Feature_extraction.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "toc_visible": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZFVhzoMRRHDI",
        "outputId": "7d57dad7-117e-44b2-e8bd-4a4957c9f148"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "2.9.2\n"
          ]
        }
      ],
      "source": [
        "import tensorflow as tf\n",
        "print(tf.__version__)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Feature Extraction Analysis\n",
        "\n",
        "Convolutional Neural Networks (CNN) are a type of deep learning architecture specialized in image-related tasks. One of the main benefits from this type of algorithms is the great performance at extracting relevant information from an image in order to classify it. We call \"relevant information\" the features of the image. Thus, for short, feature extraction is a property from CNNs that highlights meaninful pixels from an image, and discard pixels that don't convey information, or contain redundancies."
      ],
      "metadata": {
        "id": "uOjLIQB7kvwq"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 1. Objective\n",
        "\n",
        "In this notebook I will try to explore CNN as feature extractors, and I'll investigate how specialization is capable of enhancing feature extraction. More specifically, I will look at the level of specialization of two models with the same architecture, one trained with 10% of the samples in relation to the other, and see qualitatively how they perform."
      ],
      "metadata": {
        "id": "ZuN61_lkSAWT"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8pytlOw9RqBP",
        "outputId": "527095a8-f1cc-4215-be3f-35eb8440143a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import pandas as pd"
      ],
      "metadata": {
        "id": "pIvzJYrXRvq4"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "saving_direc = '/content/drive/MyDrive/Neural Networks/CNN Feature Extraction/'"
      ],
      "metadata": {
        "id": "cTCgZUw2pOZK"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 2. Selecting and absorbing a dataset\n",
        "\n",
        "The chosen dataset is the Cifar10, a dataset available at Tensorflow that has 50K images for train and 10K images for test, and consists of 32x32 pixels, RGB images distributed in 10 categories. The categories are uniformly distributed in the dataset, which means that throughout this notebook, I chose accuracy to be a relevant metric to look at. \n",
        "\n",
        "Bellow I import the data and define my train and test sets:"
      ],
      "metadata": {
        "id": "fY8YhczZVnH0"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.datasets import cifar10"
      ],
      "metadata": {
        "id": "yafABtToVl8i"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "(train_data, train_labels), (test_data, test_labels) = cifar10.load_data()"
      ],
      "metadata": {
        "id": "LL6G02CHR9rO",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c9b6d9b9-c395-42cf-d9a2-493e93badb9f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz\n",
            "170498071/170498071 [==============================] - 2s 0us/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 3. Preprocessing data\n",
        "\n",
        "There are a few of steps to facilitate the job of our classifier. Some processing is required here."
      ],
      "metadata": {
        "id": "7blay5E7f7ey"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.preprocessing import OneHotEncoder"
      ],
      "metadata": {
        "id": "HWdn0PKtx3ym"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Normalizing the data so the color values lie in the interval [0, 1]:\n",
        "train_data = train_data / 255.\n",
        "test_data = test_data / 255."
      ],
      "metadata": {
        "id": "tgt4lLRxxzvO"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Verifying the shape of our data:\n",
        "print(train_data.shape)\n",
        "print(test_data.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Obg-RzYrx11v",
        "outputId": "60172cc7-3d41-4824-ab72-52db5cb5213c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(50000, 32, 32, 3)\n",
            "(10000, 32, 32, 3)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "As can be seen above, we have 50K images for training and 10K images for testing of 32x32 pixels, as said before. Additionaly, we have 3 color channels, which accounts for the Red, Blue and Green colors.\n",
        "\n",
        "For further reference, we can plot one of them bellow. The left-most image has all 3 color channels joined. The subsequent 3 are plots for each color channel separately."
      ],
      "metadata": {
        "id": "Wi6YecfGx8Um"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "fig, axs = plt.subplots(nrows=1, ncols=4, figsize=(7,3))\n",
        "axs[0].imshow(train_data[0,...])\n",
        "axs[0].axis('off')\n",
        "axs[1].imshow(train_data[0,:,:,0], cmap='Reds')\n",
        "axs[1].axis('off')\n",
        "axs[2].imshow(train_data[0,:,:,1], cmap='Greens')\n",
        "axs[2].axis('off')\n",
        "axs[3].imshow(train_data[0,:,:,2], cmap='Blues')\n",
        "axs[3].axis('off')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 116
        },
        "id": "KKgG5f34XIAf",
        "outputId": "6cfb83b7-ae46-48e9-a535-245fd387757f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 504x216 with 4 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We can also display some labels to see what they look like:"
      ],
      "metadata": {
        "id": "XkQtAkc2cBsy"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(train_labels[:10].T)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3jCYjtkPcFhD",
        "outputId": "25294001-af15-4bf7-9b76-9f066911070c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[[6 9 9 4 1 1 2 7 8 3]]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "These are indexed labels, which is useful for classification, but not particularly enlightning. Let's create a class name for each of these indices."
