EDL_7_4_Mutation_CNN.ipynb

{
  "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_7_4_Mutation_CNN.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "FG2xwCBkZT7O"
      },
      "outputs": [],
      "source": [
        "#@title Install Packages\n",
        "!pip install livelossplot --quiet"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "id": "FFxv9NclYSvb"
      },
      "outputs": [],
      "source": [
        "#@title Imports\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras import datasets, layers, models\n",
        "import numpy as np\n",
        "import math\n",
        "import time\n",
        "import random\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "from livelossplot import PlotLossesKeras"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "id": "8d0CpCJaYXk8",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "70d79f64-2375-4c96-cdb0-381638f9c797"
      },
      "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"
          ]
        }
      ],
      "source": [
        "#@title Load Data\n",
        "dataset = datasets.fashion_mnist\n",
        "(x_train, y_train), (x_test, y_test) = dataset.load_data()\n",
        "\n",
        "# normalize and reshape data\n",
        "x_train = x_train.reshape(x_train.shape[0], 28, 28, 1).astype(\"float32\") / 255.0\n",
        "x_test = x_test.reshape(x_test.shape[0], 28, 28, 1).astype(\"float32\") / 255.0\n",
        "\n",
        "x_train = x_train[:1000]\n",
        "y_train= y_train[:1000]\n",
        "x_test = x_test[:100]\n",
        "y_test= y_test[:100]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "mUZD3klmed57"
      },
      "outputs": [],
      "source": [
        "class_names = ['T-shirt/top', 'Trouser', 'Pullover', 'Dress', 'Coat',\n",
        "               'Sandal', 'Shirt', 'Sneaker', 'Bag', 'Ankle boot']"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 589
        },
        "id": "_2WS35U8YkJQ",
        "outputId": "9c633916-0e40-43cc-fc4e-2677b70eca66"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 25 Axes>"
            ]
          },
          "metadata": {}
        }
      ],
      "source": [
        "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, x_train, y_train)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#@title Constants\n",
        "max_layers = 5\n",
        "max_neurons = 128\n",
        "min_neurons = 16\n",
        "max_kernel = 5\n",
        "min_kernel = 2\n",
        "max_pool = 3\n",
        "min_pool = 2\n",
        "\n",
        "CONV_LAYER = -1\n",
        "CONV_LAYER_LEN = 4\n",
        "POOLING_LAYER = -2\n",
        "POOLING_LAYER_LEN = 3\n",
        "BN_LAYER = -3\n",
        "BN_LAYER_LEN = 1\n",
        "DENSE_LAYER = -4\n",
        "DENSE_LAYER_LEN = 2"
      ],
      "metadata": {
        "id": "2oLtk5CwZ2Dg"
      },
      "execution_count": 6,
      "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_pooling_layer():\n",
        "  part = [POOLING_LAYER] \n",
        "  part.append(random.randint(min_pool, max_pool))\n",
        "  part.append(random.randint(min_pool, max_pool))\n",
        "  return part\n",
        "\n",
        "def generate_dense_layer():\n",
        "  part = [DENSE_LAYER] \n",
        "  part.append(generate_neurons())  \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 create_offspring():\n",
        "  ind = []\n",
        "  for i in range(max_layers):\n",
        "    if random.uniform(0,1)<.5:\n",
        "      #add convolution layer\n",
        "      ind.extend(generate_conv_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",
        "        #add max pooling layer\n",
        "        ind.extend(generate_pooling_layer())\n",
        "  ind.extend(generate_dense_layer())\n",
        "  return ind\n",
        "        \n",
        "individual = create_offspring()\n",
        "print(individual)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6pArO3IdXeH8",
        "outputId": "9c09e72d-7386-43ef-e640-0542fddbc781"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[-1, 56, 5, 5, -1, 109, 4, 5, -4, 71]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def build_model(individual):\n",
        "  model = models.Sequential()\n",
        "  il = len(individual)\n",
        "  i = 0\n",
        "  while i < il:\n",
        "    if individual[i] == CONV_LAYER: \n",
        "      n = individual[i+1]\n",
        "      k = (individual[i+2], individual[i+3])\n",
        "      i += CONV_LAYER_LEN\n",
        "      if i == 0: #first layer, add input shape      \n",
        "        model.add(layers.Conv2D(n, k, activation='relu', padding=\"same\", input_shape=(28, 28, 1)))      \n",
        "      else:\n",
        "        model.add(layers.Conv2D(n, k, activation='relu', padding=\"same\"))    \n",
        "    elif individual[i] == POOLING_LAYER: #add pooling layer\n",
        "      k = k = (individual[i+1], individual[i+2])\n",
        "      i += POOLING_LAYER_LEN\n",
        "      model.add(layers.MaxPooling2D(k, padding=\"same\"))      \n",
        "    elif individual[i] == BN_LAYER: #add batchnormal layer\n",
        "      model.add(layers.BatchNormalization())\n",
        "      i += 1      \n",
        "    elif individual[i] == DENSE_LAYER: #add dense layer\n",
        "      model.add(layers.Flatten())      \n",
        "      model.add(layers.Dense(individual[i+1], activation='relu'))\n",
        "      i += 2\n",
        "  model.add(layers.Dense(10))\n",
        "  return model\n",
        "\n",
        "model = build_model(individual) \n"
      ],
      "metadata": {
        "id": "NUlDrjKqedTA"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "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",
        "  print(ll, iv1, ch1, iv2, ch2)\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, POOLING_LAYER, POOLING_LAYER_LEN)\n",
