",
            "\n",
            "Epoch 00098: val_loss did not improve from 0.10411\n",
            "\n",
            "Epoch 00098: ReduceLROnPlateau reducing learning rate to 1.000000082740371e-09.\n",
            "Epoch 99/100\n",
            "21/21 [==============================] - 12s 561ms/step - loss: 0.0815 - dice: 0.9251 - iou: 0.8754 - recall: 0.8521 - precision: 0.9745 - val_loss: 0.1395 - val_dice: 0.8788 - val_iou: 0.8299 - val_recall: 0.8192 - val_precision: 0.9412\n",
            "\n",
            "Epoch 00099: val_loss did not improve from 0.10411\n",
            "Epoch 100/100\n",
            "21/21 [==============================] - 12s 590ms/step - loss: 0.0838 - dice: 0.9225 - iou: 0.8710 - recall: 0.8493 - precision: 0.9736 - val_loss: 0.1438 - val_dice: 0.8744 - val_iou: 0.8209 - val_recall: 0.8052 - val_precision: 0.9512\n",
            "\n",
            "Epoch 00100: val_loss did not improve from 0.10411\n",
            "Saved model and weights in the directory\n",
            "Loaded best weights:  /content/drive/MyDrive/CVDL/dataset/results/models/dbunet_BN_Glorot_e100_lr0.001/Model.hdf5\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VPk54x9RYT4A"
      },
      "source": [
        "## Evaluating"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 896
        },
        "id": "7LY5fghx609Z",
        "outputId": "b702a66c-1557-4c9f-8e72-a0e321335869"
      },
      "source": [
        "results_tf = model.evaluate(dataset_list[0]['test'], dataset_list[1]['test'], return_dict = True)\n",
        "train_figure = plots.plot_train_metric(train_history, params['nEpoches'], \"dice\")\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "2/2 [==============================] - 1s 212ms/step - loss: 0.1147 - dice: 0.8860 - iou: 0.8145 - recall: 0.8499 - precision: 0.9460\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 2304x1152 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-D-3kJvbaL9h",
        "outputId": "a7eac5af-c12d-4547-8d8d-8870d7a960b2"
      },
      "source": [
        "threshold = 0.5\n",
        "min_dice = 0.60\n",
        "results_df, avg = plots.get_results(dataset_list, params, model, threshold)\n",
        "print(results_df)\n",
        "print('\\n')\n",
        "print(f\"Threshold value: {threshold}\")\n",
        "print(f\"Best dice: {max(results_df['dice'])}\")\n",
        "print(f\"Average dice: {avg}\")\n",
        "print(f\"Minimum requested dice value (bad samples): {min_dice}\")\n",
        "good_res = results_df.loc[(results_df['dice'] >= min_dice)]\n",
        "print(f\"Average dice without bad samples: {good_res['dice'].mean()}\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "        dice\n",
            "0   0.860024\n",
            "1   0.925186\n",
            "2   0.930312\n",
            "3   0.913642\n",
            "4   0.945693\n",
            "5   0.945473\n",
            "6   0.913196\n",
            "7   0.908168\n",
            "8   0.938356\n",
            "9   0.937937\n",
            "10  0.937313\n",
            "11  0.904893\n",
            "12  0.926339\n",
            "13  0.908891\n",
            "14  0.950678\n",
            "15  0.935342\n",
            "16  0.928763\n",
            "17  0.941118\n",
            "18  0.879325\n",
            "19  0.814163\n",
            "20  0.898998\n",
            "21  0.924149\n",
            "22  0.944255\n",
            "23  0.959441\n",
            "24  0.932267\n",
            "25  0.936741\n",
            "26  0.928105\n",
            "27  0.923182\n",
            "28  0.907796\n",
            "29  0.932954\n",
            "30  0.812667\n",
            "31  0.956646\n",
            "32  0.705882\n",
            "33  0.943216\n",
            "34  0.880884\n",
            "35  0.947687\n",
            "36  0.926745\n",
            "37  0.893691\n",
            "38  0.904203\n",
            "39  0.825234\n",
            "40  0.905782\n",
            "41  0.906819\n",
            "42  0.799389\n",
            "43  0.677693\n",
            "\n",
            "\n",
            "Threshold value: 0.5\n",
            "Best dice: 0.9594405594405594\n",
            "Average dice: {'dice': 0.9027099356017142}\n",
            "Minimum requested dice value (bad samples): 0.6\n",
            "Average dice without bad samples: 0.9027099356017142\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XVaYmcK2bwVz",
        "outputId": "0397b9fe-17b6-44c1-f502-40f111355867"
      },
      "source": [
        "bad_res = results_df.loc[(results_df['dice'] < min_dice)]\n",
        "print(bad_res)\n",
        "\n",
        "num = bad_res['dice'].count()\n",
        "den = results_df['dice'].count()\n",
        "\n",
        "print(num/den)\n",
        "print('\\n')\n",
        "print(\"Loading bad samples... :\")\n",
        "print('\\n')\n",
        "for idx in bad_res.index:\n",
        "  print('\\n')\n",
        "  print('\\n')\n",
        "  print(f\"Image index {idx} - dice: {bad_res.loc[idx]}\")\n",
        "  threshold = 0.5\n",
        "  test_image = dataset_list[0]['test'][idx,:,:,0]\n",
        "  test_mask = dataset_list[1]['test'][idx,:,:,0]\n",
        "  test_results = lib.get_preditcions(test_image, test_mask, params, model, threshold)\n",
        "  results_figure = plots.plot_test_results(test_results)\n",
        "\n",
        "  plt.show()\n",
        "#results_figure.savefig(f'/content/drive/MyDrive/CVDL/trained_models/{name}/bad_samples.png')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Empty DataFrame\n",
            "Columns: [dice]\n",
            "Index: []\n",
            "0.0\n",
            "\n",
            "\n",
            "Loading bad samples... :\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c0Dn6v6UcKum",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2e65d054-5c1b-40cd-b07d-00bb173a959f"
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_1 (InputLayer)            [(None, 288, 384, 1) 0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 288, 384, 3)  0           input_1[0][0]                    \n",
