S8_Fit_ResNet50.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "colab": {
      "name": "S8_Fit_ResNet50.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/liuxx479/Outlier/blob/master/S8_Fit_ResNet50.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "stTiXCP38WMh",
        "colab_type": "code",
        "outputId": "05755b97-277c-4693-e3dd-05787d47f6b3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "%pylab inline\n",
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras.layers import *\n",
        "from tensorflow.keras.initializers import glorot_uniform\n",
        "from tensorflow.keras import Model\n",
        "import numpy as np\n",
        "import os\n",
        "import tensorflow_datasets as tfds"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Populating the interactive namespace from numpy and matplotlib\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "n1TFhPhC8-8T",
        "colab_type": "code",
        "outputId": "c700b174-d551-4f34-a02a-f7cbdbe1eaf4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "\n",
        "\n",
        "!pip install --quiet git+https://github.com/liuxx479/Outlier.git\n",
        "from outlier.datasets import gaussian_convergence\n",
        "\n"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "  Building wheel for Outlier (setup.py) ... \u001b[?25l\u001b[?25hdone\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JcfZXpBN8WM7",
        "colab_type": "code",
        "outputId": "9e19aed4-6e0a-454a-fc3f-3bcd607ebf0d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 126
        }
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')\n",
        "\n",
        "root = '/content/drive/My Drive'\n",
        "data_path = os.path.join(root,'tensorflow_datasets/')\n",
        "model_path= os.path.join(root,'models/Keras')\n",
        "\n",
        "if not os.path.isdir(model_path):\n",
        "    os.makedirs(model_path)"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&response_type=code&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "arJDnswa8WNK",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "dataset, info = tfds.load(name=\"gaussian_convergence\", split=\"train\", data_dir=data_path,with_info=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZF61hgrB8WNS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "DATASET_SIZE = info.splits['train'].num_examples"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gaoMb8IR8WNk",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def get_S8(params):\n",
        "    om  = tf.gather(params,0)\n",
        "    si8 = tf.gather(params,1)\n",
        "    S8  = si8*tf.sqrt(om/0.3)\n",
        "    return S8\n",
        "\n",
        "## Implementing a simple resnet model\n",
        "def redefine(data):\n",
        "    x    = data['map'] \n",
        "    x    = tf.expand_dims(x,-1)\n",
        "    y    = data['params']\n",
        "    y    = get_S8(y)\n",
        "    return x, y\n",
        "\n",
        "def downsample(x,y):\n",
        "    x    = tf.image.resize(x,[128,128])\n",
        "    return x, y\n",
        "  \n",
        "def augmentation(x, y):\n",
        "  x = tf.image.random_flip_left_right(x)\n",
        "  x = tf.image.random_flip_up_down(x)\n",
        "  x = x + 0.02*tf.random.normal(shape=(128,128,1))\n",
        "  return x, y"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "O8-k4TeY8WNp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "DATASET_SIZE = 20000\n",
        "BATCH_SIZE = 8\n",
        "\n",
        "train_size = int(0.8 * DATASET_SIZE)\n",
        "test_size = int(0.2 * DATASET_SIZE)\n",
        "\n",
        "full_dataset = dataset.map(redefine).map(downsample)\n",
        "\n",
        "train_dataset= full_dataset.take(train_size).repeat().map(augmentation).shuffle(buffer_size=2000).batch(BATCH_SIZE)\n",
        "test_dataset = full_dataset.skip(train_size).take(test_size).batch(BATCH_SIZE)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "b7CDt0Bs8WNu",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def identity_block(X, f, filters, stage, block):\n",
        "    \"\"\"\n",
        "    Implementation of the identity block\n",
        "    \n",
        "    Arguments:\n",
        "    X -- input tensor of shape (m, n_H_prev, n_W_prev, n_C_prev)\n",
        "    f -- integer, specifying the shape of the middle CONV's window for the main path\n",
        "    filters -- python list of integers, defining the number of filters in the CONV layers of the main path\n",
        "    stage -- integer, used to name the layers, depending on their position in the network\n",
        "    block -- string/character, used to name the layers, depending on their position in the network\n",