      ],
      "metadata": {
        "id": "IbyuHMAMcJzh"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#These class names can be referenced in the dataset documentation\n",
        "class_names = ['airplane', 'automobile', 'bird', 'cat', 'deer',\n",
        "               'dog', 'frog', 'horse', 'ship', 'truck']"
      ],
      "metadata": {
        "id": "tOEPGyurX2Z-"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now, plotting the first ten images with their respective labels:"
      ],
      "metadata": {
        "id": "1eqHvGW_cX5F"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "fig, axs = plt.subplots(nrows=1, ncols=10, figsize=(15,7))\n",
        "\n",
        "for i in range(10):\n",
        "  axs[i].imshow(train_data[i,...])\n",
        "  axs[i].title.set_text(class_names[train_labels[i][0]])\n",
        "  axs[i].axis('off')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 118
        },
        "id": "GJl4KvO9Zct1",
        "outputId": "a5cae7ae-ce4e-4344-a747-267082778436"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x504 with 10 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Since we'll create a deep learning architecture with 10 units for output layer (one unit for each class), we also have to one-hot-encode these labels for the classifier to work:"
      ],
      "metadata": {
        "id": "fC3xuzPavqkW"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "enc = OneHotEncoder()\n",
        "transformed = enc.fit_transform(train_labels)\n",
        "train_labels_encoded = transformed.toarray()"
      ],
      "metadata": {
        "id": "m_75r4Tzv_CM"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now, we labels will look like this:"
      ],
      "metadata": {
        "id": "M-cLr6X-w3WB"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "train_labels_encoded[:10]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HnqQKW8Cwz_9",
        "outputId": "35bab1b1-77b1-4ba1-8dcb-b1565a2e52aa"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0., 0., 0., 0., 0., 0., 1., 0., 0., 0.],\n",
              "       [0., 0., 0., 0., 0., 0., 0., 0., 0., 1.],\n",
              "       [0., 0., 0., 0., 0., 0., 0., 0., 0., 1.],\n",
              "       [0., 0., 0., 0., 1., 0., 0., 0., 0., 0.],\n",
              "       [0., 1., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "       [0., 1., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "       [0., 0., 1., 0., 0., 0., 0., 0., 0., 0.],\n",
              "       [0., 0., 0., 0., 0., 0., 0., 1., 0., 0.],\n",
              "       [0., 0., 0., 0., 0., 0., 0., 0., 1., 0.],\n",
              "       [0., 0., 0., 1., 0., 0., 0., 0., 0., 0.]])"
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#We can do the same thing for the test labels\n",
        "enc = OneHotEncoder()\n",
        "transformed = enc.fit_transform(test_labels)\n",
        "test_labels_encoded = transformed.toarray()"
      ],
      "metadata": {
        "id": "7bOQBOklxVp2"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "In order to make it easier for the classifier by reducing the amount of data to be processed for each individual image, we can convert the images from RGB to greyscale. In order to do that, we average the color channels for each pixel. \n",
        "\n",
        "However, when we average over the color channels, the color dimension disapears. Tensorflow works by deafult with a color dimension, so we have to re-add it with np.newaxis command, which introduces a dummy dimension to our data."
      ],
      "metadata": {
        "id": "35hj7Q9Afz1_"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "train_data = np.average(train_data, axis=-1)\n",
        "test_data = np.average(test_data, axis=-1)\n",
        "\n",
        "train_data = train_data[...,np.newaxis] \n",
        "test_data = test_data[...,np.newaxis]\n",
        "\n",
        "#We can print the shape of our data to see if it was expected\n",
        "print(train_data.shape)\n",
        "print(test_data.shape)"
      ],
      "metadata": {
        "id": "2wtLLTUCebi0",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5f2d7646-2832-4161-8a08-0e5d88d73797"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(50000, 32, 32, 1)\n",
            "(10000, 32, 32, 1)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "To see if it actually worked, let's take a look at the data shape and plot the same first 10 figures as above:"
      ],
      "metadata": {
        "id": "ZbU0Psymgb3J"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "fig, axs = plt.subplots(nrows=1, ncols=10, figsize=(15,7))\n",
        "\n",
        "for i in range(10):\n",
        "  axs[i].imshow(train_data[i,...].squeeze(), cmap='gray')\n",
        "  axs[i].title.set_text(class_names[train_labels[i][0]])\n",
        "  axs[i].axis('off')\n",
        "plt.show()\n",
        "\n",
        "print('Data shape:', train_data.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 136
        },
        "id": "HT9nFUUmgaV5",
        "outputId": "c2b46eb2-75a1-4e87-9d84-0b6fe5648f1c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x504 with 10 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Data shape: (50000, 32, 32, 1)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Something worth addressing now is that we need to retain one image from this whole dataset. This one image will serve as a test subject for our feature extraction analysis. To be fair, such image can't be part of the training process of the benchmark classifier, nor the dummy classifier."
      ],
      "metadata": {
        "id": "Lb1blFeszaXq"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "one_image_data = train_data[0,...]\n",
        "\n",
        "train_data = train_data[1:,...]\n",
        "train_labels = train_labels[1:,...]\n",
        "train_labels_encoded = train_labels_encoded[1:,...]"
      ],
      "metadata": {
        "id": "-WBgercW0Z6k"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Also, Tensorflow needs the first dimension to be the batch dimension, so in a similar fashion as we did with the color channels, we have to add an extra dummy dimension:"
      ],
      "metadata": {
        "id": "CZxe1vdFStfu"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "one_image_data = one_image_data[np.newaxis,...]\n",
        "print(one_image_data.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "p4ismO3WS3Rf",
        "outputId": "6d861c2b-cfff-4cf3-a03f-98769be11887"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(1, 32, 32, 1)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now let's check if the one_image is the frog, and the first train_data is the truck:"
      ],
      "metadata": {
        "id": "2U74Whob1gWE"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "fig,ax = plt.subplots(nrows=1, ncols=2)\n",
        "\n",
        "ax[0].imshow(one_image_data.squeeze(), cmap='gray')\n",
        "ax[0].title.set_text('One_image')\n",
        "ax[0].axis('off')\n",
        "ax[1].imshow(train_data[0,...].squeeze(), cmap='gray')\n",
        "ax[1].title.set_text('First train_data')\n",
        "ax[1].axis('off')\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 199
        },
        "id": "41khEAYl0qgA",
        "outputId": "cfd002e1-af73-4e0b-a9ab-4a98121f3718"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Ok, now that we have reserved our frog image for feature extraction analysis, we can start assembling our neural networks."
      ],
      "metadata": {
        "id": "ZaAwOE6x2H-i"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 4. Building the CNN\n",
        "\n",
        "First I will import the necessary modules to build the CNN."