            "                                                                 input_1[0][0]                    \n",
            "                                                                 input_1[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "block1_conv1 (Conv2D)           (None, 288, 384, 64) 1792        concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "block1_conv2 (Conv2D)           (None, 288, 384, 64) 36928       block1_conv1[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "block1_pool (MaxPooling2D)      (None, 144, 192, 64) 0           block1_conv2[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "block2_conv1 (Conv2D)           (None, 144, 192, 128 73856       block1_pool[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "block2_conv2 (Conv2D)           (None, 144, 192, 128 147584      block2_conv1[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "block2_pool (MaxPooling2D)      (None, 72, 96, 128)  0           block2_conv2[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "block3_conv1 (Conv2D)           (None, 72, 96, 256)  295168      block2_pool[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "block3_conv2 (Conv2D)           (None, 72, 96, 256)  590080      block3_conv1[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "block3_conv3 (Conv2D)           (None, 72, 96, 256)  590080      block3_conv2[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "block3_conv4 (Conv2D)           (None, 72, 96, 256)  590080      block3_conv3[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "block3_pool (MaxPooling2D)      (None, 36, 48, 256)  0           block3_conv4[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "block4_conv1 (Conv2D)           (None, 36, 48, 512)  1180160     block3_pool[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "block4_conv2 (Conv2D)           (None, 36, 48, 512)  2359808     block4_conv1[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "block4_conv3 (Conv2D)           (None, 36, 48, 512)  2359808     block4_conv2[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "block4_conv4 (Conv2D)           (None, 36, 48, 512)  2359808     block4_conv3[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "average_pooling2d (AveragePooli (None, 1, 1, 512)    0           block4_conv4[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv2d (Conv2D)                 (None, 1, 1, 64)     32832       average_pooling2d[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization (BatchNorma (None, 1, 1, 64)     256         conv2d[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_1 (Conv2D)               (None, 36, 48, 64)   32768       block4_conv4[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_2 (Conv2D)               (None, 36, 48, 64)   294912      block4_conv4[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_3 (Conv2D)               (None, 36, 48, 64)   294912      block4_conv4[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_4 (Conv2D)               (None, 36, 48, 64)   294912      block4_conv4[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "activation (Activation)         (None, 1, 1, 64)     0           batch_normalization[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_1 (BatchNor (None, 36, 48, 64)   256         conv2d_1[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_2 (BatchNor (None, 36, 48, 64)   256         conv2d_2[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_3 (BatchNor (None, 36, 48, 64)   256         conv2d_3[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_4 (BatchNor (None, 36, 48, 64)   256         conv2d_4[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "up_sampling2d (UpSampling2D)    (None, 36, 48, 64)   0           activation[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "activation_1 (Activation)       (None, 36, 48, 64)   0           batch_normalization_1[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "activation_2 (Activation)       (None, 36, 48, 64)   0           batch_normalization_2[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "activation_3 (Activation)       (None, 36, 48, 64)   0           batch_normalization_3[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "activation_4 (Activation)       (None, 36, 48, 64)   0           batch_normalization_4[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "concatenate_1 (Concatenate)     (None, 36, 48, 320)  0           up_sampling2d[0][0]              \n",
            "                                                                 activation_1[0][0]               \n",
            "                                                                 activation_2[0][0]               \n",
            "                                                                 activation_3[0][0]               \n",
            "                                                                 activation_4[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_5 (Conv2D)               (None, 36, 48, 64)   20480       concatenate_1[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_5 (BatchNor (None, 36, 48, 64)   256         conv2d_5[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "activation_5 (Activation)       (None, 36, 48, 64)   0           batch_normalization_5[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "up_sampling2d_1 (UpSampling2D)  (None, 72, 96, 64)   0           activation_5[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "concatenate_2 (Concatenate)     (None, 72, 96, 320)  0           up_sampling2d_1[0][0]            \n",