        "    \n",
        "    Returns:\n",
        "    X -- output of the identity block, tensor of shape (n_H, n_W, n_C)\n",
        "    \"\"\"\n",
        "    \n",
        "    # defining name basis\n",
        "    conv_name_base = 'res' + str(stage) + block + '_branch'\n",
        "    bn_name_base = 'bn' + str(stage) + block + '_branch'\n",
        "    \n",
        "    # Retrieve Filters\n",
        "    F1, F2, F3 = filters\n",
        "    \n",
        "    # Save the input value. You'll need this later to add back to the main path. \n",
        "    X_shortcut = X\n",
        "    \n",
        "    # First component of main path\n",
        "    X = Conv2D(filters = F1, kernel_size = (1, 1), strides = (1,1), padding = 'valid', name = conv_name_base + '2a', kernel_initializer = glorot_uniform(seed=0))(X)\n",
        "    X = BatchNormalization(axis = 3, name = bn_name_base + '2a')(X)\n",
        "    X = Activation('relu')(X)\n",
        "    \n",
        "    \n",
        "    # Second component of main path\n",
        "    X = Conv2D(filters = F2, kernel_size = (f, f), strides = (1,1), padding = 'same', name = conv_name_base + '2b', kernel_initializer = glorot_uniform(seed=0))(X)\n",
        "    X = BatchNormalization(axis = 3, name = bn_name_base + '2b')(X)\n",
        "    X = Activation('relu')(X)\n",
        "\n",
        "    # Third component of main path \n",
        "    X = Conv2D(filters = F3, kernel_size = (1, 1), strides = (1,1), padding = 'valid', name = conv_name_base + '2c', kernel_initializer = glorot_uniform(seed=0))(X)\n",
        "    X = BatchNormalization(axis = 3, name = bn_name_base + '2c')(X)\n",
        "\n",
        "    # Final step: Add shortcut value to main path, and pass it through a RELU activation \n",
        "    X = Add()([X, X_shortcut])\n",
        "    X = Activation('relu')(X)\n",
        "    \n",
        "    \n",
        "    return X"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zJVSNEzb9bC2",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def convolutional_block(X, f, filters, stage, block, s = 2):\n",
        "    \"\"\"\n",
        "    Implementation of the convolutional block\n",
        "    \n",
        "    Arguments:\n",
        "    X -- input tensor of shape (m, n_H_prev, n_W_prev, n_C_prev)\n",
        "    f -- integer, specifying the shape of the middle CONV's window for the main path\n",
        "    filters -- python list of integers, defining the number of filters in the CONV layers of the main path\n",
        "    stage -- integer, used to name the layers, depending on their position in the network\n",
        "    block -- string/character, used to name the layers, depending on their position in the network\n",
        "    s -- Integer, specifying the stride to be used\n",
        "    \n",
        "    Returns:\n",
        "    X -- output of the convolutional block, tensor of shape (n_H, n_W, n_C)\n",
        "    \"\"\"\n",
        "    \n",
        "    # defining name basis\n",
        "    conv_name_base = 'res' + str(stage) + block + '_branch'\n",
        "    bn_name_base = 'bn' + str(stage) + block + '_branch'\n",
        "    \n",
        "    # Retrieve Filters\n",
        "    F1, F2, F3 = filters\n",
        "    \n",
        "    # Save the input value\n",
        "    X_shortcut = X\n",
        "\n",
        "\n",
        "    ##### MAIN PATH #####\n",
        "    # First component of main path \n",
        "    X = Conv2D(F1, (1, 1), strides = (s,s), name = conv_name_base + '2a', kernel_initializer = glorot_uniform(seed=0))(X)\n",
        "    X = BatchNormalization(axis = 3, name = bn_name_base + '2a')(X)\n",
        "    X = Activation('relu')(X)\n",
        "    \n",
        "\n",
        "    # Second component of main path \n",
        "    X = Conv2D(F2, (f,f), strides = (1,1), padding = 'same', name = conv_name_base + '2b', kernel_initializer=glorot_uniform(seed =0))(X)\n",
        "    X = BatchNormalization(axis = 3, name = bn_name_base + '2b')(X)\n",
        "    X = Activation('relu') (X)\n",
        "\n",
        "    # Third component of main path \n",
        "    X = Conv2D(F3, (1,1), strides = (1,1), padding = 'valid', name = conv_name_base + '2c', kernel_initializer=glorot_uniform(seed =0))(X)\n",
        "    X = BatchNormalization(axis = 3, name = bn_name_base + '2c')(X)\n",
        "\n",
        "    ##### SHORTCUT PATH #### \n",
        "    X_shortcut = Conv2D(F3, (1,1), strides = (s,s), padding = 'valid', name = conv_name_base + '1', kernel_initializer=glorot_uniform(seed =0))(X_shortcut)\n",
        "    X_shortcut = BatchNormalization(axis = 3, name = bn_name_base + '1')(X_shortcut)\n",
