Microaneurysms.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Microaneurysms.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyOBBHogLPN+znkbWJREeyf4"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "abO4bL3pmLiq",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from os import listdir\n",
        "from numpy import asarray\n",
        "from numpy import vstack\n",
        "from keras.preprocessing.image import img_to_array\n",
        "from keras.preprocessing.image import load_img\n",
        "from numpy import savez_compressed\n",
        "import tensorflow as tf"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6NDUxOQ28loY",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "ca6bc9d5-ba4b-4602-9cde-a3ee03d9e224"
      },
      "source": [
        "device = tf.test.gpu_device_name()\n",
        "if device != '/device:GPU:0':\n",
        "  raise SystemError('GPU device not found')\n",
        "print('Found GPU at: {}'.format(device))"
      ],
      "execution_count": 69,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found GPU at: /device:GPU:0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dNmSYvIYvDl3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import numpy"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IOJz-ercm6Te",
        "colab_type": "code",
        "outputId": "1246b4c0-9d4a-4e7d-9919-e2c0a9928e54",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": 67,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EENPouxEnEMV",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def load_images(path, size=(1024,1024)):\n",
        "\tsrc_list = list()\n",
        "\t# enumerate filenames in directory, assume all are images\n",
        "\tfor filename in listdir(path):\n",
        "\t\t# load and resize the image\n",
        "\t\tpixels = load_img(path + '/'+ filename, target_size=size)\n",
        "\t\t# convert to numpy array\n",
        "\t\tpixels = img_to_array(pixels)\n",
        "\t\t# split into satellite and map\n",
        "\t\t#sat_img = pixels[:, :256]\n",
        "\t\tsrc_list.append(pixels)\n",
        "\t\t#tar_list.append(map_img)\n",
        "\treturn asarray(src_list)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "42sPulhcnPda",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "path = '/content/drive/My Drive/Retinal Images/A. Segmentation/1. Original Images/a. Training Set'\n",
        "src_image = load_images(path)\n",
        "#tar_image = load_images('/content/drive/My Drive/Retinal Images/A. Segmentation/2. All Segmentation Groundtruths/a. Training Set/1. Microaneurysms')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Tcu0_C3ouj2o",
        "colab_type": "code",
        "outputId": "feb4313b-c438-43c7-92e6-e2ab444cf953",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "src = numpy.array(src_image)\n",
        "print('Loaded: ',src.shape)"
      ],
      "execution_count": 73,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Loaded:  (54, 1024, 1024, 3)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ur_lKkWZoZaE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "path = '/content/drive/My Drive/Retinal Images/A. Segmentation/2. All Segmentation Groundtruths/a. Training Set/1. Microaneurysms'\n",
        "tar_image = load_images(path)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KjSv8AEjpGVP",
        "colab_type": "code",
        "outputId": "7588d3db-ecc4-4191-92ba-18620cc013fb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "tar = numpy.array(tar_image)\n",
        "print(\"Loaded: \",tar.shape)"
      ],
      "execution_count": 75,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Loaded:  (54, 1024, 1024, 3)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sZJS4uxnpJnP",
        "colab_type": "code",
        "outputId": "9e31e92d-a766-4051-882f-019601368502",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "print('Loaded: ',src.shape, tar.shape)"
      ],
      "execution_count": 76,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Loaded:  (54, 1024, 1024, 3) (54, 1024, 1024, 3)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "idqxiR_Dw78N",
        "colab_type": "code",
        "outputId": "298610d3-1a66-4982-e080-83dbf8dc054d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "filename = 'microaneurysms.npz'\n",
        "savez_compressed(filename,src,tar)\n",
        "print('Saved dataset: ',filename)"
      ],
      "execution_count": 77,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Saved dataset:  microaneurysms.npz\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4vPrHId3pcK7",
        "colab_type": "code",
        "outputId": "f18ac452-a60e-428c-e95e-1db842f0c2b3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        }
      },
      "source": [
        "# load the prepared dataset\n",
        "from numpy import load\n",
        "from matplotlib import pyplot\n",
        "# load the dataset\n",
        "data = load('microaneurysms.npz')\n",
        "src_images, tar_images = data['arr_0'], data['arr_1']\n",
        "print('Loaded: ', src_images.shape, tar_images.shape)\n",
        "# plot source images\n",
        "n_samples = 3\n",
        "for i in range(n_samples):\n",
        "\tpyplot.subplot(2, n_samples, 1 + i)\n",
        "\tpyplot.axis('off')\n",
        "\tpyplot.imshow(src_images[i].astype('uint8'))\n",
        "# plot target image\n",
        "for i in range(n_samples):\n",
        "\tpyplot.subplot(2, n_samples, 1 + n_samples + i)\n",
        "\tpyplot.axis('off')\n",
        "\tpyplot.imshow(tar_images[i].astype('uint8'))\n",
        "pyplot.show()"
      ],
      "execution_count": 78,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Loaded:  (54, 1024, 1024, 3) (54, 1024, 1024, 3)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 6 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ey9JZ4lVxRrE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#Example of pix2pix gan for Original Image to Groundtruth image image-to-image translation\n",
        "from numpy import load\n",
        "from numpy import zeros\n",
        "from numpy import ones\n",
        "from numpy.random import randint\n",
        "from keras.optimizers import Adam\n",
        "from keras.initializers import RandomNormal\n",
        "from keras.models import Model\n",
        "from keras.models import Input\n",
        "from keras.layers import Conv2D\n",