      ],
      "metadata": {
        "id": "Q_nAEXBXh63n"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.models import Model\n",
        "from tensorflow.keras.layers import Input, Conv2D, MaxPool2D, Dense, Dropout, Flatten\n",
        "from tensorflow.keras.callbacks import ModelCheckpoint, EarlyStopping"
      ],
      "metadata": {
        "id": "5x3ibJiahbzx"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now, let's try to build a model consisting of 13 layers, of which:\n",
        "- 0th: input layer\n",
        "- 1st and 2nd: Conv2D with 16 filters and 3x3 kernel size, ReLU activation function\n",
        "- 3th: MaxPool with pool size 2x2\n",
        "\n",
        "We repeat this process three times. Then,\n",
        "\n",
        "- 9th: Dense layer with 64 units and softmax activation function.\n",
        "- 10th: Dropout layer with rate 0.5.\n",
        "- 11th: a Flatten layer.\n",
        "- 12th: Dense layer with 10 units, softmax activation function for classification.\n",
        "\n",
        "The model will be compiled with an Adam optimizer, using as loss function the categorical cross entropy and, since the labels are hot-encoded, I'll use a simple accuracy metric."
      ],
      "metadata": {
        "id": "z2_9zt4qi-RX"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def get_new_model(input_shape):\n",
        "  '''\n",
        "  This function returns a compiled CNN with specifications given above.\n",
        "  '''\n",
        "\n",
        "  input_layer = Input(shape=input_shape, name='input')\n",
        "  h = Conv2D(filters=16, kernel_size=(3,3), \n",
        "             activation='relu', padding='same', name='conv2d_1')(input_layer)\n",
        "  h = Conv2D(filters=16, kernel_size=(3,3), \n",
        "             activation='relu', padding='same', name='conv2d_2')(h)\n",
        "\n",
        "  h = MaxPool2D(pool_size=(2,2), name='pool_1')(h)\n",
        "\n",
        "  h = Conv2D(filters=16, kernel_size=(3,3), \n",
        "             activation='relu', padding='same', name='conv2d_3')(h)\n",
        "  h = Conv2D(filters=16, kernel_size=(3,3), \n",
        "             activation='relu', padding='same', name='conv2d_4')(h)\n",
        "\n",
        "  h = MaxPool2D(pool_size=(2,2), name='pool_2')(h)\n",
        "\n",
        "  h = Conv2D(filters=16, kernel_size=(3,3), \n",
        "             activation='relu', padding='same', name='conv2d_5')(h)\n",
        "  h = Conv2D(filters=16, kernel_size=(3,3), \n",
        "             activation='relu', padding='same', name='conv2d_6')(h)\n",
        "    \n",
        "  h = Dense(64, activation='relu', name='dense_1')(h)\n",
        "  h = Dropout(0.5, name='dropout_1')(h)\n",
        "  h = Flatten(name='flatten_1')(h)\n",
        "  output_layer = Dense(10, activation='softmax', name='dense_2')(h)\n",
        "\n",
        "  model = Model(inputs=input_layer, outputs=output_layer, name='model_CNN')\n",
        "\n",
        "  model.compile(optimizer='adam',\n",
        "                loss='categorical_crossentropy',\n",
        "                metrics=['accuracy'])\n",
        "  \n",
        "  return model"
      ],
      "metadata": {
        "id": "1cTeMY8oi90M"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model_benchmark = get_new_model(input_shape=(32,32,1))\n",
        "model_benchmark.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iDCvs7fSuoqw",
        "outputId": "0bbbc5ce-5392-40c7-b550-33a9fbbd1e07"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_CNN\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input (InputLayer)          [(None, 32, 32, 1)]       0         \n",
            "                                                                 \n",
            " conv2d_1 (Conv2D)           (None, 32, 32, 16)        160       \n",
            "                                                                 \n",
            " conv2d_2 (Conv2D)           (None, 32, 32, 16)        2320      \n",
            "                                                                 \n",
            " pool_1 (MaxPooling2D)       (None, 16, 16, 16)        0         \n",
            "                                                                 \n",
            " conv2d_3 (Conv2D)           (None, 16, 16, 16)        2320      \n",
            "                                                                 \n",
            " conv2d_4 (Conv2D)           (None, 16, 16, 16)        2320      \n",
            "                                                                 \n",
            " pool_2 (MaxPooling2D)       (None, 8, 8, 16)          0         \n",
            "                                                                 \n",
            " conv2d_5 (Conv2D)           (None, 8, 8, 16)          2320      \n",
            "                                                                 \n",
            " conv2d_6 (Conv2D)           (None, 8, 8, 16)          2320      \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 8, 8, 64)          1088      \n",
            "                                                                 \n",
            " dropout_1 (Dropout)         (None, 8, 8, 64)          0         \n",
            "                                                                 \n",
            " flatten_1 (Flatten)         (None, 4096)              0         \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 10)                40970     \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 53,818\n",
            "Trainable params: 53,818\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "This will be a long-time training NN, therefore I will stablish a checkpoint callback for the best epoch, monitoring the validation set accuracy. In addition to that, I will also add an EarlyStopping callback that will monitor the validation set accuracy, with patience parameter set to 5."
      ],
      "metadata": {
        "id": "QHA3f8qonbNZ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def get_checkpoint_best_epoch(filepath):\n",
        "  checkpoint_path = filepath\n",
        "  checkpoint = ModelCheckpoint(filepath=checkpoint_path,\n",
        "                               save_weights_only=True,\n",
        "                               save_best_only=True,\n",
        "                               save_freq='epoch',\n",
        "                               monitor='val_accuracy',\n",
        "                               verbose=0)\n",
        "  return checkpoint\n",
        "\n",
        "\n",
        "def get_early_stopping():\n",
        "  callback = EarlyStopping(monitor='val_accuracy', patience=5)\n",
        "  return callback"
      ],
      "metadata": {
        "id": "dRHbUgE_ndmx"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "checkpoint_best_epoch = get_checkpoint_best_epoch(f'{saving_direc}Checkpoint_best_epoch_benchmark/checkpoint')\n",
        "early_stopping = get_early_stopping()"
      ],
      "metadata": {
        "id": "KEduzui8reBH"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now, for training: there will be 30 epochs, with validation set being the test set, and the provided callbacks. This training code will be executed only once. When the training is over and the weights of the model are tunned and saved, we can just load them whenever we want to recover the trained model."