        "\n",
        "    # Final step: Add shortcut value to main path, and pass it through a RELU activation \n",
        "    X = Add()([X, X_shortcut])\n",
        "    X = Activation('relu')(X)\n",
        "    \n",
        "    \n",
        "    return X"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "n7lPsXd69fX4",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def ResNet50(input_shape = (128, 128, 1), params = 1):\n",
        "    \"\"\"\n",
        "    Implementation of the popular ResNet50 the following architecture:\n",
        "    CONV2D -> BATCHNORM -> RELU -> MAXPOOL -> CONVBLOCK -> IDBLOCK*2 -> CONVBLOCK -> IDBLOCK*3\n",
        "    -> CONVBLOCK -> IDBLOCK*5 -> CONVBLOCK -> IDBLOCK*2 -> AVGPOOL -> TOPLAYER\n",
        "\n",
        "    Arguments:\n",
        "    input_shape -- shape of the images of the dataset\n",
        "    classes -- integer, number of classes\n",
        "\n",
        "    Returns:\n",
        "    model -- a Model() instance in Keras\n",
        "    \"\"\"\n",
        "    \n",
        "    # Define the input as a tensor with shape input_shape\n",
        "    X_input = Input(input_shape)\n",
        "\n",
        "    \n",
        "    # Zero-Padding\n",
        "    X = ZeroPadding2D((3, 3))(X_input)\n",
        "    \n",
        "    # Stage 1\n",
        "    X = Conv2D(64, (7, 7), strides = (2, 2), name = 'conv1', kernel_initializer = glorot_uniform(seed=0))(X)\n",
        "    X = BatchNormalization(axis = 3, name = 'bn_conv1')(X)\n",
        "    X = Activation('relu')(X)\n",
        "    X = MaxPooling2D((3, 3), strides=(2, 2))(X)\n",
        "\n",
        "    # Stage 2\n",
        "    X = convolutional_block(X, f = 3, filters = [64, 64, 256], stage = 2, block='a', s = 1)\n",
        "    # X = identity_block(X, 3, [64, 64, 256], stage=2, block='b')\n",
        "    # X = identity_block(X, 3, [64, 64, 256], stage=2, block='c')\n",
        "\n",
        "\n",
        "    # Stage 3 \n",
        "    X = convolutional_block(X, f = 3, filters = [128, 128, 512], stage = 3, block='a', s = 2)\n",
        "    # X = identity_block(X, 3, [128, 128, 512], stage=3, block='b')\n",
        "    # X = identity_block(X, 3, [128, 128, 512], stage=3, block='c')\n",
        "    # X = identity_block(X, 3, [128, 128, 512], stage=3, block='d')\n",
        "\n",
        "    # Stage 4 \n",
        "    X = convolutional_block(X, f = 3, filters = [256, 256, 1024], stage = 4, block='a', s = 2)\n",
        "    # X = identity_block(X, 3, [256, 256, 1024], stage=4, block='b')\n",
        "    # X = identity_block(X, 3, [256, 256, 1024], stage=4, block='c')\n",
        "    # X = identity_block(X, 3, [256, 256, 1024], stage=4, block='d')\n",
        "    # X = identity_block(X, 3, [256, 256, 1024], stage=4, block='e')\n",
        "    # X = identity_block(X, 3, [256, 256, 1024], stage=4, block='f')\n",
        "\n",
        "    # Stage 5 \n",
        "    X = convolutional_block(X, f = 3, filters = [512, 512, 2048], stage = 5, block='a', s = 2)\n",
        "    # X = identity_block(X, 3, [512, 512, 2048], stage=5, block='b')\n",
        "    # X = identity_block(X, 3, [512, 512, 2048], stage=5, block='c')\n",
        "\n",
        "    # AVGPOOL . Use \"X = AveragePooling2D(...)(X)\"\n",
        "    X = AveragePooling2D()(X)\n",
        "\n",
        "    # output layer\n",
        "    X = Flatten()(X)\n",
        "    X = Dense(params, name='fc' + str(params), kernel_initializer = glorot_uniform(seed=0))(X)\n",
        "    \n",
        "    \n",
        "    # Create model\n",
        "    model = Model(inputs = X_input, outputs = X, name='ResNet50')\n",
        "\n",
        "    return model"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nvEFW8FN-A5-",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fHnSuAMQ8WOW",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "3d19511c-c205-45eb-d75c-39b52b3e93a6"
      },
      "source": [
        "res_net_model = ResNet50()\n",
        "res_net_model.summary()"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"ResNet50\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_1 (InputLayer)            [(None, 128, 128, 1) 0                                            \n",
            "__________________________________________________________________________________________________\n",
            "zero_padding2d (ZeroPadding2D)  (None, 134, 134, 1)  0           input_1[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "conv1 (Conv2D)                  (None, 64, 64, 64)   3200        zero_padding2d[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "bn_conv1 (BatchNormalization)   (None, 64, 64, 64)   256         conv1[0][0]                      \n",