        "  ind1, ind2 = swap_layers(ind1, ind2, BN_LAYER, BN_LAYER_LEN)\n",
        "  ind1, ind2 = swap_layers(ind1, ind2, DENSE_LAYER, DENSE_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)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6lqH6fMqlYJO",
        "outputId": "849422c7-0571-40da-9974-ab7f526e8196"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[-1, 111, 4, 5, -1, 98, 4, 3, -3, -1, 92, 2, 5, -2, 2, 2, -1, 68, 2, 3, -3, -2, 2, 3, -4, 32]\n",
            "[-1, 93, 2, 2, -2, 3, 3, -1, 25, 3, 4, -2, 2, 2, -4, 100]\n",
            "4 4 [-1, 98, 4, 3] 7 [-1, 25, 3, 4]\n",
            "4 9 [-1, 92, 2, 5] 0 [-1, 93, 2, 2]\n",
            "3 13 [-2, 2, 2] 4 [-2, 3, 3]\n",
            "3 21 [-2, 2, 3] 11 [-2, 2, 2]\n",
            "2 24 [-4, 32] 14 [-4, 100]\n",
            "[-1, 111, 4, 5, -1, 25, 3, 4, -3, -1, 93, 2, 2, -2, 3, 3, -1, 68, 2, 3, -3, -2, 2, 2, -4, 100]\n",
            "[-1, 92, 2, 5, -2, 2, 2, -1, 98, 4, 3, -2, 2, 3, -4, 32]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Custom Mutation Operator"
      ],
      "metadata": {
        "id": "WuTnQ07Rh4l5"
      }
    },
    {
      "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 == POOLING_LAYER and len(part)==POOLING_LAYER_LEN:\n",
        "    part[1] = random.randint(min_kernel, max_kernel)\n",
        "    part[2] = random.randint(min_kernel, max_kernel)\n",
        "  elif layer_type == DENSE_LAYER and len(part)==DENSE_LAYER_LEN:\n",
        "    part[1] = int(part[1] * random.uniform(.9, 1.1)) \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",
        "def mutation(ind):  \n",
        "  if len(ind) > CONV_LAYER_LEN: #only mutate conv individuals\n",
        "    ind = mutate_layers(ind, CONV_LAYER, CONV_LAYER_LEN)\n",
        "    ind = mutate_layers(ind, DENSE_LAYER, DENSE_LAYER_LEN)\n",
        "    ind = mutate_layers(ind, POOLING_LAYER, POOLING_LAYER_LEN)\n",
        "  return ind,\n",
        "\n",
        "print(ind1)\n",
        "ind1 = mutation(ind1)[0]\n",
        "print(ind1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vn3-FyMQ2VHY",
        "outputId": "e08ceefb-6300-4a3b-dba1-25e6ebbb4aa0"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[-1, 116, 3, 3, -1, 22, 5, 3, -3, -1, 85, 5, 5, -2, 5, 3, -1, 58, 4, 3, -3, -2, 3, 2, -4, 92]\n",
            "[-1, 127, 3, 5, -1, 20, 3, 4, -3, -1, 93, 4, 2, -2, 5, 4, -1, 58, 5, 4, -3, -2, 2, 5, -4, 94]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model = build_model(ind1) \n",
        "\n",
        "model.compile(optimizer='adam',\n",
        "              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "history = model.fit(x_train, y_train, epochs=3, \n",
        "                    validation_data=(x_test, y_test),\n",
        "                    callbacks=[PlotLossesKeras()],\n",
        "                    verbose=0)\n",
        "\n",
        "model.summary()\n",
        "model.evaluate(x_test, y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "Ca7sBT1ZhyOK",
        "outputId": "eeaf775c-8cbb-4d65-f452-77f5096ddc6b"
      },
      "execution_count": 12,
      "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": [
            "accuracy\n",
            "\ttraining         \t (min:    0.651, max:    0.860, cur:    0.860)\n",
            "\tvalidation       \t (min:    0.380, max:    0.570, cur:    0.380)\n",
            "Loss\n",
            "\ttraining         \t (min:    0.414, max:    1.076, cur:    0.414)\n",
            "\tvalidation       \t (min:    2.038, max:    2.085, cur:    2.085)\n",
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " conv2d_2 (Conv2D)           (None, 28, 28, 111)       2331      \n",
            "                                                                 \n",
            " conv2d_3 (Conv2D)           (None, 28, 28, 25)        33325     \n",
            "                                                                 \n",
            " batch_normalization (BatchN  (None, 28, 28, 25)       100       \n",
            " ormalization)                                                   \n",
            "                                                                 \n",
            " conv2d_4 (Conv2D)           (None, 28, 28, 93)        9393      \n",
            "                                                                 \n",
            " max_pooling2d (MaxPooling2D  (None, 10, 10, 93)       0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " conv2d_5 (Conv2D)           (None, 10, 10, 68)        38012     \n",
            "                                                                 \n",
            " batch_normalization_1 (Batc  (None, 10, 10, 68)       272       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " max_pooling2d_1 (MaxPooling  (None, 5, 5, 68)         0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " flatten_1 (Flatten)         (None, 1700)              0         \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 100)               170100    \n",
            "                                                                 \n",
            " dense_3 (Dense)             (None, 10)                1010      \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 254,543\n",
            "Trainable params: 254,357\n",
            "Non-trainable params: 186\n",
            "_________________________________________________________________\n",
            "4/4 [==============================] - 0s 7ms/step - loss: 2.0853 - accuracy: 0.3800\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[2.0853333473205566, 0.3799999952316284]"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "name": "EDL_7_4_Mutation_CNN.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyMltt163nNMscsXbt2WVSLM",
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}