      ],
      "metadata": {
        "id": "JvsbT5JFnKo-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "callbacks = [checkpoint_best_epoch, early_stopping]\n",
        "\n",
        "history_benchmark = model_benchmark.fit(x=train_data,\n",
        "                                        y=train_labels_encoded,\n",
        "                                        epochs=50,batch_size=100,\n",
        "                                        validation_data=(test_data,test_labels_encoded),\n",
        "                                        callbacks=callbacks)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "S2TccS_QpjLq",
        "outputId": "718c8638-295e-4153-f0dc-262ee5a5a6e5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/50\n",
            "500/500 [==============================] - 126s 252ms/step - loss: 0.6679 - accuracy: 0.7644 - val_loss: 0.7894 - val_accuracy: 0.7294\n",
            "Epoch 2/50\n",
            "500/500 [==============================] - 129s 257ms/step - loss: 0.6621 - accuracy: 0.7652 - val_loss: 0.8274 - val_accuracy: 0.7181\n",
            "Epoch 3/50\n",
            "500/500 [==============================] - 127s 253ms/step - loss: 0.6610 - accuracy: 0.7676 - val_loss: 0.7994 - val_accuracy: 0.7319\n",
            "Epoch 4/50\n",
            "500/500 [==============================] - 129s 257ms/step - loss: 0.6573 - accuracy: 0.7689 - val_loss: 0.7852 - val_accuracy: 0.7316\n",
            "Epoch 5/50\n",
            "500/500 [==============================] - 126s 251ms/step - loss: 0.6553 - accuracy: 0.7690 - val_loss: 0.8173 - val_accuracy: 0.7191\n",
            "Epoch 6/50\n",
            "500/500 [==============================] - 128s 256ms/step - loss: 0.6475 - accuracy: 0.7735 - val_loss: 0.7983 - val_accuracy: 0.7285\n",
            "Epoch 7/50\n",
            "500/500 [==============================] - 125s 250ms/step - loss: 0.6511 - accuracy: 0.7703 - val_loss: 0.7958 - val_accuracy: 0.7334\n",
            "Epoch 8/50\n",
            "500/500 [==============================] - 129s 258ms/step - loss: 0.6386 - accuracy: 0.7765 - val_loss: 0.8018 - val_accuracy: 0.7308\n",
            "Epoch 9/50\n",
            "500/500 [==============================] - 124s 248ms/step - loss: 0.6404 - accuracy: 0.7738 - val_loss: 0.7845 - val_accuracy: 0.7338\n",
            "Epoch 10/50\n",
            "500/500 [==============================] - 130s 261ms/step - loss: 0.6373 - accuracy: 0.7738 - val_loss: 0.8055 - val_accuracy: 0.7307\n",
            "Epoch 11/50\n",
            "500/500 [==============================] - 127s 254ms/step - loss: 0.6290 - accuracy: 0.7784 - val_loss: 0.7819 - val_accuracy: 0.7353\n",
            "Epoch 12/50\n",
            "500/500 [==============================] - 127s 255ms/step - loss: 0.6290 - accuracy: 0.7775 - val_loss: 0.7845 - val_accuracy: 0.7376\n",
            "Epoch 13/50\n",
            "500/500 [==============================] - 125s 249ms/step - loss: 0.6226 - accuracy: 0.7797 - val_loss: 0.8012 - val_accuracy: 0.7306\n",
            "Epoch 14/50\n",
            "500/500 [==============================] - 129s 258ms/step - loss: 0.6237 - accuracy: 0.7809 - val_loss: 0.8212 - val_accuracy: 0.7282\n",
            "Epoch 15/50\n",
            "500/500 [==============================] - 124s 247ms/step - loss: 0.6219 - accuracy: 0.7809 - val_loss: 0.8138 - val_accuracy: 0.7313\n",
            "Epoch 16/50\n",
            "500/500 [==============================] - 132s 265ms/step - loss: 0.6075 - accuracy: 0.7852 - val_loss: 0.7915 - val_accuracy: 0.7360\n",
            "Epoch 17/50\n",
            "500/500 [==============================] - 125s 251ms/step - loss: 0.6064 - accuracy: 0.7841 - val_loss: 0.7891 - val_accuracy: 0.7329\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We can use the history variable created to access the evolution of the loss function and the accuracy during the epochs:"
      ],
      "metadata": {
        "id": "1twfdlt3KvpT"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#Using the history method we can convert the history object created into\n",
        "#a Pandas dataframe\n",
        "df_benchmark = pd.DataFrame(history_benchmark.history)\n",
        "df_benchmark.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "YF5x64YWPojI",
        "outputId": "0d8143c9-f540-4cfc-e17d-bf1cef8dd51e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "       loss  accuracy  val_loss  val_accuracy\n",
              "0  0.667899  0.764415  0.789441        0.7294\n",
              "1  0.662114  0.765235  0.827421        0.7181\n",
              "2  0.661000  0.767575  0.799365        0.7319\n",
              "3  0.657292  0.768915  0.785238        0.7316\n",
              "4  0.655326  0.769015  0.817274        0.7191"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-cda3340c-d196-4800-b91f-eb9f729a0001\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>loss</th>\n",
              "      <th>accuracy</th>\n",
              "      <th>val_loss</th>\n",
              "      <th>val_accuracy</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0.667899</td>\n",
              "      <td>0.764415</td>\n",
              "      <td>0.789441</td>\n",
              "      <td>0.7294</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0.662114</td>\n",
              "      <td>0.765235</td>\n",
              "      <td>0.827421</td>\n",
              "      <td>0.7181</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0.661000</td>\n",
              "      <td>0.767575</td>\n",
              "      <td>0.799365</td>\n",
              "      <td>0.7319</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0.657292</td>\n",
              "      <td>0.768915</td>\n",
              "      <td>0.785238</td>\n",
              "      <td>0.7316</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0.655326</td>\n",
              "      <td>0.769015</td>\n",
              "      <td>0.817274</td>\n",
              "      <td>0.7191</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-cda3340c-d196-4800-b91f-eb9f729a0001')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-cda3340c-d196-4800-b91f-eb9f729a0001 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-cda3340c-d196-4800-b91f-eb9f729a0001');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 89