            "__________________________________________________________________________________________________\n",
            "activation (Activation)         (None, 64, 64, 64)   0           bn_conv1[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D)    (None, 31, 31, 64)   0           activation[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "res2a_branch2a (Conv2D)         (None, 31, 31, 64)   4160        max_pooling2d[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "bn2a_branch2a (BatchNormalizati (None, 31, 31, 64)   256         res2a_branch2a[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "activation_1 (Activation)       (None, 31, 31, 64)   0           bn2a_branch2a[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "res2a_branch2b (Conv2D)         (None, 31, 31, 64)   36928       activation_1[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "bn2a_branch2b (BatchNormalizati (None, 31, 31, 64)   256         res2a_branch2b[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "activation_2 (Activation)       (None, 31, 31, 64)   0           bn2a_branch2b[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "res2a_branch2c (Conv2D)         (None, 31, 31, 256)  16640       activation_2[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "res2a_branch1 (Conv2D)          (None, 31, 31, 256)  16640       max_pooling2d[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "bn2a_branch2c (BatchNormalizati (None, 31, 31, 256)  1024        res2a_branch2c[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "bn2a_branch1 (BatchNormalizatio (None, 31, 31, 256)  1024        res2a_branch1[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "add (Add)                       (None, 31, 31, 256)  0           bn2a_branch2c[0][0]              \n",
            "                                                                 bn2a_branch1[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "activation_3 (Activation)       (None, 31, 31, 256)  0           add[0][0]                        \n",
            "__________________________________________________________________________________________________\n",
            "res3a_branch2a (Conv2D)         (None, 16, 16, 128)  32896       activation_3[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "bn3a_branch2a (BatchNormalizati (None, 16, 16, 128)  512         res3a_branch2a[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "activation_4 (Activation)       (None, 16, 16, 128)  0           bn3a_branch2a[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "res3a_branch2b (Conv2D)         (None, 16, 16, 128)  147584      activation_4[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "bn3a_branch2b (BatchNormalizati (None, 16, 16, 128)  512         res3a_branch2b[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "activation_5 (Activation)       (None, 16, 16, 128)  0           bn3a_branch2b[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "res3a_branch2c (Conv2D)         (None, 16, 16, 512)  66048       activation_5[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "res3a_branch1 (Conv2D)          (None, 16, 16, 512)  131584      activation_3[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "bn3a_branch2c (BatchNormalizati (None, 16, 16, 512)  2048        res3a_branch2c[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "bn3a_branch1 (BatchNormalizatio (None, 16, 16, 512)  2048        res3a_branch1[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "add_1 (Add)                     (None, 16, 16, 512)  0           bn3a_branch2c[0][0]              \n",
            "                                                                 bn3a_branch1[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "activation_6 (Activation)       (None, 16, 16, 512)  0           add_1[0][0]                      \n",
            "__________________________________________________________________________________________________\n",
            "res4a_branch2a (Conv2D)         (None, 8, 8, 256)    131328      activation_6[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "bn4a_branch2a (BatchNormalizati (None, 8, 8, 256)    1024        res4a_branch2a[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "activation_7 (Activation)       (None, 8, 8, 256)    0           bn4a_branch2a[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "res4a_branch2b (Conv2D)         (None, 8, 8, 256)    590080      activation_7[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "bn4a_branch2b (BatchNormalizati (None, 8, 8, 256)    1024        res4a_branch2b[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "activation_8 (Activation)       (None, 8, 8, 256)    0           bn4a_branch2b[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "res4a_branch2c (Conv2D)         (None, 8, 8, 1024)   263168      activation_8[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "res4a_branch1 (Conv2D)          (None, 8, 8, 1024)   525312      activation_6[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "bn4a_branch2c (BatchNormalizati (None, 8, 8, 1024)   4096        res4a_branch2c[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "bn4a_branch1 (BatchNormalizatio (None, 8, 8, 1024)   4096        res4a_branch1[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "add_2 (Add)                     (None, 8, 8, 1024)   0           bn4a_branch2c[0][0]              \n",