            "                                                                 block3_conv4[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_6 (Conv2D)               (None, 72, 96, 128)  368768      concatenate_2[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_6 (BatchNor (None, 72, 96, 128)  512         conv2d_6[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "activation_6 (Activation)       (None, 72, 96, 128)  0           batch_normalization_6[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_7 (Conv2D)               (None, 72, 96, 128)  147584      activation_6[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_7 (BatchNor (None, 72, 96, 128)  512         conv2d_7[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "activation_7 (Activation)       (None, 72, 96, 128)  0           batch_normalization_7[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "global_average_pooling2d (Globa (None, 128)          0           activation_7[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 128)    0           global_average_pooling2d[0][0]   \n",
            "__________________________________________________________________________________________________\n",
            "dense (Dense)                   (None, 1, 1, 16)     2048        reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "dense_1 (Dense)                 (None, 1, 1, 128)    2048        dense[0][0]                      \n",
            "__________________________________________________________________________________________________\n",
            "multiply (Multiply)             (None, 72, 96, 128)  0           activation_7[0][0]               \n",
            "                                                                 dense_1[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "up_sampling2d_2 (UpSampling2D)  (None, 144, 192, 128 0           multiply[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "concatenate_3 (Concatenate)     (None, 144, 192, 256 0           up_sampling2d_2[0][0]            \n",
            "                                                                 block2_conv2[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_8 (Conv2D)               (None, 144, 192, 64) 147520      concatenate_3[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_8 (BatchNor (None, 144, 192, 64) 256         conv2d_8[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "activation_8 (Activation)       (None, 144, 192, 64) 0           batch_normalization_8[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_9 (Conv2D)               (None, 144, 192, 64) 36928       activation_8[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_9 (BatchNor (None, 144, 192, 64) 256         conv2d_9[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "activation_9 (Activation)       (None, 144, 192, 64) 0           batch_normalization_9[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "global_average_pooling2d_1 (Glo (None, 64)           0           activation_9[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "reshape_1 (Reshape)             (None, 1, 1, 64)     0           global_average_pooling2d_1[0][0] \n",
            "__________________________________________________________________________________________________\n",
            "dense_2 (Dense)                 (None, 1, 1, 8)      512         reshape_1[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_3 (Dense)                 (None, 1, 1, 64)     512         dense_2[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "multiply_1 (Multiply)           (None, 144, 192, 64) 0           activation_9[0][0]               \n",
            "                                                                 dense_3[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "up_sampling2d_3 (UpSampling2D)  (None, 288, 384, 64) 0           multiply_1[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "concatenate_4 (Concatenate)     (None, 288, 384, 128 0           up_sampling2d_3[0][0]            \n",
            "                                                                 block1_conv2[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_10 (Conv2D)              (None, 288, 384, 32) 36896       concatenate_4[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_10 (BatchNo (None, 288, 384, 32) 128         conv2d_10[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_10 (Activation)      (None, 288, 384, 32) 0           batch_normalization_10[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_11 (Conv2D)              (None, 288, 384, 32) 9248        activation_10[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_11 (BatchNo (None, 288, 384, 32) 128         conv2d_11[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_11 (Activation)      (None, 288, 384, 32) 0           batch_normalization_11[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "global_average_pooling2d_2 (Glo (None, 32)           0           activation_11[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "reshape_2 (Reshape)             (None, 1, 1, 32)     0           global_average_pooling2d_2[0][0] \n",
            "__________________________________________________________________________________________________\n",
            "dense_4 (Dense)                 (None, 1, 1, 4)      128         reshape_2[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_5 (Dense)                 (None, 1, 1, 32)     128         dense_4[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "multiply_2 (Multiply)           (None, 288, 384, 32) 0           activation_11[0][0]              \n",
            "                                                                 dense_5[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_12 (Conv2D)              (None, 288, 384, 2)  578         multiply_2[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "activation_12 (Activation)      (None, 288, 384, 2)  0           conv2d_12[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_13 (Conv2D)              (None, 288, 384, 1)  3           activation_12[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "activation_13 (Activation)      (None, 288, 384, 1)  0           conv2d_13[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "tf.math.multiply (TFOpLambda)   (None, 288, 384, 1)  0           input_1[0][0]                    \n",
