EDL_8_3_EvoAutoencoder.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "EDL_8_3_EvoAutoencoder.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyNlCVLhxWdrV5wmiweUdxoD",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/cxbxmxcx/EvolutionaryDeepLearning/blob/main/EDL_8_3_EvoAutoencoder.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SMGfjIfw7rmD"
      },
      "source": [
        "Imports and other setup"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install livelossplot --quiet\n",
        "!pip install deap --quiet"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GLvPw4XcYfVW",
        "outputId": "e041df51-6226-4bda-c857-74a055217124"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[?25l\r\u001b[K     |██                              | 10 kB 34.9 MB/s eta 0:00:01\r\u001b[K     |████                            | 20 kB 31.3 MB/s eta 0:00:01\r\u001b[K     |██████                          | 30 kB 19.8 MB/s eta 0:00:01\r\u001b[K     |████████▏                       | 40 kB 17.0 MB/s eta 0:00:01\r\u001b[K     |██████████▏                     | 51 kB 17.2 MB/s eta 0:00:01\r\u001b[K     |████████████▏                   | 61 kB 15.3 MB/s eta 0:00:01\r\u001b[K     |██████████████▎                 | 71 kB 15.3 MB/s eta 0:00:01\r\u001b[K     |████████████████▎               | 81 kB 16.6 MB/s eta 0:00:01\r\u001b[K     |██████████████████▎             | 92 kB 15.2 MB/s eta 0:00:01\r\u001b[K     |████████████████████▍           | 102 kB 13.9 MB/s eta 0:00:01\r\u001b[K     |██████████████████████▍         | 112 kB 13.9 MB/s eta 0:00:01\r\u001b[K     |████████████████████████▍       | 122 kB 13.9 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████▌     | 133 kB 13.9 MB/s eta 0:00:01\r\u001b[K     |████████████████████████████▌   | 143 kB 13.9 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████████▌ | 153 kB 13.9 MB/s eta 0:00:01\r\u001b[K     |████████████████████████████████| 160 kB 13.9 MB/s \n",
            "\u001b[?25h"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pDtDvs-_7qN1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "4eead7a3-0023-45fd-d202-06705c3c8904"
      },
      "source": [
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "import random\n",
        "\n",
        "from tensorflow.keras.datasets import mnist\n",
        "from tensorflow.keras import layers, models, Input, Model\n",
        "from tensorflow.keras.callbacks import EarlyStopping\n",
        "\n",
        "from IPython import display\n",
        "from IPython.display import clear_output\n",
        "import matplotlib.pyplot as plt\n",
        "from tensorflow.keras.utils import plot_model\n",
        "from livelossplot import PlotLosses\n",
        "plt.gray()\n",
        "\n",
        "#DEAP\n",
        "from deap import algorithms\n",
        "from deap import base\n",
        "from deap import benchmarks\n",
        "from deap import creator\n",
        "from deap import tools"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "M0RfbGcU8q6D"
      },
      "source": [
        "CONSTANTS"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ew02NVRk8gzm"
      },
      "source": [
        "Load Fashion Data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WVt0CcAj8jii",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2acbfecd-ae85-4e07-cdea-419029a5c639"
      },
      "source": [
        "# load dataset\n",
        "(train_images, train_labels), (test_images, test_labels) = tf.keras.datasets.fashion_mnist.load_data()\n",
        "\n",
        "# split dataset\n",
        "train_images = train_images.reshape(train_images.shape[0], 28, 28, 1).astype(\"float32\") / 255.0\n",
        "test_images = test_images.reshape(test_images.shape[0], 28, 28, 1).astype(\"float32\") / 255.0"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-labels-idx1-ubyte.gz\n",
            "32768/29515 [=================================] - 0s 0us/step\n",
            "40960/29515 [=========================================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-images-idx3-ubyte.gz\n",
            "26427392/26421880 [==============================] - 0s 0us/step\n",
            "26435584/26421880 [==============================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-labels-idx1-ubyte.gz\n",
            "16384/5148 [===============================================================================================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-images-idx3-ubyte.gz\n",
            "4423680/4422102 [==============================] - 0s 0us/step\n",
            "4431872/4422102 [==============================] - 0s 0us/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "G4a6JK-amsHi"
      },
      "source": [
        "Setup class names and labels for visualization, not training"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "m_tNwyn0mxJr"
      },
      "source": [
        "class_names = ['T-shirt/top', 'Trouser', 'Pullover', 'Dress', 'Coat',\n",