            "                                                                 bn4a_branch1[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "activation_9 (Activation)       (None, 8, 8, 1024)   0           add_2[0][0]                      \n",
            "__________________________________________________________________________________________________\n",
            "res5a_branch2a (Conv2D)         (None, 4, 4, 512)    524800      activation_9[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "bn5a_branch2a (BatchNormalizati (None, 4, 4, 512)    2048        res5a_branch2a[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "activation_10 (Activation)      (None, 4, 4, 512)    0           bn5a_branch2a[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "res5a_branch2b (Conv2D)         (None, 4, 4, 512)    2359808     activation_10[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "bn5a_branch2b (BatchNormalizati (None, 4, 4, 512)    2048        res5a_branch2b[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "activation_11 (Activation)      (None, 4, 4, 512)    0           bn5a_branch2b[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "res5a_branch2c (Conv2D)         (None, 4, 4, 2048)   1050624     activation_11[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "res5a_branch1 (Conv2D)          (None, 4, 4, 2048)   2099200     activation_9[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "bn5a_branch2c (BatchNormalizati (None, 4, 4, 2048)   8192        res5a_branch2c[0][0]             \n",
            "__________________________________________________________________________________________________\n",
            "bn5a_branch1 (BatchNormalizatio (None, 4, 4, 2048)   8192        res5a_branch1[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "add_3 (Add)                     (None, 4, 4, 2048)   0           bn5a_branch2c[0][0]              \n",
            "                                                                 bn5a_branch1[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "activation_12 (Activation)      (None, 4, 4, 2048)   0           add_3[0][0]                      \n",
            "__________________________________________________________________________________________________\n",
            "average_pooling2d (AveragePooli (None, 2, 2, 2048)   0           activation_12[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "flatten (Flatten)               (None, 8192)         0           average_pooling2d[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "fc1 (Dense)                     (None, 1)            8193        flatten[0][0]                    \n",
            "==================================================================================================\n",
            "Total params: 8,046,849\n",
            "Trainable params: 8,027,521\n",
            "Non-trainable params: 19,328\n",
            "__________________________________________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vB-aiL9N8WOb",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "checkpoint_file = os.path.join(model_path+'/weights.{epoch:02d}-{val_loss:.2f}.hdf5')\n",
        "ckpt_callback   = keras.callbacks.ModelCheckpoint(filepath=checkpoint_file, monitor='val_loss')\n",
        "hists = []"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6_XoDyr28WOe",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def newest(path):\n",
        "    files = os.listdir(path)\n",
        "    paths = [os.path.join(path, basename) for basename in files]\n",
        "    try:\n",
        "        return max(paths, key=os.path.getctime)\n",
        "    except:\n",
        "        return None"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VxKKDzjQ8WOi",
        "colab_type": "code",
        "outputId": "a7317384-999e-44f7-a287-2cc291d28eb8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "latest_file=newest(model_path)\n",
        "print(latest_file)"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "None\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0t58Sc2N8WOl",