        "from keras.layers import Conv2DTranspose\n",
        "from keras.layers import LeakyReLU\n",
        "from keras.layers import Activation\n",
        "from keras.layers import Concatenate\n",
        "from keras.layers import Dropout\n",
        "from keras.layers import BatchNormalization\n",
        "from keras.layers import LeakyReLU\n",
        "from matplotlib import pyplot"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Q8AFpWnPzI4o",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def define_discriminator(image_shape):\n",
        "\t# weight initialization\n",
        "\tinit = RandomNormal(stddev=0.02)\n",
        "\t# source image input\n",
        "\tin_src_image = Input(shape=image_shape)\n",
        "\t# target image input\n",
        "\tin_target_image = Input(shape=image_shape)\n",
        "\t# concatenate images channel-wise\n",
        "\tmerged = Concatenate()([in_src_image, in_target_image])\n",
        "\t# C64\n",
        "\td = Conv2D(64, (4,4), strides=(2,2), padding='same', kernel_initializer=init)(merged)\n",
        "\td = LeakyReLU(alpha=0.2)(d)\n",
        "\t# C128\n",
        "\td = Conv2D(128, (4,4), strides=(2,2), padding='same', kernel_initializer=init)(d)\n",
        "\td = BatchNormalization()(d)\n",
        "\td = LeakyReLU(alpha=0.2)(d)\n",
        "\t# C256\n",
        "\td = Conv2D(256, (4,4), strides=(2,2), padding='same', kernel_initializer=init)(d)\n",
        "\td = BatchNormalization()(d)\n",
        "\td = LeakyReLU(alpha=0.2)(d)\n",
        "\t# C512\n",
        "\td = Conv2D(512, (4,4), strides=(2,2), padding='same', kernel_initializer=init)(d)\n",
        "\td = BatchNormalization()(d)\n",
        "\td = LeakyReLU(alpha=0.2)(d)\n",
        "\t# second last output layer\n",
        "\td = Conv2D(512, (4,4), padding='same', kernel_initializer=init)(d)\n",
        "\td = BatchNormalization()(d)\n",
        "\td = LeakyReLU(alpha=0.2)(d)\n",
        "\t# patch output\n",
        "\td = Conv2D(1, (4,4), padding='same', kernel_initializer=init)(d)\n",
        "\tpatch_out = Activation('sigmoid')(d)\n",
        "\t# define model\n",
        "\tmodel = Model([in_src_image, in_target_image], patch_out)\n",
        "\t# compile model\n",
        "\topt = Adam(lr=0.0002, beta_1=0.5)\n",
        "\tmodel.compile(loss='binary_crossentropy', optimizer=opt, loss_weights=[0.5])\n",
        "\treturn model"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "G3VE4TDnzNVf",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def define_encoder_block(layer_in, n_filters, batchnorm=True):\n",
        "\t# weight initialization\n",
        "\tinit = RandomNormal(stddev=0.02)\n",
        "\t# add downsampling layer\n",
        "\tg = Conv2D(n_filters, (4,4), strides=(2,2), padding='same', kernel_initializer=init)(layer_in)\n",
        "\t# conditionally add batch normalization\n",
        "\tif batchnorm:\n",
        "\t\tg = BatchNormalization()(g, training=True)\n",
        "\t# leaky relu activation\n",
        "\tg = LeakyReLU(alpha=0.2)(g)\n",
        "\treturn g\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1bfu_OVfzQg8",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def decoder_block(layer_in, skip_in, n_filters, dropout=True):\n",
        "\t# weight initialization\n",
        "\tinit = RandomNormal(stddev=0.02)\n",
        "\t# add upsampling layer\n",
        "\tg = Conv2DTranspose(n_filters, (4,4), strides=(2,2), padding='same', kernel_initializer=init)(layer_in)\n",
        "\t# add batch normalization\n",
        "\tg = BatchNormalization()(g, training=True)\n",
        "\t# conditionally add dropout\n",
        "\tif dropout:\n",
        "\t\tg = Dropout(0.5)(g, training=True)\n",
        "\t# merge with skip connection\n",
        "\tg = Concatenate()([g, skip_in])\n",
        "\t# relu activation\n",
        "\tg = Activation('relu')(g)\n",
        "\treturn g"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mlB50OsOzS0x",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#U NET Model\n",
        "def define_generator(image_shape=(1024,1024,3)):\n",
        "\t# weight initialization\n",
        "\tinit = RandomNormal(stddev=0.02)\n",
        "\t# image input\n",
        "\tin_image = Input(shape=image_shape)\n",
        "\t# encoder model\n",
        "\te1 = define_encoder_block(in_image, 64, batchnorm=False)\n",
        "\te2 = define_encoder_block(e1, 128)\n",
        "\te3 = define_encoder_block(e2, 256)\n",
        "\te4 = define_encoder_block(e3, 512)\n",
        "\te5 = define_encoder_block(e4, 512)\n",
        "\te6 = define_encoder_block(e5, 512)\n",
        "\te7 = define_encoder_block(e6, 512)\n",
        "\t# bottleneck, no batch norm and relu\n",
        "\tb = Conv2D(512, (4,4), strides=(2,2), padding='same', kernel_initializer=init)(e7)\n",
        "\tb = Activation('relu')(b)\n",
        "\t# decoder model\n",
        "\td1 = decoder_block(b, e7, 512)\n",
        "\td2 = decoder_block(d1, e6, 512)\n",
        "\td3 = decoder_block(d2, e5, 512)\n",
        "\td4 = decoder_block(d3, e4, 512, dropout=False)\n",
        "\td5 = decoder_block(d4, e3, 256, dropout=False)\n",
        "\td6 = decoder_block(d5, e2, 128, dropout=False)\n",
        "\td7 = decoder_block(d6, e1, 64, dropout=False)\n",
        "\t# output\n",
        "\tg = Conv2DTranspose(3, (4,4), strides=(2,2), padding='same', kernel_initializer=init)(d7)\n",
        "\tout_image = Activation('tanh')(g)\n",
        "\t# define model\n",
        "\tmodel = Model(in_image, out_image)\n",
        "\treturn model"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8mKkhZGQzXpD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# define the combined generator and discriminator model, for updating the generator\n",
        "def define_gan(g_model, d_model, image_shape):\n",
        "\t# make weights in the discriminator not trainable\n",
        "\td_model.trainable = False\n",
        "\t# define the source image\n",
        "\tin_src = Input(shape=image_shape)\n",
        "\t# connect the source image to the generator input\n",
        "\tgen_out = g_model(in_src)\n",
        "\t# connect the source input and generator output to the discriminator input\n",
        "\tdis_out = d_model([in_src, gen_out])\n",
        "\t# src image as input, generated image and classification output\n",
        "\tmodel = Model(in_src, [dis_out, gen_out])\n",
        "\t# compile model\n",