        "               'Sandal', 'Shirt', 'Sneaker', 'Bag', 'Ankle boot']"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zoUkmn5Z9XpW"
      },
      "source": [
        "Plot some images."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "n5hq3JD79aQv",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 589
        },
        "outputId": "80d2d02d-58a5-4339-ce82-d4d4b2f2f2ce"
      },
      "source": [
        "import math\n",
        "\n",
        "def plot_data(num_images, images, labels):\n",
        "  grid = math.ceil(math.sqrt(num_images))\n",
        "  plt.figure(figsize=(grid*2,grid*2))\n",
        "  for i in range(num_images):\n",
        "      plt.subplot(grid,grid,i+1)\n",
        "      plt.xticks([])\n",
        "      plt.yticks([])\n",
        "      plt.grid(False)     \n",
        "      plt.imshow(images[i].reshape(28,28))\n",
        "      plt.xlabel(class_names[labels[i]])      \n",
        "  plt.show()\n",
        "\n",
        "plot_data(25, train_images, train_labels)\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 25 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "N8odmpoCwtIO"
      },
      "source": [
        "STAGE 1: Auto-encoders"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hRQ6OzSNpLJx"
      },
      "source": [
        "Build the Encoder"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "efscot4HpN3i"
      },
      "source": [
        "# input layer\n",
        "input_layer = Input(shape=(28, 28, 1))\n",
        "\n",
        "# encoding architecture\n",
        "encoded_layer1 = layers.Conv2D(64, (3, 3), activation='relu', padding='same')(input_layer)\n",
        "encoded_layer1 = layers.MaxPool2D( (2, 2), padding='same')(encoded_layer1)\n",
        "encoded_layer2 = layers.Conv2D(32, (3, 3), activation='relu', padding='same')(encoded_layer1)\n",
        "encoded_layer2 = layers.MaxPool2D( (2, 2), padding='same')(encoded_layer2)\n",
        "encoded_layer3 = layers.Conv2D(16, (3, 3), activation='relu', padding='same')(encoded_layer2)\n",
        "latent_view    = layers.MaxPool2D( (2, 2), padding='same')(encoded_layer3)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "D4QFajN_qIKP"
      },
      "source": [
        "Build the Decoder"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PXuiAj9IqKbb"
      },
      "source": [
        "#decoding architecture\n",
        "decoded_layer1 = layers.Conv2D(16, (3, 3), activation='relu', padding='same')(latent_view)\n",
        "decoded_layer1 = layers.UpSampling2D((2, 2))(decoded_layer1)\n",
        "decoded_layer2 = layers.Conv2D(32, (3, 3), activation='relu', padding='same')(decoded_layer1)\n",
        "decoded_layer2 = layers.UpSampling2D((2, 2))(decoded_layer2)\n",
        "decoded_layer3 = layers.Conv2D(64, (3, 3), activation='relu')(decoded_layer2)\n",
        "decoded_layer3 = layers.UpSampling2D((2, 2))(decoded_layer3)\n",
        "#output layer\n",
        "output_layer   = layers.Conv2D(1, (3, 3), padding='same')(decoded_layer3)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ks0eEtdbqas0"
      },
      "source": [
        "Build the Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8eV6oxGlqdWp",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "319d3cf4-7f41-4b13-e236-170fde36fcf5"
      },
      "source": [
        "# compile the model\n",
        "model = Model(input_layer, output_layer)\n",
        "model.compile(optimizer='adam', loss='mse')\n",
        "model.summary()\n",
        "plot_model(model)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_1 (InputLayer)        [(None, 28, 28, 1)]       0         \n",
            "                                                                 \n",
            " conv2d (Conv2D)             (None, 28, 28, 64)        640       \n",
            "                                                                 \n",
            " max_pooling2d (MaxPooling2D  (None, 14, 14, 64)       0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " conv2d_1 (Conv2D)           (None, 14, 14, 32)        18464     \n",
            "                                                                 \n",
            " max_pooling2d_1 (MaxPooling  (None, 7, 7, 32)         0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_2 (Conv2D)           (None, 7, 7, 16)          4624      \n",
            "                                                                 \n",
            " max_pooling2d_2 (MaxPooling  (None, 4, 4, 16)         0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_3 (Conv2D)           (None, 4, 4, 16)          2320      \n",
            "                                                                 \n",
            " up_sampling2d (UpSampling2D  (None, 8, 8, 16)         0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " conv2d_4 (Conv2D)           (None, 8, 8, 32)          4640      \n",
            "                                                                 \n",
            " up_sampling2d_1 (UpSampling  (None, 16, 16, 32)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_5 (Conv2D)           (None, 14, 14, 64)        18496     \n",