            "                                                                 activation_13[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_14 (Conv2D)              (None, 288, 384, 32) 320         tf.math.multiply[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_12 (BatchNo (None, 288, 384, 32) 128         conv2d_14[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_14 (Activation)      (None, 288, 384, 32) 0           batch_normalization_12[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_15 (Conv2D)              (None, 288, 384, 32) 9248        activation_14[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_13 (BatchNo (None, 288, 384, 32) 128         conv2d_15[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_15 (Activation)      (None, 288, 384, 32) 0           batch_normalization_13[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "global_average_pooling2d_3 (Glo (None, 32)           0           activation_15[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "reshape_3 (Reshape)             (None, 1, 1, 32)     0           global_average_pooling2d_3[0][0] \n",
            "__________________________________________________________________________________________________\n",
            "dense_6 (Dense)                 (None, 1, 1, 4)      128         reshape_3[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_7 (Dense)                 (None, 1, 1, 32)     128         dense_6[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "multiply_3 (Multiply)           (None, 288, 384, 32) 0           activation_15[0][0]              \n",
            "                                                                 dense_7[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D)    (None, 144, 192, 32) 0           multiply_3[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_16 (Conv2D)              (None, 144, 192, 64) 18496       max_pooling2d[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_14 (BatchNo (None, 144, 192, 64) 256         conv2d_16[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_16 (Activation)      (None, 144, 192, 64) 0           batch_normalization_14[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_17 (Conv2D)              (None, 144, 192, 64) 36928       activation_16[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_15 (BatchNo (None, 144, 192, 64) 256         conv2d_17[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_17 (Activation)      (None, 144, 192, 64) 0           batch_normalization_15[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "global_average_pooling2d_4 (Glo (None, 64)           0           activation_17[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "reshape_4 (Reshape)             (None, 1, 1, 64)     0           global_average_pooling2d_4[0][0] \n",
            "__________________________________________________________________________________________________\n",
            "dense_8 (Dense)                 (None, 1, 1, 8)      512         reshape_4[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_9 (Dense)                 (None, 1, 1, 64)     512         dense_8[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "multiply_4 (Multiply)           (None, 144, 192, 64) 0           activation_17[0][0]              \n",
            "                                                                 dense_9[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling2d_1 (MaxPooling2D)  (None, 72, 96, 64)   0           multiply_4[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_18 (Conv2D)              (None, 72, 96, 128)  73856       max_pooling2d_1[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_16 (BatchNo (None, 72, 96, 128)  512         conv2d_18[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_18 (Activation)      (None, 72, 96, 128)  0           batch_normalization_16[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_19 (Conv2D)              (None, 72, 96, 128)  147584      activation_18[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_17 (BatchNo (None, 72, 96, 128)  512         conv2d_19[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_19 (Activation)      (None, 72, 96, 128)  0           batch_normalization_17[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "global_average_pooling2d_5 (Glo (None, 128)          0           activation_19[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "reshape_5 (Reshape)             (None, 1, 1, 128)    0           global_average_pooling2d_5[0][0] \n",
            "__________________________________________________________________________________________________\n",
            "dense_10 (Dense)                (None, 1, 1, 16)     2048        reshape_5[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_11 (Dense)                (None, 1, 1, 128)    2048        dense_10[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "multiply_5 (Multiply)           (None, 72, 96, 128)  0           activation_19[0][0]              \n",
            "                                                                 dense_11[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling2d_2 (MaxPooling2D)  (None, 36, 48, 128)  0           multiply_5[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "average_pooling2d_1 (AveragePoo (None, 1, 1, 128)    0           max_pooling2d_2[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_20 (Conv2D)              (None, 1, 1, 64)     8256        average_pooling2d_1[0][0]        \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_18 (BatchNo (None, 1, 1, 64)     256         conv2d_20[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_21 (Conv2D)              (None, 36, 48, 64)   8192        max_pooling2d_2[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_22 (Conv2D)              (None, 36, 48, 64)   73728       max_pooling2d_2[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_23 (Conv2D)              (None, 36, 48, 64)   73728       max_pooling2d_2[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_24 (Conv2D)              (None, 36, 48, 64)   73728       max_pooling2d_2[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "activation_20 (Activation)      (None, 1, 1, 64)     0           batch_normalization_18[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_19 (BatchNo (None, 36, 48, 64)   256         conv2d_21[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_20 (BatchNo (None, 36, 48, 64)   256         conv2d_22[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_21 (BatchNo (None, 36, 48, 64)   256         conv2d_23[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_22 (BatchNo (None, 36, 48, 64)   256         conv2d_24[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "up_sampling2d_4 (UpSampling2D)  (None, 36, 48, 64)   0           activation_20[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "activation_21 (Activation)      (None, 36, 48, 64)   0           batch_normalization_19[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_22 (Activation)      (None, 36, 48, 64)   0           batch_normalization_20[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_23 (Activation)      (None, 36, 48, 64)   0           batch_normalization_21[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "activation_24 (Activation)      (None, 36, 48, 64)   0           batch_normalization_22[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "concatenate_5 (Concatenate)     (None, 36, 48, 320)  0           up_sampling2d_4[0][0]            \n",
            "                                                                 activation_21[0][0]              \n",
            "                                                                 activation_22[0][0]              \n",
            "                                                                 activation_23[0][0]              \n",
            "                                                                 activation_24[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_25 (Conv2D)              (None, 36, 48, 64)   20480       concatenate_5[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_23 (BatchNo (None, 36, 48, 64)   256         conv2d_25[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_25 (Activation)      (None, 36, 48, 64)   0           batch_normalization_23[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_26 (Conv2D)              (None, 36, 48, 256)  147712      activation_25[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_24 (BatchNo (None, 36, 48, 256)  1024        conv2d_26[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_26 (Activation)      (None, 36, 48, 256)  0           batch_normalization_24[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_27 (Conv2D)              (None, 36, 48, 256)  590080      activation_26[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_25 (BatchNo (None, 36, 48, 256)  1024        conv2d_27[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_27 (Activation)      (None, 36, 48, 256)  0           batch_normalization_25[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "up_sampling2d_5 (UpSampling2D)  (None, 72, 96, 256)  0           activation_27[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "concatenate_6 (Concatenate)     (None, 72, 96, 640)  0           up_sampling2d_5[0][0]            \n",
            "                                                                 block3_conv4[0][0]               \n",
            "                                                                 multiply_5[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_28 (Conv2D)              (None, 72, 96, 128)  737408      concatenate_6[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_26 (BatchNo (None, 72, 96, 128)  512         conv2d_28[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_28 (Activation)      (None, 72, 96, 128)  0           batch_normalization_26[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_29 (Conv2D)              (None, 72, 96, 128)  147584      activation_28[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_27 (BatchNo (None, 72, 96, 128)  512         conv2d_29[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_29 (Activation)      (None, 72, 96, 128)  0           batch_normalization_27[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "global_average_pooling2d_6 (Glo (None, 128)          0           activation_29[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "reshape_6 (Reshape)             (None, 1, 1, 128)    0           global_average_pooling2d_6[0][0] \n",
            "__________________________________________________________________________________________________\n",
            "dense_12 (Dense)                (None, 1, 1, 16)     2048        reshape_6[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_13 (Dense)                (None, 1, 1, 128)    2048        dense_12[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "multiply_6 (Multiply)           (None, 72, 96, 128)  0           activation_29[0][0]              \n",
            "                                                                 dense_13[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "up_sampling2d_6 (UpSampling2D)  (None, 144, 192, 128 0           multiply_6[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "concatenate_7 (Concatenate)     (None, 144, 192, 320 0           up_sampling2d_6[0][0]            \n",
            "                                                                 block2_conv2[0][0]               \n",