        "\topt = Adam(lr=0.0002, beta_1=0.5)\n",
        "\tmodel.compile(loss=['binary_crossentropy', 'mae'], optimizer=opt, loss_weights=[1,100])\n",
        "\treturn model"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cYbxv8K5zhwf",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# load and prepare training images\n",
        "def load_real_samples(filename):\n",
        "\t# load compressed arrays\n",
        "\tdata = load(filename)\n",
        "\t# unpack arrays\n",
        "\tX1, X2 = data['arr_0'], data['arr_1']\n",
        "\t# scale from [0,255] to [-1,1]\n",
        "\tX1 = (X1 - 127.5) / 127.5\n",
        "\tX2 = (X2 - 127.5) / 127.5\n",
        "\treturn [X1, X2]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WmDDo3ZazkTZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# select a batch of random samples, returns images and target\n",
        "def generate_real_samples(dataset, n_samples, patch_shape):\n",
        "\t# unpack dataset\n",
        "\ttrainA, trainB = dataset\n",
        "\t# choose random instances\n",
        "\tix = randint(0, trainA.shape[0], n_samples)\n",
        "\t# retrieve selected images\n",
        "\tX1, X2 = trainA[ix], trainB[ix]\n",
        "\t# generate 'real' class labels (1)\n",
        "\ty = ones((n_samples, patch_shape, patch_shape, 1))\n",
        "\treturn [X1, X2], y"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Dr0qdPcTzm5p",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# generate a batch of images, returns images and targets\n",
        "def generate_fake_samples(g_model, samples, patch_shape):\n",
        "\t# generate fake instance\n",
        "\tX = g_model.predict(samples)\n",
        "\t# create 'fake' class labels (0)\n",
        "\ty = zeros((len(X), patch_shape, patch_shape, 1))\n",
        "\treturn X, y\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h79wKYNpzoqP",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# generate samples and save as a plot and save the model\n",
        "def summarize_performance(step, g_model, dataset, n_samples=3):\n",
        "\t# select a sample of input images\n",
        "\t[X_realA, X_realB], _ = generate_real_samples(dataset, n_samples, 1)\n",
        "\t# generate a batch of fake samples\n",
        "\tX_fakeB, _ = generate_fake_samples(g_model, X_realA, 1)\n",
        "\t# scale all pixels from [-1,1] to [0,1]\n",
        "\tX_realA = (X_realA + 1) / 2.0\n",
        "\tX_realB = (X_realB + 1) / 2.0\n",
        "\tX_fakeB = (X_fakeB + 1) / 2.0\n",
        "\t# plot real source images\n",
        "\tfor i in range(n_samples):\n",
        "\t\tpyplot.subplot(3, n_samples, 1 + i)\n",
        "\t\tpyplot.axis('off')\n",
        "\t\tpyplot.imshow(X_realA[i])\n",
        "\t# plot generated target image\n",
        "\tfor i in range(n_samples):\n",
        "\t\tpyplot.subplot(3, n_samples, 1 + n_samples + i)\n",
        "\t\tpyplot.axis('off')\n",
        "\t\tpyplot.imshow(X_fakeB[i])\n",
        "\t# plot real target image\n",
        "\tfor i in range(n_samples):\n",
        "\t\tpyplot.subplot(3, n_samples, 1 + n_samples*2 + i)\n",
        "\t\tpyplot.axis('off')\n",
        "\t\tpyplot.imshow(X_realB[i])\n",
        "\t# save plot to file\n",
        "\tfilename1 = 'plot_%06d.png' % (step+1)\n",
        "\tpyplot.savefig(filename1)\n",
        "\tpyplot.close()\n",
        "\t# save the generator model\n",
        "\tfilename2 = 'model_%06d.h5' % (step+1)\n",
        "\tg_model.save(filename2)\n",
        "\tprint('>Saved: %s and %s' % (filename1, filename2))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VDYF8be6zrr_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# train pix2pix models\n",
        "def train(d_model, g_model, gan_model, dataset, n_epochs=10, n_batch=1):\n",
        "\t# determine the output square shape of the discriminator\n",
        "\tn_patch = d_model.output_shape[1]\n",
        "\t# unpack dataset\n",
        "\ttrainA, trainB = dataset\n",
        "\t# calculate the number of batches per training epoch\n",
        "\tbat_per_epo = int(len(trainA) / n_batch)\n",
        "\t# calculate the number of training iterations\n",
        "\tn_steps = bat_per_epo * n_epochs\n",
        "\t# manually enumerate epochs\n",
        "\tfor i in range(n_steps):\n",
        "\t\t# select a batch of real samples\n",
        "\t\t[X_realA, X_realB], y_real = generate_real_samples(dataset, n_batch, n_patch)\n",
        "\t\t# generate a batch of fake samples\n",
        "\t\tX_fakeB, y_fake = generate_fake_samples(g_model, X_realA, n_patch)\n",
        "\t\t# update discriminator for real samples\n",
        "\t\td_loss1 = d_model.train_on_batch([X_realA, X_realB], y_real)\n",
        "\t\t# update discriminator for generated samples\n",
        "\t\td_loss2 = d_model.train_on_batch([X_realA, X_fakeB], y_fake)\n",
        "\t\t# update the generator\n",
        "\t\tg_loss, _, _ = gan_model.train_on_batch(X_realA, [y_real, X_realB])\n",
        "\t\t# summarize performance\n",
        "\t\tprint('>%d, d1[%.3f] d2[%.3f] g[%.3f]' % (i+1, d_loss1, d_loss2, g_loss))\n",
        "\t\t# summarize model performance\n",
        "\t\tif (i+1) % (bat_per_epo * 10) == 0:\n",
        "\t\t\tsummarize_performance(i, g_model, dataset)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "28CgoUAezvrp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "dataset = load_real_samples('microaneurysms.npz')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vwVPSRqRz72-",
        "colab_type": "code",
        "outputId": "3b9ddb9e-ddc0-45b7-94f6-25ee028fd710",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "print('Loaded', dataset[0].shape, dataset[1].shape)"
      ],
      "execution_count": 105,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Loaded (54, 1024, 1024, 3) (54, 1024, 1024, 3)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "liv57nP9z_j6",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "f640150d-4e98-48e7-dfcf-37429456eb1a"
      },
      "source": [
        "image_shape = dataset[0].shape[1:]\n",
        "print(image_shape)"
      ],
      "execution_count": 106,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(1024, 1024, 3)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-iJ0nPCc0Ckp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "d_model = define_discriminator(image_shape)\n",
        "g_model = define_generator(image_shape)\n",