            "                                                                 \n",
            " up_sampling2d_2 (UpSampling  (None, 28, 28, 64)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_6 (Conv2D)           (None, 28, 28, 1)         577       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 49,761\n",
            "Trainable params: 49,761\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "history_loss = []\n",
        "history_val_loss = []\n",
        "\n",
        "def add_history(history):\n",
        "  history_loss.append(history.history[\"loss\"])\n",
        "  history_val_loss.append(history.history[\"val_loss\"])\n",
        "\n",
        "def reset_history():\n",
        "  global history_loss\n",
        "  global history_val_loss\n",
        "  history_loss = []\n",
        "  history_val_loss = []\n",
        "  return []\n",
        "\n",
        "def plot_results(num_images, images, labels, history):\n",
        "  add_history(history)\n",
        "  grid = math.ceil(math.sqrt(num_images))\n",
        "  plt.figure(figsize=(grid*2,grid*2))\n",
        "  for i in range(num_images):\n",
        "      plt.subplot(grid,grid,i+1)\n",
        "      plt.xticks([])\n",
        "      plt.yticks([])\n",
        "      plt.grid(False)     \n",
        "      plt.imshow(images[i].reshape(28,28))\n",
        "      plt.xlabel(class_names[labels[i]])      \n",
        "  plt.show()\n",
        "  plt.plot(history_loss, label='loss')\n",
        "  plt.plot(history_val_loss, label='val_loss')\n",
        "  plt.legend()\n",
        "  plt.show()  "
      ],
      "metadata": {
        "id": "0NDsxe9LnQX-"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "EPOCHS = 3\n",
        "history = reset_history()\n",
        "\n",
        "for i in range(EPOCHS):\n",
        "  history = model.fit(train_images, train_images, epochs=1, batch_size=2048, validation_data=(test_images, test_images))\n",
        "  pred_images = model.predict(test_images[:25])\n",
        "  clear_output()\n",
        "  plot_results(25, pred_images[:25], test_labels[:25], history)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 837
        },
        "id": "Km0n4knukEc6",
        "outputId": "51f633e3-e8fe-46b5-cd52-6080ade5f60c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 25 Axes>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#@title Constants\n",
        "max_layers = 10\n",
        "max_neurons = 128\n",
        "min_neurons = 16\n",
        "max_kernel = 3\n",
        "min_kernel = 3\n",
        "max_pool = 2\n",
        "min_pool = 2\n",
        "\n",
        "CONV_LAYER = -1\n",
        "CONV_LAYER_LEN = 4\n",
        "BN_LAYER = -3\n",
        "BN_LAYER_LEN = 1\n",
        "DROPOUT_LAYER = -4\n",
        "DROPOUT_LAYER_LEN = 2\n",
        "UPCONV_LAYER = -2\n",
        "UPCONV_LAYER_LEN = 4"
      ],
      "metadata": {
        "id": "klAM5onsvSx4"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#@title Encoding scheme\n",
        "def generate_neurons():\n",
        "  return random.randint(min_neurons, max_neurons)\n",
        "\n",
        "def generate_kernel():\n",
        "  part = []\n",
        "  part.append(random.randint(min_kernel, max_kernel))\n",
        "  part.append(random.randint(min_kernel, max_kernel))\n",
        "  return part\n",
        "\n",
        "def generate_bn_layer():\n",
        "  part = [BN_LAYER] \n",
        "  return part\n",
        "\n",
        "def generate_dropout_layer():\n",
        "  part = [DROPOUT_LAYER] \n",
        "  part.append(random.uniform(0,.5))  \n",
        "  return part\n",
        "\n",
        "def generate_conv_layer():\n",
        "  part = [CONV_LAYER] \n",
        "  part.append(generate_neurons())\n",
        "  part.extend(generate_kernel())  \n",
        "  return part\n",
        "\n",
        "def generate_upconv_layer():\n",
        "  part = [UPCONV_LAYER] \n",
        "  part.append(generate_neurons())\n",
        "  part.extend(generate_kernel())  \n",
        "  return part\n",
        "\n",
        "def create_offspring():\n",
        "  ind = []\n",
        "  layers = 0\n",
        "  for i in range(max_layers):\n",
        "    if i==0: #first layer always convolutational\n",
        "      ind.extend(generate_conv_layer())      \n",
        "      layers += 1\n",
        "    elif random.uniform(0,1)<.5:\n",
        "      #add convolution layer\n",
        "      ind.extend(generate_conv_layer())\n",
        "      layers += 1\n",
        "      if random.uniform(0,1)<.5:\n",
        "        #add batchnormalization\n",
        "        ind.extend(generate_bn_layer())\n",
        "      if random.uniform(0,1) < .5:\n",
        "        ind.extend(generate_dropout_layer()) \n",
        "  for i in range(layers):\n",
        "    ind.extend(generate_upconv_layer())\n",
        "    if random.uniform(0,1)<.5:\n",
        "      #add batchnormalization\n",
        "      ind.extend(generate_bn_layer())\n",