            "                                                                 multiply_4[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_30 (Conv2D)              (None, 144, 192, 64) 184384      concatenate_7[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_28 (BatchNo (None, 144, 192, 64) 256         conv2d_30[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_30 (Activation)      (None, 144, 192, 64) 0           batch_normalization_28[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_31 (Conv2D)              (None, 144, 192, 64) 36928       activation_30[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_29 (BatchNo (None, 144, 192, 64) 256         conv2d_31[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_31 (Activation)      (None, 144, 192, 64) 0           batch_normalization_29[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "global_average_pooling2d_7 (Glo (None, 64)           0           activation_31[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "reshape_7 (Reshape)             (None, 1, 1, 64)     0           global_average_pooling2d_7[0][0] \n",
            "__________________________________________________________________________________________________\n",
            "dense_14 (Dense)                (None, 1, 1, 8)      512         reshape_7[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_15 (Dense)                (None, 1, 1, 64)     512         dense_14[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "multiply_7 (Multiply)           (None, 144, 192, 64) 0           activation_31[0][0]              \n",
            "                                                                 dense_15[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "up_sampling2d_7 (UpSampling2D)  (None, 288, 384, 64) 0           multiply_7[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "concatenate_8 (Concatenate)     (None, 288, 384, 160 0           up_sampling2d_7[0][0]            \n",
            "                                                                 block1_conv2[0][0]               \n",
            "                                                                 multiply_3[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_32 (Conv2D)              (None, 288, 384, 32) 46112       concatenate_8[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_30 (BatchNo (None, 288, 384, 32) 128         conv2d_32[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_32 (Activation)      (None, 288, 384, 32) 0           batch_normalization_30[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_33 (Conv2D)              (None, 288, 384, 32) 9248        activation_32[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_31 (BatchNo (None, 288, 384, 32) 128         conv2d_33[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "activation_33 (Activation)      (None, 288, 384, 32) 0           batch_normalization_31[0][0]     \n",
            "__________________________________________________________________________________________________\n",
            "global_average_pooling2d_8 (Glo (None, 32)           0           activation_33[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "reshape_8 (Reshape)             (None, 1, 1, 32)     0           global_average_pooling2d_8[0][0] \n",
            "__________________________________________________________________________________________________\n",
            "dense_16 (Dense)                (None, 1, 1, 4)      128         reshape_8[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_17 (Dense)                (None, 1, 1, 32)     128         dense_16[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "multiply_8 (Multiply)           (None, 288, 384, 32) 0           activation_33[0][0]              \n",
            "                                                                 dense_17[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_34 (Conv2D)              (None, 288, 384, 2)  578         multiply_8[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "activation_34 (Activation)      (None, 288, 384, 2)  0           conv2d_34[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "conv2d_35 (Conv2D)              (None, 288, 384, 1)  3           activation_34[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "activation_35 (Activation)      (None, 288, 384, 1)  0           conv2d_35[0][0]                  \n",
            "==================================================================================================\n",
            "Total params: 14,774,698\n",
            "Trainable params: 14,769,450\n",
            "Non-trainable params: 5,248\n",
            "__________________________________________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "w2Pjy39gYnI5"
      },
      "source": [
        "## Saving"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eYmRJJ9LCkRn"
      },
      "source": [
        "name = 'doubleunet_d0.9027'\n",
        "model.save(f'/content/drive/MyDrive/CVDL/trained_models/{name}/model.h5', save_format='h5')\n",
        "lib.create_dump(train_history.history, f'/content/drive/MyDrive/CVDL/trained_models/{name}/train_history')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_C_dy_XSOFG9",
        "outputId": "c11e58ea-2ac1-4dcd-babc-a7e52dd79a57"
      },
      "source": [
        "from tensorflow import keras\n",
        "\n",
        "c_objects = {\n",
        "    \"bce_dice_loss\": metrics.bce_dice_loss,\n",
        "    \"dice\": metrics.dice,\n",
        "    \"iou\": metrics.iou,\n",
        "    \"precision\": tf.keras.metrics.Precision(),\n",
        "    \"recall\": tf.keras.metrics.Recall(),\n",
        "}\n",
        "\n",
        "loaded_model = keras.models.load_model(f'/content/drive/MyDrive/CVDL/trained_models/{name}/model.h5', custom_objects=c_objects)\n",
        "\n",
        "loaded_model.compile(loss=metrics.bce_dice_loss, optimizer='adam', metrics=metric)\n",
        "\n",
        "results_tf = loaded_model.evaluate(dataset_list[0]['test'], dataset_list[1]['test'], return_dict = True)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "2/2 [==============================] - 4s 212ms/step - loss: 0.1147 - dice: 0.8860 - iou: 0.8145 - recall: 0.8499 - precision: 0.9460\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}