        "colab_type": "code",
        "outputId": "ba1f0956-1511-40bd-fd5c-089a108d7280",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 124
        }
      },
      "source": [
        "callbacks     = [ckpt_callback]\n",
        "if latest_file is None:\n",
        "    res_net_model.compile(optimizer=keras.optimizers.Adam(),loss='mse',metrics=['mse'])   \n",
        "else:\n",
        "    res_net_model = keras.models.load_model(latest_file)\n",
        "history = res_net_model.fit(train_dataset, epochs=3, batch_size=BATCH_SIZE,\n",
        "                    validation_data=test_dataset, verbose=1, \n",
        "                    steps_per_epoch=int(0.8*20000/BATCH_SIZE)-1,\n",
        "                    validation_steps=2)"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/3\n",
            "1999/1999 [==============================] - 207s 104ms/step - loss: 0.5923 - mse: 0.5923 - val_loss: 0.0910 - val_mse: 0.0910\n",
            "Epoch 2/3\n",
            "1999/1999 [==============================] - 169s 84ms/step - loss: 0.0842 - mse: 0.0842 - val_loss: 0.0697 - val_mse: 0.0697\n",
            "Epoch 3/3\n",
            "1999/1999 [==============================] - 188s 94ms/step - loss: 0.0534 - mse: 0.0534 - val_loss: 0.0585 - val_mse: 0.0585\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1LVumhpn8WO1",
        "colab_type": "code",
        "outputId": "0ec01ea8-6f29-460a-d910-fa982b0f9391",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        }
      },
      "source": [
        "plt.plot(history.history['loss'], marker='o')"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7fe1e238f748>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 23
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "j7KetQVq8WPE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "truth = list(test_dataset.as_numpy_iterator()) \n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rLDg_Lf5C2MX",
        "colab_type": "code",
        "outputId": "229831d9-f2c1-4e24-87bb-19a254132612",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "truth[0][0].shape"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(32, 128, 128, 1)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 59
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3R8zHvck8WPI",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "\n",
        "\n",
        "predictions = res_net_model.predict(x=truth[0][0])\n",
        "\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8RuYuDI4Cm3X",
        "colab_type": "code",
        "outputId": "2d87c5f8-ed82-4810-e44e-c10220a4281f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 53
        }
      },
      "source": [
        "print(predictions.shape)\n",
        "print(truth[1][1].shape)"
      ],
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(8, 1)\n",
            "(8,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bAdkGOalo7Hb",
        "colab_type": "code",
        "outputId": "ce52250e-e40c-42f3-88d8-bbaf9dd28f01",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 196
        }
      },
      "source": [
        "print(truth[1][1])\n",
        "print(predictions)\n"
      ],
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[1.0081496  0.5585804  0.6934614  0.9299813  0.8084125  0.5019638\n",
            " 0.8331738  0.30797407]\n",
            "[[0.61770606]\n",
            " [0.5937376 ]\n",
            " [0.6619957 ]\n",
            " [0.6215235 ]\n",
            " [0.6061509 ]\n",
            " [0.6184734 ]\n",
            " [0.66192716]\n",
            " [0.60767615]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Vkl7TtwJiRNN",
        "colab_type": "code",
        "outputId": "9e42d302-814f-4abc-93f3-6523a7b5a4b1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "np.mean((predictions-truth[1][1])**2)"
      ],
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.055967823"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 32
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xtaaC7Sp8WPN",
        "colab_type": "code",
        "outputId": "88a32682-5455-404f-dfc5-8674b05d19ae",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 285
        }
      },
      "source": [
        "plt.plot(truth[1][1],predictions, marker='o', ls='')\n",
        "plt.ylim(min(truth[1][1]),max(truth[1][1]))"
      ],
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.30797407031059265, 1.0081496238708496)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 33
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hx1g-lmO8WPh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "plt."
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}