        "    if random.uniform(0,1) < .5:\n",
        "      ind.extend(generate_dropout_layer())\n",
        "  return ind\n",
        "        \n",
        "individual = create_offspring()\n",
        "print(individual)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZRlGpaMYvZK8",
        "outputId": "0f92b64a-ab30-46f9-ff9f-dd82e03f1d37"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[-1, 36, 3, 3, -1, 87, 3, 3, -3, -4, 0.3412580264132386, -1, 39, 3, 3, -4, 0.0158711234430432, -1, 128, 3, 3, -4, 0.12518939732131767, -1, 42, 3, 3, -3, -2, 111, 3, 3, -2, 62, 3, 3, -3, -4, 0.4405968326159026, -2, 122, 3, 3, -2, 81, 3, 3, -2, 121, 3, 3]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def padding(gene):\n",
        "  return \"same\" if gene == 1 else \"valid\"\n",
        "\n",
        "def build_model(individual):\n",
        "  input_layer = Input(shape=(28, 28, 1))  \n",
        "  il = len(individual)\n",
        "  i = 0\n",
        "  x = input_layer\n",
        "  while i < il:    \n",
        "    if individual[i] == CONV_LAYER:      \n",
        "      pad=\"same\" \n",
        "      n = individual[i+1]\n",
        "      k = (individual[i+2], individual[i+3])\n",
        "      i += CONV_LAYER_LEN        \n",
        "      x = layers.Conv2D(n, k, activation='relu', padding=pad)(x) \n",
        "      if x.shape[1] > 7:\n",
        "        x = layers.MaxPool2D( (2, 2), padding='same')(x)\n",
        "    elif individual[i] == BN_LAYER: #add batchnormal layer\n",
        "      x = layers.BatchNormalization()(x)\n",
        "      i += BN_LAYER_LEN      \n",
        "    elif individual[i] == DROPOUT_LAYER: #add dropout layer            \n",
        "      x = layers.Dropout(individual[i+1])(x) \n",
        "      i += DROPOUT_LAYER_LEN\n",
        "    elif individual[i] == UPCONV_LAYER:\n",
        "      pad=\"same\"\n",
        "      n = individual[i+1]\n",
        "      k = (individual[i+2], individual[i+3])        \n",
        "      x = layers.Conv2D(n, k, activation='relu', padding=pad)(x)   \n",
        "      x = layers.UpSampling2D((2, 2))(x)   \n",
        "      i += CONV_LAYER_LEN   \n",
        "      if x.shape[1] == (28):\n",
        "        break #model is complete\n",
        "    else:\n",
        "      break\n",
        "  if x.shape[1] == 14:\n",
        "    x = layers.UpSampling2D((2, 2))(x)\n",
        "    \n",
        "  output_layer = layers.Conv2D(1, (3, 3), padding='same')(x)\n",
        "  model = Model(input_layer, output_layer)\n",
        "  model.compile(optimizer='adam', loss='mse')\n",
        "  return model\n",
        "\n",
        "model = build_model(individual) \n",
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BLU1Y1650PEe",
        "outputId": "619cc411-f6ce-4694-a2e4-43ab92144e32"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_1\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_2 (InputLayer)        [(None, 28, 28, 1)]       0         \n",
            "                                                                 \n",
            " conv2d_7 (Conv2D)           (None, 28, 28, 36)        360       \n",
            "                                                                 \n",
            " max_pooling2d_3 (MaxPooling  (None, 14, 14, 36)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_8 (Conv2D)           (None, 14, 14, 87)        28275     \n",
            "                                                                 \n",
            " max_pooling2d_4 (MaxPooling  (None, 7, 7, 87)         0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " batch_normalization (BatchN  (None, 7, 7, 87)         348       \n",
            " ormalization)                                                   \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 7, 7, 87)          0         \n",
            "                                                                 \n",
            " conv2d_9 (Conv2D)           (None, 7, 7, 39)          30576     \n",
            "                                                                 \n",
            " dropout_1 (Dropout)         (None, 7, 7, 39)          0         \n",
            "                                                                 \n",
            " conv2d_10 (Conv2D)          (None, 7, 7, 128)         45056     \n",
            "                                                                 \n",
            " dropout_2 (Dropout)         (None, 7, 7, 128)         0         \n",
            "                                                                 \n",
            " conv2d_11 (Conv2D)          (None, 7, 7, 42)          48426     \n",
            "                                                                 \n",
            " batch_normalization_1 (Batc  (None, 7, 7, 42)         168       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " conv2d_12 (Conv2D)          (None, 7, 7, 111)         42069     \n",
            "                                                                 \n",