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_benchmark['loss'].plot(legend=True)\n",
        "df_benchmark['val_loss'].plot(title='Losses', grid=True,\n",
        "                              xlabel='Epochs',\n",
        "                              ylabel='Loss',\n",
        "                              legend=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 313
        },
        "id": "80wo6VAmQBR4",
        "outputId": "543650ac-adbd-41f4-eab4-d74b4a214153"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f00321bc700>"
            ]
          },
          "metadata": {},
          "execution_count": 99
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_benchmark['accuracy'].plot(legend=True)\n",
        "df_benchmark['val_accuracy'].plot(title='Accuracies', grid=True,\n",
        "                              xlabel='Epochs',\n",
        "                              ylabel='Accuracy',\n",
        "                              legend=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 313
        },
        "id": "nQBmmFmaRM2F",
        "outputId": "607a3a7e-9513-4e74-f773-2baf3a22fc58"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f002fea5880>"
            ]
          },
          "metadata": {},
          "execution_count": 100
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We can see a discrepancy of the losses and the accuracy between the train and validation sets. The loss function shows a discrepancy a little higher than I anticipated, but this will do for now. The accuracies are within the expected, since the scale of the image provides a false sensation of a validation set much worse than the training set, but they are close in performance.\n",
        "\n",
        "Also, something worth noticing is that these metrics don't seem to improve much during training. This is because I had previously trained the model and had to stop the training step, and re-start it later. The actual training took more than 20 epochs and improved a lot in its course."
      ],
      "metadata": {
        "id": "xAGGQ6HrTDq5"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now, creating a new benchmark model with tunned parameters is easy, since we just have to load the trained weights:"
      ],
      "metadata": {
        "id": "IpWWauHyD1nW"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model_benchmark = get_new_model(input_shape=train_data.shape[1:])\n",
        "model_benchmark.load_weights(filepath=f'{saving_direc}Checkpoint_best_epoch_benchmark/checkpoint')"
      ],
      "metadata": {
        "id": "xFrfeoSSUNbZ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "cdb11bd3-1f85-4991-dac9-ab61611ebf82"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.training.tracking.util.CheckpointLoadStatus at 0x7f8a7f3f9340>"
            ]
          },
          "metadata": {},
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 5. Dummy CNN\n",
        "\n",
        "Now I will focus on a second CNN, which will be a CNN with same architecture as our benchmark, but will be trained with significantly less data.\n",
        "\n",
        "Creating a new model is as easy as calling the get_new_model function. However, we need to take care of our data first."
      ],
      "metadata": {
        "id": "fOd0uzWzEW87"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 5.1 Taking care of our data first\n",
        "\n",
        "The cifar10 database says in its documentation that the labels are uniformly distributed, e.g. this is a balanced dataset. Let's confirm that:"
      ],
      "metadata": {
        "id": "wealudakFid-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "unique_train, counts_train = np.unique(train_labels, return_counts=True)\n",
        "unique_test, counts_test = np.unique(test_labels, return_counts=True)\n",
        "\n",
        "print('Train set:',dict(zip(unique_train, counts_train)))\n",
        "print('Test set:',dict(zip(unique_test, counts_test)))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Rt30VxjmDy12",
        "outputId": "d3f3cf76-3365-45f1-e707-7bd73e04bfb0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Train set: {0: 5000, 1: 5000, 2: 5000, 3: 5000, 4: 5000, 5: 5000, 6: 4999, 7: 5000, 8: 5000, 9: 5000}\n",
            "Test set: {0: 1000, 1: 1000, 2: 1000, 3: 1000, 4: 1000, 5: 1000, 6: 1000, 7: 1000, 8: 1000, 9: 1000}\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "As we can see, the distribution of labels is perfectly uniform, with the exception of the frog (label 6) we retained for analysis. We need to preserve that for our \"dummy\" classifier. Instead of subsampling our train data and try to replicate the proportions, I will simply use the test data to train the dummy model, and then use the train data to test it. Since the database is the same, we don't have a problem with the nature of the data used, and we benefit from a training dataset that contains 10% of the images present in the benchmark classifier."
      ],
      "metadata": {
        "id": "dY-CWbLVGpzR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#Redefining the test to be train, and vice versa:\n",
        "\n",
        "train_data_dummy = test_data\n",
        "train_labels_encoded_dummy = test_labels_encoded\n",
        "\n",
        "test_data_dummy = train_data\n",
        "test_labels_encoded_dummy = train_labels_encoded"
      ],
      "metadata": {
        "id": "zuwSJaCfG4dy"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(train_data_dummy.shape)\n",
        "print(test_data_dummy.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WfQLHBd7ZJoA",
        "outputId": "80946dae-119b-4537-c781-9e8602e8b1ee"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(10000, 32, 32, 1)\n",
            "(49999, 32, 32, 1)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 5.2 Training the dummy CNN\n",
        "\n",
        "Now I will build a compiled CNN model with same architecture as the benchmark, and train it exactly with the same parameters as before."