        "# define the composite model\n",
        "gan_model = define_gan(g_model, d_model, image_shape)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oh7IBD-20GmU",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "c8ddf835-4b3c-44e3-e444-cc3aa6684952"
      },
      "source": [
        "train(d_model, g_model, gan_model, dataset)"
      ],
      "execution_count": 108,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/keras/engine/training.py:297: UserWarning: Discrepancy between trainable weights and collected trainable weights, did you set `model.trainable` without calling `model.compile` after ?\n",
            "  'Discrepancy between trainable weights and collected trainable'\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            ">1, d1[0.236] d2[2.047] g[101.701]\n",
            ">2, d1[0.260] d2[1.051] g[98.829]\n",
            ">3, d1[0.511] d2[0.908] g[95.216]\n",
            ">4, d1[0.381] d2[0.914] g[92.025]\n",
            ">5, d1[0.253] d2[0.609] g[88.981]\n",
            ">6, d1[0.261] d2[0.476] g[86.139]\n",
            ">7, d1[0.283] d2[0.377] g[83.790]\n",
            ">8, d1[0.291] d2[0.316] g[81.413]\n",
            ">9, d1[0.298] d2[0.299] g[78.866]\n",
            ">10, d1[0.287] d2[0.381] g[76.742]\n",
            ">11, d1[0.297] d2[0.514] g[74.257]\n",
            ">12, d1[0.338] d2[0.345] g[71.270]\n",
            ">13, d1[0.288] d2[0.329] g[70.157]\n",
            ">14, d1[0.267] d2[0.297] g[66.470]\n",
            ">15, d1[0.274] d2[0.248] g[64.297]\n",
            ">16, d1[0.213] d2[0.295] g[61.685]\n",
            ">17, d1[0.263] d2[0.471] g[58.844]\n",
            ">18, d1[0.253] d2[0.278] g[58.244]\n",
            ">19, d1[0.185] d2[0.220] g[54.516]\n",
            ">20, d1[0.206] d2[0.167] g[53.410]\n",
            ">21, d1[0.272] d2[0.284] g[50.119]\n",
            ">22, d1[0.321] d2[0.348] g[47.692]\n",
            ">23, d1[0.156] d2[0.266] g[45.270]\n",
            ">24, d1[0.192] d2[0.187] g[43.622]\n",
            ">25, d1[0.143] d2[0.095] g[42.513]\n",
            ">26, d1[0.095] d2[0.086] g[39.047]\n",
            ">27, d1[0.106] d2[0.197] g[35.496]\n",
            ">28, d1[0.464] d2[0.334] g[34.973]\n",
            ">29, d1[0.302] d2[0.213] g[33.512]\n",
            ">30, d1[0.097] d2[0.176] g[31.543]\n",
            ">31, d1[0.140] d2[0.217] g[31.048]\n",
            ">32, d1[0.092] d2[0.105] g[27.708]\n",
            ">33, d1[0.268] d2[0.183] g[26.359]\n",
            ">34, d1[0.654] d2[0.706] g[22.374]\n",
            ">35, d1[0.484] d2[0.317] g[25.926]\n",
            ">36, d1[0.365] d2[0.382] g[22.251]\n",
            ">37, d1[0.302] d2[0.338] g[21.340]\n",
            ">38, d1[0.254] d2[0.315] g[22.375]\n",
            ">39, d1[0.667] d2[0.297] g[18.003]\n",
            ">40, d1[0.177] d2[0.402] g[16.982]\n",
            ">41, d1[0.505] d2[0.260] g[16.554]\n",
            ">42, d1[0.183] d2[0.766] g[15.352]\n",
            ">43, d1[0.468] d2[0.175] g[14.411]\n",
            ">44, d1[0.455] d2[0.381] g[13.620]\n",
            ">45, d1[0.259] d2[0.339] g[17.571]\n",
            ">46, d1[0.510] d2[0.317] g[14.147]\n",
            ">47, d1[0.248] d2[0.380] g[13.735]\n",
            ">48, d1[0.508] d2[0.361] g[13.302]\n",
            ">49, d1[0.261] d2[0.402] g[13.205]\n",
            ">50, d1[0.497] d2[0.280] g[10.103]\n",
            ">51, d1[0.378] d2[0.521] g[9.997]\n",
            ">52, d1[0.352] d2[0.348] g[10.072]\n",
            ">53, d1[0.438] d2[0.332] g[9.235]\n",
            ">54, d1[0.337] d2[0.396] g[10.920]\n",
            ">55, d1[0.431] d2[0.363] g[9.151]\n",
            ">56, d1[0.306] d2[0.489] g[7.456]\n",
            ">57, d1[0.432] d2[0.287] g[7.560]\n",
            ">58, d1[0.442] d2[0.334] g[7.271]\n",
            ">59, d1[0.362] d2[0.402] g[6.903]\n",
            ">60, d1[0.500] d2[0.354] g[7.432]\n",
            ">61, d1[0.292] d2[0.469] g[5.942]\n",
            ">62, d1[0.425] d2[0.338] g[5.733]\n",
            ">63, d1[0.336] d2[0.404] g[6.414]\n",
            ">64, d1[0.384] d2[0.355] g[6.061]\n",
            ">65, d1[0.490] d2[0.336] g[6.278]\n",
            ">66, d1[0.328] d2[0.439] g[5.383]\n",
            ">67, d1[0.392] d2[0.398] g[4.658]\n",
            ">68, d1[0.428] d2[0.331] g[5.224]\n",
            ">69, d1[0.361] d2[0.402] g[4.412]\n",
            ">70, d1[0.410] d2[0.353] g[5.412]\n",
            ">71, d1[0.378] d2[0.364] g[6.610]\n",
            ">72, d1[0.344] d2[0.408] g[5.752]\n",
            ">73, d1[0.387] d2[0.371] g[4.022]\n",
            ">74, d1[0.388] d2[0.371] g[5.456]\n",
            ">75, d1[0.395] d2[0.366] g[3.920]\n",
            ">76, d1[0.322] d2[0.400] g[6.751]\n",
            ">77, d1[0.434] d2[0.327] g[4.200]\n",
            ">78, d1[0.368] d2[0.403] g[3.819]\n",
            ">79, d1[0.411] d2[0.384] g[3.208]\n",
            ">80, d1[0.413] d2[0.352] g[3.722]\n",
            ">81, d1[0.366] d2[0.381] g[3.663]\n",
            ">82, d1[0.366] d2[0.380] g[3.573]\n",
            ">83, d1[0.363] d2[0.385] g[3.294]\n",
            ">84, d1[0.363] d2[0.388] g[4.066]\n",
            ">85, d1[0.390] d2[0.369] g[3.282]\n",
            ">86, d1[0.409] d2[0.371] g[2.771]\n",
            ">87, d1[0.380] d2[0.374] g[2.923]\n",
            ">88, d1[0.362] d2[0.407] g[2.581]\n",
            ">89, d1[0.362] d2[0.387] g[2.792]\n",
            ">90, d1[0.403] d2[0.350] g[2.760]\n",
            ">91, d1[0.423] d2[0.359] g[2.346]\n",
            ">92, d1[0.409] d2[0.363] g[2.547]\n",
            ">93, d1[0.341] d2[0.427] g[1.914]\n",
            ">94, d1[0.361] d2[0.389] g[1.856]\n",
            ">95, d1[0.404] d2[0.345] g[3.332]\n",
            ">96, d1[0.367] d2[0.392] g[2.633]\n",
            ">97, d1[0.405] d2[0.355] g[3.320]\n",
            ">98, d1[0.374] d2[0.388] g[2.008]\n",
            ">99, d1[0.366] d2[0.379] g[2.955]\n",
            ">100, d1[0.395] d2[0.355] g[2.233]\n",
            ">101, d1[0.366] d2[0.394] g[1.685]\n",
            ">102, d1[0.373] d2[0.379] g[1.914]\n",
            ">103, d1[0.417] d2[0.348] g[2.095]\n",
            ">104, d1[0.380] d2[0.376] g[1.873]\n",
            ">105, d1[0.360] d2[0.394] g[1.555]\n",
            ">106, d1[0.401] d2[0.355] g[1.995]\n",
            ">107, d1[0.376] d2[0.377] g[2.372]\n",
            ">108, d1[0.369] d2[0.378] g[2.092]\n",
            ">109, d1[0.358] d2[0.388] g[1.983]\n",
            ">110, d1[0.372] d2[0.374] g[1.913]\n",