            " up_sampling2d_3 (UpSampling  (None, 14, 14, 111)      0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_13 (Conv2D)          (None, 14, 14, 62)        62000     \n",
            "                                                                 \n",
            " up_sampling2d_4 (UpSampling  (None, 28, 28, 62)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_14 (Conv2D)          (None, 28, 28, 1)         559       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 257,837\n",
            "Trainable params: 257,579\n",
            "Non-trainable params: 258\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "individual = create_offspring() \n",
        "model = build_model(individual)\n",
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8MTbYKpYBBTz",
        "outputId": "66040d9b-a5f7-4edb-d69d-29d35d5082ac"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_2\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_3 (InputLayer)        [(None, 28, 28, 1)]       0         \n",
            "                                                                 \n",
            " conv2d_15 (Conv2D)          (None, 28, 28, 92)        920       \n",
            "                                                                 \n",
            " max_pooling2d_5 (MaxPooling  (None, 14, 14, 92)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_16 (Conv2D)          (None, 14, 14, 30)        24870     \n",
            "                                                                 \n",
            " max_pooling2d_6 (MaxPooling  (None, 7, 7, 30)         0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_17 (Conv2D)          (None, 7, 7, 38)          10298     \n",
            "                                                                 \n",
            " batch_normalization_2 (Batc  (None, 7, 7, 38)         152       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " dropout_3 (Dropout)         (None, 7, 7, 38)          0         \n",
            "                                                                 \n",
            " conv2d_18 (Conv2D)          (None, 7, 7, 30)          10290     \n",
            "                                                                 \n",
            " batch_normalization_3 (Batc  (None, 7, 7, 30)         120       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " conv2d_19 (Conv2D)          (None, 7, 7, 42)          11382     \n",
            "                                                                 \n",
            " conv2d_20 (Conv2D)          (None, 7, 7, 22)          8338      \n",
            "                                                                 \n",
            " up_sampling2d_5 (UpSampling  (None, 14, 14, 22)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " dropout_4 (Dropout)         (None, 14, 14, 22)        0         \n",
            "                                                                 \n",
            " conv2d_21 (Conv2D)          (None, 14, 14, 26)        5174      \n",
            "                                                                 \n",
            " up_sampling2d_6 (UpSampling  (None, 28, 28, 26)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_22 (Conv2D)          (None, 28, 28, 1)         235       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 71,779\n",
            "Trainable params: 71,643\n",
            "Non-trainable params: 136\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Creating Mate/Mutation Operators"
      ],
      "metadata": {
        "id": "yIQD2pN7fID0"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#@title Start of Mate/Mutation Operators\n",
        "def get_layers(ind, layer_type):\n",
        "  return [a for a in range(len(ind)) if ind[a] == layer_type]\n",
        "\n",
        "def swap(ind1, iv1, ind2, iv2, ll):\n",
        "  ch1 = ind1[iv1:iv1+ll]\n",
        "  ch2 = ind2[iv2:iv2+ll] \n",
        "  ind1[iv1:iv1+ll] = ch2\n",
        "  ind2[iv2:iv2+ll] = ch1\n",
        "  return ind1, ind2\n",
        "\n",
        "def swap_layers(ind1, ind2, layer_type, layer_len):\n",
        "  c1, c2 = get_layers(ind1, layer_type), get_layers(ind2, layer_type) \n",
        "  min_c = min(len(c1), len(c2))\n",
        "  for i in range(min_c):\n",
        "    if random.random() < 1:\n",
        "      i1 = random.randint(0, len(c1)-1)\n",
        "      i2 = random.randint(0, len(c2)-1)      \n",
        "      iv1 = c1.pop(i1)\n",
        "      iv2 = c2.pop(i2)    \n",
        "      ind1, ind2 = swap(ind1, iv1, ind2, iv2, layer_len) \n",
        "  return ind1, ind2 \n",
        "\n",
        "def crossover(ind1, ind2):  \n",
        "  ind1, ind2 = swap_layers(ind1, ind2, CONV_LAYER, CONV_LAYER_LEN)\n",
        "  ind1, ind2 = swap_layers(ind1, ind2, UPCONV_LAYER, UPCONV_LAYER_LEN)\n",
        "  ind1, ind2 = swap_layers(ind1, ind2, BN_LAYER, BN_LAYER_LEN)\n",
        "  ind1, ind2 = swap_layers(ind1, ind2, DROPOUT_LAYER, DROPOUT_LAYER_LEN)\n",
        "  return ind1, ind2  \n",
        "\n",