      ],
      "metadata": {
        "id": "aaZxJfCDYjKU"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model_dummy = get_new_model(input_shape=train_data_dummy.shape[1:])\n",
        "model_dummy.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1LVb9tCSZ2wT",
        "outputId": "eb0d4103-855c-4b06-861a-50d2185b84f2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_CNN\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input (InputLayer)          [(None, 32, 32, 1)]       0         \n",
            "                                                                 \n",
            " conv2d_1 (Conv2D)           (None, 32, 32, 16)        160       \n",
            "                                                                 \n",
            " conv2d_2 (Conv2D)           (None, 32, 32, 16)        2320      \n",
            "                                                                 \n",
            " pool_1 (MaxPooling2D)       (None, 16, 16, 16)        0         \n",
            "                                                                 \n",
            " conv2d_3 (Conv2D)           (None, 16, 16, 16)        2320      \n",
            "                                                                 \n",
            " conv2d_4 (Conv2D)           (None, 16, 16, 16)        2320      \n",
            "                                                                 \n",
            " pool_2 (MaxPooling2D)       (None, 8, 8, 16)          0         \n",
            "                                                                 \n",
            " conv2d_5 (Conv2D)           (None, 8, 8, 16)          2320      \n",
            "                                                                 \n",
            " conv2d_6 (Conv2D)           (None, 8, 8, 16)          2320      \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 8, 8, 64)          1088      \n",
            "                                                                 \n",
            " dropout_1 (Dropout)         (None, 8, 8, 64)          0         \n",
            "                                                                 \n",
            " flatten_1 (Flatten)         (None, 4096)              0         \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 10)                40970     \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 53,818\n",
            "Trainable params: 53,818\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "checkpoint_best_epoch = get_checkpoint_best_epoch(f'{saving_direc}Checkpoint_best_epoch_dummy/checkpoint')\n",
        "early_stopping = get_early_stopping()"
      ],
      "metadata": {
        "id": "Z1F59sSTZ31l"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "callbacks = [checkpoint_best_epoch, early_stopping]\n",
        "\n",
        "history_dummy = model_dummy.fit(x=train_data_dummy,\n",
        "                                y=train_labels_encoded_dummy,\n",
        "                                epochs=50,\n",
        "                                batch_size=100,\n",
        "                                validation_data=(test_data_dummy,test_labels_encoded_dummy),\n",
        "                                callbacks=callbacks)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7cpiYvgRInhr",
        "outputId": "1fac3dc9-cbe9-4b8e-a5ff-37cda799460e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/50\n",
            "100/100 [==============================] - 55s 541ms/step - loss: 2.1473 - accuracy: 0.2014 - val_loss: 1.9539 - val_accuracy: 0.2924\n",
            "Epoch 2/50\n",
            "100/100 [==============================] - 56s 566ms/step - loss: 1.8619 - accuracy: 0.3306 - val_loss: 1.7821 - val_accuracy: 0.3593\n",
            "Epoch 3/50\n",
            "100/100 [==============================] - 58s 585ms/step - loss: 1.6917 - accuracy: 0.3913 - val_loss: 1.6289 - val_accuracy: 0.4137\n",
            "Epoch 4/50\n",
            "100/100 [==============================] - 54s 542ms/step - loss: 1.6204 - accuracy: 0.4190 - val_loss: 1.5964 - val_accuracy: 0.4293\n",
            "Epoch 5/50\n",
            "100/100 [==============================] - 52s 519ms/step - loss: 1.5444 - accuracy: 0.4443 - val_loss: 1.5228 - val_accuracy: 0.4594\n",
            "Epoch 6/50\n",
            "100/100 [==============================] - 56s 564ms/step - loss: 1.4914 - accuracy: 0.4707 - val_loss: 1.5191 - val_accuracy: 0.4576\n",
            "Epoch 7/50\n",
            "100/100 [==============================] - 51s 518ms/step - loss: 1.4475 - accuracy: 0.4866 - val_loss: 1.4832 - val_accuracy: 0.4736\n",
            "Epoch 8/50\n",
            "100/100 [==============================] - 54s 540ms/step - loss: 1.3804 - accuracy: 0.5119 - val_loss: 1.4209 - val_accuracy: 0.4958\n",
            "Epoch 9/50\n",
            "100/100 [==============================] - 51s 515ms/step - loss: 1.3437 - accuracy: 0.5229 - val_loss: 1.4215 - val_accuracy: 0.4995\n",
            "Epoch 10/50\n",
            "100/100 [==============================] - 54s 542ms/step - loss: 1.3064 - accuracy: 0.5403 - val_loss: 1.3779 - val_accuracy: 0.5114\n",
            "Epoch 11/50\n",
            "100/100 [==============================] - 53s 537ms/step - loss: 1.2761 - accuracy: 0.5472 - val_loss: 1.3654 - val_accuracy: 0.5205\n",
            "Epoch 12/50\n",
            "100/100 [==============================] - 54s 542ms/step - loss: 1.2182 - accuracy: 0.5696 - val_loss: 1.3550 - val_accuracy: 0.5231\n",
            "Epoch 13/50\n",
            "100/100 [==============================] - 51s 517ms/step - loss: 1.2000 - accuracy: 0.5764 - val_loss: 1.3646 - val_accuracy: 0.5201\n",
            "Epoch 14/50\n",
            "100/100 [==============================] - 55s 556ms/step - loss: 1.1580 - accuracy: 0.5891 - val_loss: 1.3288 - val_accuracy: 0.5380\n",
            "Epoch 15/50\n",
            "100/100 [==============================] - 51s 517ms/step - loss: 1.1288 - accuracy: 0.6025 - val_loss: 1.2973 - val_accuracy: 0.5458\n",
            "Epoch 16/50\n",
            "100/100 [==============================] - 56s 565ms/step - loss: 1.1039 - accuracy: 0.6052 - val_loss: 1.3184 - val_accuracy: 0.5416\n",
            "Epoch 17/50\n",
            "100/100 [==============================] - 51s 516ms/step - loss: 1.0659 - accuracy: 0.6235 - val_loss: 1.3541 - val_accuracy: 0.5368\n",
            "Epoch 18/50\n",
            "100/100 [==============================] - 54s 540ms/step - loss: 1.0444 - accuracy: 0.6292 - val_loss: 1.2790 - val_accuracy: 0.5575\n",