            ">111, d1[0.383] d2[0.367] g[2.581]\n",
            ">112, d1[0.346] d2[0.390] g[3.624]\n",
            ">113, d1[0.381] d2[0.364] g[1.866]\n",
            ">114, d1[0.377] d2[0.369] g[2.779]\n",
            ">115, d1[0.344] d2[0.412] g[1.924]\n",
            ">116, d1[0.410] d2[0.340] g[2.918]\n",
            ">117, d1[0.400] d2[0.360] g[1.662]\n",
            ">118, d1[0.369] d2[0.380] g[1.846]\n",
            ">119, d1[0.394] d2[0.362] g[1.518]\n",
            ">120, d1[0.385] d2[0.369] g[1.479]\n",
            ">121, d1[0.368] d2[0.385] g[1.414]\n",
            ">122, d1[0.368] d2[0.377] g[2.071]\n",
            ">123, d1[0.363] d2[0.387] g[1.525]\n",
            ">124, d1[0.368] d2[0.381] g[1.517]\n",
            ">125, d1[0.364] d2[0.387] g[1.323]\n",
            ">126, d1[0.371] d2[0.379] g[1.318]\n",
            ">127, d1[0.414] d2[0.340] g[1.719]\n",
            ">128, d1[0.397] d2[0.352] g[1.548]\n",
            ">129, d1[0.366] d2[0.380] g[1.534]\n",
            ">130, d1[0.378] d2[0.371] g[1.491]\n",
            ">131, d1[0.375] d2[0.381] g[1.243]\n",
            ">132, d1[0.378] d2[0.370] g[1.463]\n",
            ">133, d1[0.380] d2[0.371] g[1.928]\n",
            ">134, d1[0.351] d2[0.392] g[1.792]\n",
            ">135, d1[0.379] d2[0.360] g[1.578]\n",
            ">136, d1[0.376] d2[0.372] g[1.431]\n",
            ">137, d1[0.371] d2[0.372] g[1.430]\n",
            ">138, d1[0.386] d2[0.367] g[1.560]\n",
            ">139, d1[0.384] d2[0.355] g[1.425]\n",
            ">140, d1[0.379] d2[0.373] g[1.849]\n",
            ">141, d1[0.372] d2[0.380] g[1.240]\n",
            ">142, d1[0.373] d2[0.375] g[1.493]\n",
            ">143, d1[0.354] d2[0.397] g[1.125]\n",
            ">144, d1[0.375] d2[0.367] g[1.417]\n",
            ">145, d1[0.371] d2[0.380] g[1.064]\n",
            ">146, d1[0.406] d2[0.347] g[1.330]\n",
            ">147, d1[0.379] d2[0.369] g[1.293]\n",
            ">148, d1[0.364] d2[0.380] g[1.091]\n",
            ">149, d1[0.363] d2[0.374] g[1.336]\n",
            ">150, d1[0.381] d2[0.365] g[1.358]\n",
            ">151, d1[0.351] d2[0.399] g[1.328]\n",
            ">152, d1[0.377] d2[0.362] g[1.322]\n",
            ">153, d1[0.369] d2[0.379] g[1.391]\n",
            ">154, d1[0.384] d2[0.374] g[1.225]\n",
            ">155, d1[0.394] d2[0.363] g[1.261]\n",
            ">156, d1[0.376] d2[0.366] g[1.271]\n",
            ">157, d1[0.374] d2[0.366] g[1.038]\n",
            ">158, d1[0.361] d2[0.385] g[1.346]\n",
            ">159, d1[0.390] d2[0.373] g[1.043]\n",
            ">160, d1[0.368] d2[0.370] g[1.270]\n",
            ">161, d1[0.388] d2[0.358] g[1.056]\n",
            ">162, d1[0.391] d2[0.359] g[1.059]\n",
            ">163, d1[0.369] d2[0.385] g[1.006]\n",
            ">164, d1[0.346] d2[0.386] g[2.319]\n",
            ">165, d1[0.364] d2[0.363] g[1.263]\n",
            ">166, d1[0.389] d2[0.354] g[1.362]\n",
            ">167, d1[0.364] d2[0.365] g[1.059]\n",
            ">168, d1[0.379] d2[0.367] g[1.226]\n",
            ">169, d1[0.356] d2[0.375] g[1.232]\n",
            ">170, d1[0.385] d2[0.362] g[1.198]\n",
            ">171, d1[0.374] d2[0.366] g[1.162]\n",
            ">172, d1[0.351] d2[0.383] g[1.275]\n",
            ">173, d1[0.362] d2[0.364] g[1.261]\n",
            ">174, d1[0.383] d2[0.370] g[1.116]\n",
            ">175, d1[0.399] d2[0.371] g[1.008]\n",
            ">176, d1[0.375] d2[0.373] g[1.170]\n",
            ">177, d1[0.376] d2[0.376] g[0.972]\n",
            ">178, d1[0.381] d2[0.356] g[1.162]\n",
            ">179, d1[0.363] d2[0.368] g[1.869]\n",
            ">180, d1[0.370] d2[0.360] g[1.352]\n",
            ">181, d1[0.379] d2[0.364] g[1.140]\n",
            ">182, d1[0.357] d2[0.355] g[1.220]\n",
            ">183, d1[0.371] d2[0.364] g[1.113]\n",
            ">184, d1[0.338] d2[0.380] g[1.342]\n",
            ">185, d1[0.350] d2[0.370] g[1.299]\n",
            ">186, d1[0.399] d2[0.358] g[1.122]\n",
            ">187, d1[0.358] d2[0.352] g[1.205]\n",
            ">188, d1[0.399] d2[0.353] g[1.470]\n",
            ">189, d1[0.377] d2[0.369] g[1.037]\n",
            ">190, d1[0.364] d2[0.379] g[1.021]\n",
            ">191, d1[0.359] d2[0.366] g[1.294]\n",
            ">192, d1[0.346] d2[0.357] g[1.179]\n",
            ">193, d1[0.389] d2[0.368] g[1.038]\n",
            ">194, d1[0.360] d2[0.358] g[1.253]\n",
            ">195, d1[0.366] d2[0.388] g[0.991]\n",
            ">196, d1[0.362] d2[0.353] g[1.120]\n",
            ">197, d1[0.401] d2[0.355] g[1.006]\n",
            ">198, d1[0.374] d2[0.368] g[1.009]\n",
            ">199, d1[0.338] d2[0.373] g[1.130]\n",
            ">200, d1[0.342] d2[0.371] g[1.223]\n",
            ">201, d1[0.401] d2[0.358] g[0.983]\n",
            ">202, d1[0.371] d2[0.346] g[1.295]\n",
            ">203, d1[0.353] d2[0.340] g[1.609]\n",
            ">204, d1[0.370] d2[0.375] g[1.371]\n",
            ">205, d1[0.391] d2[0.363] g[0.991]\n",
            ">206, d1[0.329] d2[0.351] g[1.169]\n",
            ">207, d1[0.333] d2[0.346] g[1.830]\n",
            ">208, d1[0.364] d2[0.406] g[0.913]\n",
            ">209, d1[0.379] d2[0.338] g[1.145]\n",
            ">210, d1[0.387] d2[0.348] g[1.101]\n",
            ">211, d1[0.331] d2[0.323] g[1.822]\n",
            ">212, d1[0.431] d2[0.379] g[0.934]\n",
            ">213, d1[0.313] d2[0.345] g[1.245]\n",
            ">214, d1[0.328] d2[0.317] g[1.299]\n",
            ">215, d1[0.465] d2[0.380] g[1.236]\n",
            ">216, d1[0.404] d2[0.389] g[1.167]\n",
            ">217, d1[0.315] d2[0.385] g[0.966]\n",
            ">218, d1[0.346] d2[0.380] g[0.912]\n",
            ">219, d1[0.408] d2[0.350] g[1.005]\n",
            ">220, d1[0.361] d2[0.337] g[1.421]\n",
            ">221, d1[0.338] d2[0.338] g[1.267]\n",
            ">222, d1[0.363] d2[0.340] g[1.082]\n",
            ">223, d1[0.349] d2[0.318] g[0.987]\n",
            ">224, d1[0.356] d2[0.440] g[0.967]\n",
            ">225, d1[0.312] d2[0.326] g[1.003]\n",
            ">226, d1[0.367] d2[0.426] g[1.127]\n",
            ">227, d1[0.320] d2[0.341] g[1.213]\n",
            ">228, d1[0.360] d2[0.363] g[1.016]\n",
            ">229, d1[0.422] d2[0.317] g[1.365]\n",
            ">230, d1[0.368] d2[0.336] g[1.131]\n",
            ">231, d1[0.370] d2[0.324] g[1.138]\n",
            ">232, d1[0.271] d2[0.290] g[1.234]\n",
            ">233, d1[0.312] d2[0.338] g[0.971]\n",
            ">234, d1[0.304] d2[0.350] g[1.711]\n",