        "ind1 = create_offspring()\n",
        "ind2 = create_offspring()\n",
        "print(ind1)\n",
        "print(ind2)\n",
        "\n",
        "ind1, ind2 = crossover(ind1, ind2)\n",
        "print(ind1)\n",
        "print(ind2)\n",
        "\n",
        "model = build_model(ind1)\n",
        "model.summary()\n",
        "model = build_model(ind2)\n",
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VkiLVueJWBdx",
        "outputId": "08a19cb7-7f0e-488c-a8c7-747746f2aef5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[-1, 20, 3, 3, -1, 39, 3, 3, -3, -1, 124, 3, 3, -1, 102, 3, 3, -3, -2, 29, 3, 3, -3, -2, 73, 3, 3, -4, 0.03315006221009481, -2, 128, 3, 3, -3, -4, 0.06349085416824679, -2, 44, 3, 3, -4, 0.24023849961836757]\n",
            "[-1, 97, 3, 3, -1, 32, 3, 3, -3, -2, 60, 3, 3, -3, -4, 0.3982435397232512, -2, 106, 3, 3, -3]\n",
            "[-1, 97, 3, 3, -1, 39, 3, 3, -3, -1, 124, 3, 3, -1, 32, 3, 3, -3, -2, 60, 3, 3, -3, -2, 73, 3, 3, -4, 0.03315006221009481, -2, 106, 3, 3, -3, -4, 0.06349085416824679, -2, 44, 3, 3, -4, 0.3982435397232512]\n",
            "[-1, 20, 3, 3, -1, 102, 3, 3, -3, -2, 29, 3, 3, -3, -4, 0.24023849961836757, -2, 128, 3, 3, -3]\n",
            "Model: \"model_3\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_4 (InputLayer)        [(None, 28, 28, 1)]       0         \n",
            "                                                                 \n",
            " conv2d_23 (Conv2D)          (None, 28, 28, 97)        970       \n",
            "                                                                 \n",
            " max_pooling2d_7 (MaxPooling  (None, 14, 14, 97)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_24 (Conv2D)          (None, 14, 14, 39)        34086     \n",
            "                                                                 \n",
            " max_pooling2d_8 (MaxPooling  (None, 7, 7, 39)         0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " batch_normalization_4 (Batc  (None, 7, 7, 39)         156       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " conv2d_25 (Conv2D)          (None, 7, 7, 124)         43648     \n",
            "                                                                 \n",
            " conv2d_26 (Conv2D)          (None, 7, 7, 32)          35744     \n",
            "                                                                 \n",
            " batch_normalization_5 (Batc  (None, 7, 7, 32)         128       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " conv2d_27 (Conv2D)          (None, 7, 7, 60)          17340     \n",
            "                                                                 \n",
            " up_sampling2d_7 (UpSampling  (None, 14, 14, 60)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " batch_normalization_6 (Batc  (None, 14, 14, 60)       240       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " conv2d_28 (Conv2D)          (None, 14, 14, 73)        39493     \n",
            "                                                                 \n",
            " up_sampling2d_8 (UpSampling  (None, 28, 28, 73)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_29 (Conv2D)          (None, 28, 28, 1)         658       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 172,463\n",
            "Trainable params: 172,201\n",
            "Non-trainable params: 262\n",
            "_________________________________________________________________\n",
            "Model: \"model_4\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_5 (InputLayer)        [(None, 28, 28, 1)]       0         \n",
            "                                                                 \n",
            " conv2d_30 (Conv2D)          (None, 28, 28, 20)        200       \n",
            "                                                                 \n",
            " max_pooling2d_9 (MaxPooling  (None, 14, 14, 20)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_31 (Conv2D)          (None, 14, 14, 102)       18462     \n",
            "                                                                 \n",
            " max_pooling2d_10 (MaxPoolin  (None, 7, 7, 102)        0         \n",
            " g2D)                                                            \n",
            "                                                                 \n",
            " batch_normalization_7 (Batc  (None, 7, 7, 102)        408       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " conv2d_32 (Conv2D)          (None, 7, 7, 29)          26651     \n",
            "                                                                 \n",
            " up_sampling2d_9 (UpSampling  (None, 14, 14, 29)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " batch_normalization_8 (Batc  (None, 14, 14, 29)       116       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " dropout_5 (Dropout)         (None, 14, 14, 29)        0         \n",
            "                                                                 \n",