            "Epoch 19/50\n",
            "100/100 [==============================] - 54s 539ms/step - loss: 1.0341 - accuracy: 0.6351 - val_loss: 1.3052 - val_accuracy: 0.5446\n",
            "Epoch 20/50\n",
            "100/100 [==============================] - 52s 519ms/step - loss: 1.0016 - accuracy: 0.6437 - val_loss: 1.2733 - val_accuracy: 0.5591\n",
            "Epoch 21/50\n",
            "100/100 [==============================] - 54s 541ms/step - loss: 0.9778 - accuracy: 0.6561 - val_loss: 1.2796 - val_accuracy: 0.5612\n",
            "Epoch 22/50\n",
            "100/100 [==============================] - 54s 540ms/step - loss: 0.9666 - accuracy: 0.6603 - val_loss: 1.2708 - val_accuracy: 0.5632\n",
            "Epoch 23/50\n",
            "100/100 [==============================] - 52s 523ms/step - loss: 0.9436 - accuracy: 0.6638 - val_loss: 1.2682 - val_accuracy: 0.5670\n",
            "Epoch 24/50\n",
            "100/100 [==============================] - 53s 534ms/step - loss: 0.9162 - accuracy: 0.6804 - val_loss: 1.3227 - val_accuracy: 0.5477\n",
            "Epoch 25/50\n",
            "100/100 [==============================] - 53s 536ms/step - loss: 0.9060 - accuracy: 0.6790 - val_loss: 1.2919 - val_accuracy: 0.5646\n",
            "Epoch 26/50\n",
            "100/100 [==============================] - 52s 526ms/step - loss: 0.8795 - accuracy: 0.6843 - val_loss: 1.2876 - val_accuracy: 0.5659\n",
            "Epoch 27/50\n",
            "100/100 [==============================] - 56s 564ms/step - loss: 0.8678 - accuracy: 0.6882 - val_loss: 1.2977 - val_accuracy: 0.5600\n",
            "Epoch 28/50\n",
            "100/100 [==============================] - 51s 517ms/step - loss: 0.8535 - accuracy: 0.7019 - val_loss: 1.3120 - val_accuracy: 0.5648\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_dummy = pd.DataFrame(history_dummy.history)\n",
        "df_dummy.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "8ZHIkAJ1R8Qx",
        "outputId": "bec61b82-ffeb-4c0e-e2bf-76e26dc9b42e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "       loss  accuracy  val_loss  val_accuracy\n",
              "0  2.147257    0.2014  1.953896      0.292386\n",
              "1  1.861860    0.3306  1.782106      0.359347\n",
              "2  1.691717    0.3913  1.628928      0.413668\n",
              "3  1.620391    0.4190  1.596420      0.429349\n",
              "4  1.544360    0.4443  1.522831      0.459389"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-af7e1e06-4723-4e1d-9133-30decd6c6e8d\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>loss</th>\n",
              "      <th>accuracy</th>\n",
              "      <th>val_loss</th>\n",
              "      <th>val_accuracy</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>2.147257</td>\n",
              "      <td>0.2014</td>\n",
              "      <td>1.953896</td>\n",
              "      <td>0.292386</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1.861860</td>\n",
              "      <td>0.3306</td>\n",
              "      <td>1.782106</td>\n",
              "      <td>0.359347</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1.691717</td>\n",
              "      <td>0.3913</td>\n",
              "      <td>1.628928</td>\n",
              "      <td>0.413668</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1.620391</td>\n",
              "      <td>0.4190</td>\n",
              "      <td>1.596420</td>\n",
              "      <td>0.429349</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1.544360</td>\n",
              "      <td>0.4443</td>\n",
              "      <td>1.522831</td>\n",
              "      <td>0.459389</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-af7e1e06-4723-4e1d-9133-30decd6c6e8d')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-af7e1e06-4723-4e1d-9133-30decd6c6e8d button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-af7e1e06-4723-4e1d-9133-30decd6c6e8d');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 36
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_dummy['loss'].plot(legend=True)\n",
        "df_dummy['val_loss'].plot(title='Losses', grid=True,\n",
        "                              xlabel='Epochs',\n",
        "                              ylabel='Loss',\n",
        "                              legend=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 313
        },
        "id": "59WBanAjSHet",
        "outputId": "2281a6ea-3aa1-4e5b-f5b3-073ced308b56"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f8a7ed16310>"
            ]
          },
          "metadata": {},
          "execution_count": 37
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_dummy['accuracy'].plot(legend=True)\n",
        "df_dummy['val_accuracy'].plot(title='Accuracies', grid=True,\n",
        "                              xlabel='Epochs',\n",
        "                              ylabel='Accuracy',\n",
        "                              legend=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 313
        },
        "id": "qv4p0p3jSPvk",
        "outputId": "39c83973-df46-40d9-8c50-7a8d766e18d2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f8a7f164280>"
            ]
          },
          "metadata": {},
          "execution_count": 38
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We see a different evolution in relation to the benchmark classifier. This is because I didn't have to stop the training process in the middle of it. Now we can clearly see the improvement of the metrics over the epochs.\n",
        "\n",
        "Also, the performance of the dummy classifier is clearly worse than the benchmark, so certainly these are statistically different classifiers."
      ],
      "metadata": {
        "id": "3xRPYEE_XcsS"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model_dummy.load_weights(f'{saving_direc}Checkpoint_best_epoch_dummy/checkpoint')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DHOtvzuDqCYe",
        "outputId": "7ceaeb92-3406-46e6-e2a7-3c490489e18c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.training.tracking.util.CheckpointLoadStatus at 0x7f8a87fd0190>"
            ]
          },
          "metadata": {},
          "execution_count": 39
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 6. Feature extraction\n",
        "Now that we have our two models: one benchmark and one \"dummy\" that performs sensibly worse than the benchmark, we can see how both of these perform on feature extraction. We'll be looking at some key layers and comparing them."