            ">235, d1[0.339] d2[0.243] g[1.331]\n",
            ">236, d1[0.342] d2[0.401] g[1.302]\n",
            ">237, d1[0.337] d2[0.251] g[1.315]\n",
            ">238, d1[0.281] d2[0.269] g[1.617]\n",
            ">239, d1[0.373] d2[0.421] g[1.395]\n",
            ">240, d1[0.275] d2[0.425] g[1.194]\n",
            ">241, d1[0.329] d2[0.288] g[1.600]\n",
            ">242, d1[0.478] d2[0.309] g[1.163]\n",
            ">243, d1[0.384] d2[0.343] g[1.071]\n",
            ">244, d1[0.345] d2[0.318] g[1.270]\n",
            ">245, d1[0.318] d2[0.293] g[1.272]\n",
            ">246, d1[0.210] d2[0.362] g[1.286]\n",
            ">247, d1[0.308] d2[0.341] g[1.499]\n",
            ">248, d1[0.361] d2[0.257] g[1.464]\n",
            ">249, d1[0.227] d2[0.211] g[1.623]\n",
            ">250, d1[0.351] d2[0.296] g[1.038]\n",
            ">251, d1[0.450] d2[0.672] g[0.916]\n",
            ">252, d1[0.226] d2[0.441] g[1.139]\n",
            ">253, d1[0.320] d2[0.323] g[1.215]\n",
            ">254, d1[0.308] d2[0.298] g[1.335]\n",
            ">255, d1[0.473] d2[0.333] g[1.361]\n",
            ">256, d1[0.258] d2[0.263] g[1.825]\n",
            ">257, d1[0.286] d2[0.212] g[1.756]\n",
            ">258, d1[0.235] d2[0.220] g[1.711]\n",
            ">259, d1[0.325] d2[0.228] g[1.755]\n",
            ">260, d1[0.267] d2[0.314] g[1.856]\n",
            ">261, d1[0.336] d2[0.169] g[1.797]\n",
            ">262, d1[0.163] d2[0.140] g[2.102]\n",
            ">263, d1[0.282] d2[0.302] g[1.818]\n",
            ">264, d1[0.479] d2[0.456] g[1.342]\n",
            ">265, d1[0.158] d2[0.336] g[1.870]\n",
            ">266, d1[0.218] d2[0.185] g[2.156]\n",
            ">267, d1[0.189] d2[0.104] g[2.491]\n",
            ">268, d1[0.166] d2[0.100] g[2.298]\n",
            ">269, d1[0.209] d2[0.146] g[2.205]\n",
            ">270, d1[0.133] d2[0.115] g[1.376]\n",
            ">271, d1[0.189] d2[0.434] g[2.585]\n",
            ">272, d1[0.155] d2[0.110] g[2.767]\n",
            ">273, d1[0.197] d2[0.089] g[3.278]\n",
            ">274, d1[0.112] d2[0.095] g[2.592]\n",
            ">275, d1[0.124] d2[0.176] g[2.675]\n",
            ">276, d1[0.191] d2[0.138] g[2.452]\n",
            ">277, d1[0.130] d2[0.122] g[2.732]\n",
            ">278, d1[0.122] d2[0.108] g[3.073]\n",
            ">279, d1[0.139] d2[0.054] g[3.182]\n",
            ">280, d1[0.059] d2[0.024] g[3.528]\n",
            ">281, d1[0.062] d2[0.040] g[3.400]\n",
            ">282, d1[0.039] d2[0.020] g[3.709]\n",
            ">283, d1[0.031] d2[0.036] g[3.503]\n",
            ">284, d1[0.386] d2[0.937] g[1.008]\n",
            ">285, d1[0.113] d2[0.884] g[2.078]\n",
            ">286, d1[0.393] d2[0.364] g[2.412]\n",
            ">287, d1[0.312] d2[0.232] g[2.332]\n",
            ">288, d1[0.274] d2[0.087] g[2.995]\n",
            ">289, d1[0.275] d2[0.077] g[3.134]\n",
            ">290, d1[0.147] d2[0.053] g[3.131]\n",
            ">291, d1[0.450] d2[0.825] g[1.559]\n",
            ">292, d1[0.256] d2[0.441] g[1.289]\n",
            ">293, d1[0.264] d2[0.277] g[1.418]\n",
            ">294, d1[0.300] d2[0.208] g[1.839]\n",
            ">295, d1[0.220] d2[0.169] g[1.997]\n",
            ">296, d1[0.465] d2[0.294] g[1.268]\n",
            ">297, d1[0.149] d2[0.216] g[1.850]\n",
            ">298, d1[0.129] d2[0.160] g[2.098]\n",
            ">299, d1[0.181] d2[0.184] g[2.199]\n",
            ">300, d1[0.160] d2[0.144] g[2.501]\n",
            ">301, d1[0.149] d2[0.072] g[2.734]\n",
            ">302, d1[0.092] d2[0.052] g[2.899]\n",
            ">303, d1[0.087] d2[0.054] g[3.140]\n",
            ">304, d1[0.060] d2[0.056] g[3.449]\n",
            ">305, d1[0.049] d2[0.063] g[3.584]\n",
            ">306, d1[0.110] d2[0.166] g[1.808]\n",
            ">307, d1[0.101] d2[0.133] g[2.925]\n",
            ">308, d1[0.098] d2[0.114] g[3.114]\n",
            ">309, d1[0.083] d2[0.045] g[3.375]\n",
            ">310, d1[0.326] d2[1.216] g[2.171]\n",
            ">311, d1[0.163] d2[0.298] g[2.115]\n",
            ">312, d1[0.175] d2[0.180] g[2.530]\n",
            ">313, d1[0.184] d2[0.191] g[2.291]\n",
            ">314, d1[0.171] d2[0.075] g[2.856]\n",
            ">315, d1[0.218] d2[0.112] g[2.318]\n",
            ">316, d1[0.099] d2[0.130] g[2.720]\n",
            ">317, d1[0.095] d2[0.054] g[3.340]\n",
            ">318, d1[0.237] d2[0.162] g[2.637]\n",
            ">319, d1[0.071] d2[0.106] g[3.709]\n",
            ">320, d1[0.119] d2[0.106] g[3.059]\n",
            ">321, d1[0.068] d2[0.083] g[3.412]\n",
            ">322, d1[0.063] d2[0.040] g[3.136]\n",
            ">323, d1[0.041] d2[0.039] g[3.723]\n",
            ">324, d1[0.054] d2[0.025] g[4.084]\n",
            ">325, d1[0.047] d2[0.015] g[4.236]\n",
            ">326, d1[0.037] d2[0.013] g[4.176]\n",
            ">327, d1[0.035] d2[0.028] g[4.229]\n",
            ">328, d1[0.030] d2[0.011] g[4.388]\n",
            ">329, d1[0.028] d2[0.010] g[4.269]\n",
            ">330, d1[0.045] d2[0.011] g[3.957]\n",
            ">331, d1[0.013] d2[0.012] g[4.260]\n",
            ">332, d1[0.025] d2[0.111] g[3.886]\n",
            ">333, d1[0.033] d2[0.014] g[4.499]\n",
            ">334, d1[0.062] d2[0.011] g[4.101]\n",
            ">335, d1[0.024] d2[0.053] g[3.781]\n",
            ">336, d1[0.049] d2[0.008] g[4.506]\n",
            ">337, d1[0.018] d2[0.026] g[3.810]\n",
            ">338, d1[0.017] d2[0.064] g[3.703]\n",
            ">339, d1[0.040] d2[0.033] g[4.437]\n",
            ">340, d1[0.044] d2[0.130] g[7.616]\n",
            ">341, d1[0.044] d2[0.211] g[3.646]\n",
            ">342, d1[0.190] d2[0.321] g[3.862]\n",
            ">343, d1[1.038] d2[0.166] g[3.475]\n",
            ">344, d1[0.054] d2[0.511] g[3.085]\n",
            ">345, d1[0.071] d2[0.067] g[3.989]\n",
            ">346, d1[0.117] d2[0.034] g[3.966]\n",
            ">347, d1[0.081] d2[0.019] g[3.851]\n",
            ">348, d1[0.038] d2[0.017] g[4.166]\n",
            ">349, d1[0.031] d2[0.088] g[3.228]\n",
            ">350, d1[0.044] d2[0.116] g[3.309]\n",
            ">351, d1[0.073] d2[0.176] g[3.109]\n",
            ">352, d1[0.104] d2[0.133] g[3.257]\n",
            ">353, d1[0.148] d2[0.040] g[3.555]\n",
            ">354, d1[0.296] d2[1.034] g[2.146]\n",
            ">355, d1[0.248] d2[0.121] g[2.655]\n",
            ">356, d1[0.470] d2[0.236] g[2.035]\n",
            ">357, d1[0.115] d2[0.120] g[2.585]\n",
            ">358, d1[0.091] d2[0.503] g[3.345]\n",