            " conv2d_33 (Conv2D)          (None, 14, 14, 128)       33536     \n",
            "                                                                 \n",
            " up_sampling2d_10 (UpSamplin  (None, 28, 28, 128)      0         \n",
            " g2D)                                                            \n",
            "                                                                 \n",
            " conv2d_34 (Conv2D)          (None, 28, 28, 1)         1153      \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 80,526\n",
            "Trainable params: 80,264\n",
            "Non-trainable params: 262\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#@title Mutation\n",
        "def mutate(part, layer_type):\n",
        "  if layer_type == CONV_LAYER and len(part)==CONV_LAYER_LEN:\n",
        "    part[1] = int(part[1] * random.uniform(.9, 1.1))\n",
        "    part[2] = random.randint(min_kernel, max_kernel)\n",
        "    part[3] = random.randint(min_kernel, max_kernel)\n",
        "  elif layer_type == UPCONV_LAYER and len(part)==UPCONV_LAYER_LEN:\n",
        "    part[1] = random.randint(min_kernel, max_kernel)\n",
        "    part[2] = random.randint(min_kernel, max_kernel)\n",
        "  elif layer_type == DROPOUT_LAYER and len(part)==DROPOUT_LAYER_LEN:\n",
        "    part[1] = random.uniform(0, .5) \n",
        "  else:\n",
        "    error = f\"mutate ERROR {part}\"    \n",
        "    raise Exception(error) \n",
        "  return part\n",
        "\n",
        "def mutate_layers(ind, layer_type, layer_len):\n",
        "  layers = get_layers(ind1, layer_type)\n",
        "  for layer in layers:\n",
        "    if random.random() < 1:\n",
        "      try:\n",
        "        ind[layer:layer+layer_len] = mutate(\n",
        "            ind[layer:layer+layer_len], layer_type) \n",
        "      except:\n",
        "        print(layers)\n",
        "  return ind      \n",
        "\n",
        "print(ind1)\n",
        "\n",
        "def mutation(ind):    \n",
        "  ind = mutate_layers(ind, CONV_LAYER, CONV_LAYER_LEN)\n",
        "  ind = mutate_layers(ind, DROPOUT_LAYER, DROPOUT_LAYER_LEN)    \n",
        "  ind = mutate_layers(ind, UPCONV_LAYER, UPCONV_LAYER_LEN)\n",
        "  return ind,\n",
        "\n",
        "ind, = mutation(ind1)\n",
        "print(ind)\n",
        "model = build_model(ind1)\n",
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sImATzc4Wjp8",
        "outputId": "02809e52-5460-462b-b3c2-bbd7507594bf"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[-1, 97, 3, 3, -1, 39, 3, 3, -3, -1, 124, 3, 3, -1, 32, 3, 3, -3, -2, 60, 3, 3, -3, -2, 73, 3, 3, -4, 0.03315006221009481, -2, 106, 3, 3, -3, -4, 0.06349085416824679, -2, 44, 3, 3, -4, 0.3982435397232512]\n",
            "[-1, 100, 3, 3, -1, 39, 3, 3, -3, -1, 113, 3, 3, -1, 34, 3, 3, -3, -2, 3, 3, 3, -3, -2, 3, 3, 3, -4, 0.10881617530032173, -2, 3, 3, 3, -3, -4, 0.03324032570163665, -2, 3, 3, 3, -4, 0.45543010327726324]\n",
            "Model: \"model_5\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_6 (InputLayer)        [(None, 28, 28, 1)]       0         \n",
            "                                                                 \n",
            " conv2d_35 (Conv2D)          (None, 28, 28, 100)       1000      \n",
            "                                                                 \n",
            " max_pooling2d_11 (MaxPoolin  (None, 14, 14, 100)      0         \n",
            " g2D)                                                            \n",
            "                                                                 \n",
            " conv2d_36 (Conv2D)          (None, 14, 14, 39)        35139     \n",
            "                                                                 \n",
            " max_pooling2d_12 (MaxPoolin  (None, 7, 7, 39)         0         \n",
            " g2D)                                                            \n",
            "                                                                 \n",
            " batch_normalization_9 (Batc  (None, 7, 7, 39)         156       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " conv2d_37 (Conv2D)          (None, 7, 7, 113)         39776     \n",
            "                                                                 \n",
            " conv2d_38 (Conv2D)          (None, 7, 7, 34)          34612     \n",
            "                                                                 \n",
            " batch_normalization_10 (Bat  (None, 7, 7, 34)         136       \n",
            " chNormalization)                                                \n",
            "                                                                 \n",
            " conv2d_39 (Conv2D)          (None, 7, 7, 3)           921       \n",
            "                                                                 \n",
            " up_sampling2d_11 (UpSamplin  (None, 14, 14, 3)        0         \n",
            " g2D)                                                            \n",
            "                                                                 \n",
            " batch_normalization_11 (Bat  (None, 14, 14, 3)        12        \n",
            " chNormalization)                                                \n",
            "                                                                 \n",
            " conv2d_40 (Conv2D)          (None, 14, 14, 3)         84        \n",
            "                                                                 \n",
            " up_sampling2d_12 (UpSamplin  (None, 28, 28, 3)        0         \n",
            " g2D)                                                            \n",
            "                                                                 \n",
            " conv2d_41 (Conv2D)          (None, 28, 28, 1)         28        \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 111,864\n",
            "Trainable params: 111,712\n",
            "Non-trainable params: 152\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#@title Setting up the Creator\n",
        "creator.create(\"FitnessMax\", base.Fitness, weights=(-1.0,))\n",
        "creator.create(\"Individual\", list, fitness=creator.FitnessMax)"
      ],
      "metadata": {
        "id": "P37ruhkcYy7x"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#@title Create Individual and Population\n",
        "toolbox = base.Toolbox()\n",
        "toolbox.register(\"autoencoder\", create_offspring)\n",
        "toolbox.register(\"individual\", tools.initIterate, creator.Individual, toolbox.autoencoder)\n",
        "toolbox.register(\"population\", tools.initRepeat, list, toolbox.individual)\n",
        "\n",
        "toolbox.register(\"select\", tools.selTournament, tournsize=5)"
      ],
      "metadata": {
        "id": "-rWTwio4Y7tK"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#@title Register Crossover and Mutation\n",
        "toolbox.register(\"mate\", crossover)\n",
        "toolbox.register(\"mutate\", mutation)"
      ],
      "metadata": {
        "id": "1YXkl6VGZBHE"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#@title Register Evaluation\n",
        "def clamp(num, min_value, max_value):\n",
        "   return max(min(num, max_value), min_value)\n",
        "\n",
        "def train(model):  \n",
        "  model.fit(train_images, train_images, epochs=3,                     \n",
        "                    verbose=0)\n",
        "  return model\n",
        "\n",
        "max_fitness = 0\n",
        "\n",
        "def evaluate(individual):  \n",
        "  try:\n",
        "    model = build_model(individual)\n",
        "    model = train(model)    \n",
        "    eval = model.evaluate(test_images, test_images, verbose=0)\n",
        "    print(\".\", end='')\n",
        "    \n",
        "    return clamp(eval, 0, 1000),\n",
        "  except:\n",
        "    return 1000,  \n",
        "\n",
        "toolbox.register(\"evaluate\", evaluate)  "
      ],
      "metadata": {
        "id": "znaZtRebZKxJ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#@title Optimize the Weights { run: \"auto\" }\n",
        "MU = 5 #@param {type:\"slider\", min:5, max:1000, step:5}\n",
        "NGEN = 5 #@param {type:\"slider\", min:5, max:100, step:1}\n",
        "RGEN = 1 #@param {type:\"slider\", min:1, max:100, step:1}\n",
        "CXPB = .6\n",
        "MUTPB = .3\n",
        "\n",
        "random.seed(64)\n",
        "\n",
        "pop = toolbox.population(n=MU)\n",
        "hof = tools.HallOfFame(1)\n",
        "stats = tools.Statistics(lambda ind: ind.fitness.values)\n",
        "stats.register(\"avg\", np.mean)\n",
        "stats.register(\"std\", np.std)\n",
        "stats.register(\"min\", np.min)\n",
        "stats.register(\"max\", np.max)"
      ],
      "metadata": {
        "id": "tkfBghYPZlSp"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "best = None\n",
        "groups = { \"fitness\" : {\"min\", \"max\"}}\n",
        "plotlosses = PlotLosses(groups=groups)\n",
        "\n",
        "for g in range(NGEN):\n",
        "  pop, logbook = algorithms.eaSimple(pop, toolbox, \n",
        "            cxpb=CXPB, mutpb=MUTPB, ngen=RGEN, stats=stats, halloffame=hof, verbose=False)\n",
        "  best = hof[0] \n",
        "  \n",
        "  print(f\"Gen ({(g+1)*RGEN})\")      \n",
        "  for l in logbook:\n",
        "    plotlosses.update({'min': l[\"min\"], 'max': l[\"max\"]})\n",
        "  plotlosses.send()  # draw, update logs, etc"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 637
        },
        "id": "-YeaIXkIZqGp",
        "outputId": "e9fb6d59-d82a-406c-a400-476c9be17d2f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x576 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "fitness\n",
            "\tmax              \t (min:    0.007, max: 1000.000, cur: 1000.000)\n",
            "\tmin              \t (min:    0.006, max:    0.007, cur:    0.006)\n"
          ]
        }
      ]
    }
  ]
}