            ">359, d1[0.307] d2[0.817] g[1.547]\n",
            ">360, d1[0.183] d2[0.350] g[2.256]\n",
            ">361, d1[0.963] d2[0.210] g[1.524]\n",
            ">362, d1[0.247] d2[0.600] g[1.140]\n",
            ">363, d1[0.242] d2[0.396] g[1.680]\n",
            ">364, d1[0.256] d2[0.376] g[1.956]\n",
            ">365, d1[0.292] d2[0.182] g[2.131]\n",
            ">366, d1[0.333] d2[0.212] g[2.002]\n",
            ">367, d1[0.146] d2[0.384] g[2.513]\n",
            ">368, d1[0.200] d2[0.242] g[1.879]\n",
            ">369, d1[0.333] d2[0.481] g[1.423]\n",
            ">370, d1[0.378] d2[0.726] g[1.364]\n",
            ">371, d1[0.805] d2[0.537] g[0.934]\n",
            ">372, d1[0.660] d2[0.456] g[1.339]\n",
            ">373, d1[0.279] d2[0.499] g[1.349]\n",
            ">374, d1[0.341] d2[0.287] g[1.556]\n",
            ">375, d1[0.345] d2[0.242] g[1.541]\n",
            ">376, d1[0.452] d2[0.339] g[1.219]\n",
            ">377, d1[0.377] d2[0.354] g[1.468]\n",
            ">378, d1[0.300] d2[0.361] g[1.484]\n",
            ">379, d1[0.260] d2[0.372] g[1.433]\n",
            ">380, d1[0.510] d2[0.300] g[1.325]\n",
            ">381, d1[0.208] d2[0.315] g[1.517]\n",
            ">382, d1[0.446] d2[0.312] g[1.326]\n",
            ">383, d1[0.290] d2[0.287] g[1.417]\n",
            ">384, d1[0.352] d2[0.271] g[1.416]\n",
            ">385, d1[0.181] d2[0.419] g[1.575]\n",
            ">386, d1[0.237] d2[0.274] g[1.594]\n",
            ">387, d1[0.284] d2[0.407] g[1.304]\n",
            ">388, d1[0.245] d2[0.155] g[2.074]\n",
            ">389, d1[0.307] d2[0.410] g[1.801]\n",
            ">390, d1[0.222] d2[0.182] g[2.079]\n",
            ">391, d1[0.732] d2[0.348] g[1.475]\n",
            ">392, d1[0.162] d2[0.255] g[1.838]\n",
            ">393, d1[0.114] d2[0.194] g[2.324]\n",
            ">394, d1[0.598] d2[0.482] g[1.466]\n",
            ">395, d1[0.203] d2[0.314] g[1.801]\n",
            ">396, d1[0.308] d2[0.345] g[1.698]\n",
            ">397, d1[0.542] d2[0.523] g[1.252]\n",
            ">398, d1[0.399] d2[0.389] g[1.148]\n",
            ">399, d1[0.343] d2[0.355] g[1.210]\n",
            ">400, d1[0.561] d2[0.404] g[1.013]\n",
            ">401, d1[0.370] d2[0.399] g[1.198]\n",
            ">402, d1[0.477] d2[0.388] g[1.066]\n",
            ">403, d1[0.363] d2[0.360] g[1.096]\n",
            ">404, d1[0.311] d2[0.321] g[1.151]\n",
            ">405, d1[0.309] d2[0.328] g[1.188]\n",
            ">406, d1[0.280] d2[0.271] g[1.501]\n",
            ">407, d1[0.327] d2[0.296] g[1.594]\n",
            ">408, d1[0.309] d2[0.295] g[1.744]\n",
            ">409, d1[0.268] d2[0.221] g[1.557]\n",
            ">410, d1[0.255] d2[0.211] g[1.630]\n",
            ">411, d1[0.240] d2[0.210] g[1.838]\n",
            ">412, d1[0.142] d2[0.137] g[1.983]\n",
            ">413, d1[0.138] d2[0.128] g[2.195]\n",
            ">414, d1[0.068] d2[0.089] g[2.627]\n",
            ">415, d1[0.092] d2[0.075] g[2.531]\n",
            ">416, d1[0.070] d2[0.184] g[2.928]\n",
            ">417, d1[0.173] d2[0.041] g[3.146]\n",
            ">418, d1[0.290] d2[0.703] g[2.606]\n",
            ">419, d1[0.279] d2[0.087] g[2.370]\n",
            ">420, d1[0.175] d2[0.177] g[2.388]\n",
            ">421, d1[0.898] d2[0.589] g[1.709]\n",
            ">422, d1[0.209] d2[0.231] g[2.160]\n",
            ">423, d1[0.255] d2[0.210] g[2.335]\n",
            ">424, d1[0.352] d2[0.253] g[1.920]\n",
            ">425, d1[0.224] d2[0.266] g[1.991]\n",
            ">426, d1[0.134] d2[0.463] g[1.711]\n",
            ">427, d1[0.319] d2[0.315] g[1.523]\n",
            ">428, d1[0.485] d2[0.424] g[1.376]\n",
            ">429, d1[0.416] d2[0.313] g[1.118]\n",
            ">430, d1[0.289] d2[0.443] g[1.316]\n",
            ">431, d1[0.234] d2[0.367] g[1.506]\n",
            ">432, d1[0.336] d2[0.360] g[1.766]\n",
            ">433, d1[0.391] d2[0.824] g[1.087]\n",
            ">434, d1[0.327] d2[0.340] g[1.678]\n",
            ">435, d1[0.541] d2[0.385] g[1.351]\n",
            ">436, d1[0.323] d2[0.365] g[1.309]\n",
            ">437, d1[0.276] d2[0.242] g[1.722]\n",
            ">438, d1[0.392] d2[0.222] g[1.546]\n",
            ">439, d1[0.273] d2[0.289] g[1.633]\n",
            ">440, d1[0.226] d2[0.233] g[1.929]\n",
            ">441, d1[0.200] d2[0.242] g[2.003]\n",
            ">443, d1[0.249] d2[0.223] g[2.293]\n",
            ">444, d1[0.269] d2[0.293] g[2.254]\n",
            ">445, d1[0.298] d2[0.232] g[2.290]\n",
            ">446, d1[0.471] d2[0.273] g[1.968]\n",
            ">447, d1[0.197] d2[0.140] g[2.437]\n",
            ">448, d1[0.464] d2[0.761] g[1.723]\n",
            ">449, d1[0.394] d2[0.812] g[0.981]\n",
            ">450, d1[0.320] d2[0.317] g[1.548]\n",
            ">451, d1[0.632] d2[0.329] g[1.322]\n",
            ">452, d1[0.453] d2[0.315] g[1.409]\n",
            ">453, d1[0.257] d2[0.359] g[2.077]\n",
            ">454, d1[0.558] d2[0.295] g[1.314]\n",
            ">455, d1[0.203] d2[0.223] g[1.831]\n",
            ">456, d1[0.253] d2[0.180] g[1.756]\n",
            ">457, d1[0.167] d2[0.188] g[1.970]\n",
            ">458, d1[0.151] d2[0.175] g[2.279]\n",
            ">459, d1[0.233] d2[0.145] g[2.294]\n",
            ">460, d1[0.124] d2[0.131] g[2.445]\n",
            ">461, d1[0.189] d2[0.274] g[3.203]\n",
            ">462, d1[0.384] d2[0.262] g[2.143]\n",
            ">463, d1[0.150] d2[0.363] g[2.511]\n",
            ">464, d1[0.337] d2[0.410] g[1.908]\n",
            ">465, d1[0.316] d2[0.382] g[1.664]\n",
            ">466, d1[0.268] d2[0.492] g[1.586]\n",
            ">467, d1[0.516] d2[0.704] g[1.240]\n",
            ">468, d1[0.335] d2[0.413] g[1.726]\n",
            ">469, d1[0.475] d2[0.329] g[2.079]\n",
            ">470, d1[0.599] d2[0.345] g[1.361]\n",
            ">471, d1[0.321] d2[0.304] g[1.508]\n",
            ">472, d1[0.544] d2[0.321] g[1.326]\n",
            ">473, d1[0.180] d2[0.246] g[1.840]\n",
            ">474, d1[0.287] d2[0.183] g[1.739]\n",
            ">475, d1[0.265] d2[0.215] g[1.738]\n",
            ">476, d1[0.280] d2[0.263] g[1.815]\n",
            ">477, d1[0.170] d2[0.187] g[2.519]\n",
            ">478, d1[0.207] d2[0.138] g[2.105]\n",
            ">479, d1[0.110] d2[0.124] g[2.156]\n",
            ">480, d1[0.080] d2[0.087] g[2.465]\n",
            ">481, d1[0.080] d2[0.085] g[2.356]\n",
            ">482, d1[0.117] d2[0.054] g[2.754]\n",
            ">483, d1[0.093] d2[0.055] g[2.708]\n",
            ">484, d1[0.082] d2[0.055] g[2.894]\n",
            ">485, d1[0.048] d2[0.044] g[3.259]\n",
            ">486, d1[0.075] d2[0.035] g[3.121]\n",
            ">487, d1[0.025] d2[0.028] g[3.326]\n",
            ">488, d1[0.046] d2[0.028] g[3.378]\n",
            ">489, d1[0.063] d2[0.021] g[3.324]\n",
            ">490, d1[0.026] d2[0.026] g[3.497]\n",
            ">491, d1[0.024] d2[0.043] g[3.307]\n",
            ">492, d1[0.025] d2[0.099] g[3.361]\n",
            ">493, d1[0.130] d2[0.277] g[3.915]\n",
            ">494, d1[0.063] d2[0.115] g[3.948]\n",
            ">495, d1[0.287] d2[0.061] g[3.310]\n",
            ">496, d1[0.043] d2[0.063] g[3.076]\n",
            ">497, d1[0.052] d2[0.172] g[2.564]\n",
            ">498, d1[0.167] d2[0.679] g[1.588]\n",
            ">499, d1[0.199] d2[0.225] g[1.902]\n",
            ">500, d1[0.238] d2[0.199] g[1.784]\n",
            ">501, d1[0.161] d2[0.204] g[1.726]\n",
            ">502, d1[0.141] d2[0.259] g[1.673]\n",
            ">503, d1[0.424] d2[0.450] g[1.560]\n",
            ">504, d1[0.087] d2[0.605] g[1.418]\n",
            ">505, d1[0.222] d2[0.281] g[1.637]\n",
            ">506, d1[0.275] d2[0.197] g[1.734]\n",
            ">507, d1[0.170] d2[0.173] g[1.848]\n",
            ">508, d1[0.171] d2[0.192] g[1.847]\n",
            ">509, d1[0.149] d2[0.146] g[2.049]\n",
            ">510, d1[0.259] d2[0.186] g[2.405]\n",
            ">511, d1[0.180] d2[0.127] g[1.981]\n",
            ">512, d1[0.085] d2[0.151] g[2.302]\n",
            ">513, d1[0.139] d2[0.121] g[2.041]\n",
            ">514, d1[0.084] d2[0.113] g[2.202]\n",
            ">515, d1[0.091] d2[0.087] g[2.338]\n",
            ">516, d1[0.067] d2[0.097] g[2.400]\n",
            ">517, d1[0.151] d2[0.155] g[2.388]\n",
            ">518, d1[0.559] d2[0.497] g[1.535]\n",
            ">519, d1[0.052] d2[0.949] g[2.508]\n",
            ">520, d1[0.591] d2[0.428] g[1.795]\n",
            ">521, d1[0.512] d2[0.191] g[1.892]\n",
            ">522, d1[0.339] d2[0.436] g[1.391]\n",
            ">523, d1[0.222] d2[0.307] g[1.688]\n",
            ">524, d1[0.229] d2[0.247] g[2.324]\n",
            ">525, d1[0.257] d2[0.124] g[2.311]\n",
            ">526, d1[0.886] d2[0.485] g[1.354]\n",
            ">527, d1[0.102] d2[0.251] g[2.055]\n",
            ">528, d1[0.145] d2[0.129] g[2.722]\n",
            ">529, d1[0.182] d2[0.343] g[2.247]\n",
            ">530, d1[0.250] d2[0.160] g[2.668]\n",
            ">531, d1[0.140] d2[0.050] g[3.194]\n",
            ">532, d1[0.089] d2[0.049] g[3.423]\n",
            ">533, d1[0.066] d2[0.034] g[3.486]\n",
            ">534, d1[0.040] d2[0.046] g[3.758]\n",
            ">535, d1[0.071] d2[0.022] g[3.578]\n",
            ">536, d1[0.041] d2[0.019] g[3.626]\n",
            ">537, d1[0.027] d2[0.040] g[3.425]\n",
            ">538, d1[0.024] d2[0.144] g[2.510]\n",
            ">539, d1[0.038] d2[0.355] g[2.292]\n",
            ">540, d1[0.420] d2[0.172] g[2.194]\n",
            ">Saved: plot_000540.png and model_000540.h5\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5q7enDTG0JAG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# load and prepare training images\n",
        "def load_real_samples(filename):\n",
        "\t# load compressed ararys\n",
        "\tdata = load(filename)\n",
        "\t# unpack arrays\n",
        "\tX1, X2 = data['arr_0'], data['arr_1']\n",
        "\t# scale from [0,1024] to [-1,1]\n",
        "\tX1 = (X1 - 512) / 512\n",
        "\tX2 = (X2 - 512) / 512\n",
        "\treturn [X1, X2]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uK1pnHzf0LZj",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "1f1f308c-c4d3-459a-ed89-6d812c083b5f"
      },
      "source": [
        "# load dataset\n",
        "[X1, X2] = load_real_samples('microaneurysms.npz')\n",
        "print('Loaded', X1.shape, X2.shape)"
      ],
      "execution_count": 112,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Loaded (54, 1024, 1024, 3) (54, 1024, 1024, 3)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "upqbxP8sPjXD",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 50
        },
        "outputId": "724af151-a5be-43be-a903-7d71742d7000"
      },
      "source": [
        "# load model\n",
        "from keras import models\n",
        "model = models.load_model('model_000540.h5')"
      ],
      "execution_count": 114,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/keras/engine/saving.py:341: UserWarning: No training configuration found in save file: the model was *not* compiled. Compile it manually.\n",
            "  warnings.warn('No training configuration found in save file: '\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8KuiVPEVPz5u",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# select random example\n",
        "ix = randint(0, len(X1), 1)\n",
        "src_image, tar_image = X1[ix], X2[ix]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ss7DqRhiQE_2",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "gen_image = model.predict(src_image)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dVKBIK5EQHv0",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# plot source, generated and target images\n",
        "def plot_images(src_img, gen_img, tar_img):\n",
        "\timages = vstack((src_img, gen_img, tar_img))\n",
        "\t# scale from [-1,1] to [0,1]\n",
        "\timages = (images + 1) / 2.0\n",
        "\ttitles = ['Source', 'Generated', 'Expected']\n",
        "\t# plot images row by row\n",
        "\tfor i in range(len(images)):\n",
        "\t\t# define subplot\n",
        "\t\tpyplot.subplot(1, 3, 1 + i)\n",
        "\t\t# turn off axis\n",
        "\t\tpyplot.axis('off')\n",
        "\t\t# plot raw pixel data\n",
        "\t\tpyplot.imshow(images[i])\n",
        "\t\t# show title\n",
        "\t\tpyplot.title(titles[i])\n",
        "\tpyplot.show()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dr7ky-AHQRwF",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 145
        },
        "outputId": "79a78420-ccba-4885-d6a1-39d7d077cc09"
      },
      "source": [
        "# plot all three images\n",
        "plot_images(src_image, gen_image, tar_image)"
      ],
      "execution_count": 122,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 3 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "B6wiVdbcQUrz",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}