      ],
      "metadata": {
        "id": "sO0v_lumqpVI"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 6.2 Analysis of the benchmark classifier"
      ],
      "metadata": {
        "id": "EvYo-NixwOzH"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model_benchmark.summary()"
      ],
      "metadata": {
        "id": "zuM2LyUtw6az",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "3c453d5e-2a64-4137-f79d-b4716fe536cf"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_CNN\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input (InputLayer)          [(None, 32, 32, 1)]       0         \n",
            "                                                                 \n",
            " conv2d_1 (Conv2D)           (None, 32, 32, 16)        160       \n",
            "                                                                 \n",
            " conv2d_2 (Conv2D)           (None, 32, 32, 16)        2320      \n",
            "                                                                 \n",
            " pool_1 (MaxPooling2D)       (None, 16, 16, 16)        0         \n",
            "                                                                 \n",
            " conv2d_3 (Conv2D)           (None, 16, 16, 16)        2320      \n",
            "                                                                 \n",
            " conv2d_4 (Conv2D)           (None, 16, 16, 16)        2320      \n",
            "                                                                 \n",
            " pool_2 (MaxPooling2D)       (None, 8, 8, 16)          0         \n",
            "                                                                 \n",
            " conv2d_5 (Conv2D)           (None, 8, 8, 16)          2320      \n",
            "                                                                 \n",
            " conv2d_6 (Conv2D)           (None, 8, 8, 16)          2320      \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 8, 8, 64)          1088      \n",
            "                                                                 \n",
            " dropout_1 (Dropout)         (None, 8, 8, 64)          0         \n",
            "                                                                 \n",
            " flatten_1 (Flatten)         (None, 4096)              0         \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 10)                40970     \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 53,818\n",
            "Trainable params: 53,818\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Here I'm slicing the layers of the CNN and looking at their outputs. I can create a Model that return each layer's output."
      ],
      "metadata": {
        "id": "x_f00baiYc0u"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "benchmark_layers = model_benchmark.layers\n",
        "benchmark_input = model_benchmark.input\n",
        "\n",
        "layer_outputs_benchmark = [layer.output for layer in benchmark_layers]\n",
        "features_benchmark = Model(inputs=benchmark_input, outputs=layer_outputs_benchmark)"
      ],
      "metadata": {
        "id": "7gK0OwUIxn52"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "First, we look at the output of the first layer (i.e. the input layer) as a sanity check, since it has to match the input given:"
      ],
      "metadata": {
        "id": "MPdWRzVJZQ6E"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "extracted_benchmark = features_benchmark(one_image_data)\n",
        "\n",
        "f1_benchmark = extracted_benchmark[0]\n",
        "\n",
        "print('\\n f1_benchmark shape:', f1_benchmark.shape)\n",
        "\n",
        "imgs = f1_benchmark[0,...]\n",
        "\n",
        "plt.imshow(imgs[...,0], cmap='gray')\n",
        "plt.axis('off')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 284
        },
        "id": "kTLxw57a0Th6",
        "outputId": "a5703689-2f20-492e-c7c8-49968187cd1c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            " f1_benchmark shape: (1, 32, 32, 1)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We can see that it matches our frog pal. Now, we can look at a deeper layer, after the first convolution:"
      ],
      "metadata": {
        "id": "rDG2tunOZY5g"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "f2_benchmark = extracted_benchmark[1]\n",
        "\n",
        "print('\\n f2_benchmark shape:', f2_benchmark.shape)\n",
        "\n",
        "imgs = f2_benchmark[0,...]\n",
        "plt.figure(figsize=(15,15))\n",
        "for n in range(16):\n",
        "    ax = plt.subplot(4, 4, n+1)\n",
        "    plt.imshow(imgs[...,n], cmap='gray')\n",
        "    plt.axis('off')\n",
        "plt.subplots_adjust(wspace=0.01, hspace=0.01)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 882
        },
        "id": "tsxvi0PAUFNb",
        "outputId": "f3dd7a26-47bd-4c2e-ed47-96f2faf29ecb"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            " f2_benchmark shape: (1, 32, 32, 16)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Since there are 16 filters on the convolutional layers, we observe 16 images as result of the first convolution. We already have a strong processing of the image, highlighting key aspects of the frog, like its eyes and paws.\n",
        "\n",
        "We can look at a super deep layer now that will be heavily processed. Here we see the output of the last convolutional layer. At this stage, only the most meaningful pixels appear in the image, and the information contained is super non-redundant:"
      ],
      "metadata": {
        "id": "p0zzIsutZtBS"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "f3_benchmark = extracted_benchmark[9]\n",
        "\n",
        "print('\\n f3_benchmark shape:', f3_benchmark.shape)\n",
        "\n",
        "imgs = f3_benchmark[0,...]\n",
        "plt.figure(figsize=(15,15))\n",
        "for n in range(16):\n",
        "    ax = plt.subplot(4, 4, n+1)\n",
        "    plt.imshow(imgs[...,n], cmap='gray')\n",
        "    plt.axis('off')\n",
        "plt.subplots_adjust(wspace=0.01, hspace=0.01)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 882
        },
        "id": "VL-HZpEr5LXM",
        "outputId": "87cfd236-eb5f-4a45-dab4-c71d7c01b8b4"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            " f3_benchmark shape: (1, 8, 8, 64)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "A nice thing to notice is the drastic reduction on the image size due to the poolings performed through the network. We can see how it affects interpretability of the images, since we can't recognize any trace of the initial frog."
      ],
      "metadata": {
        "id": "Pxo2LmcygQ8-"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now, let's look the output of the model, as a probability distribution over the classes:"
      ],
      "metadata": {
        "id": "FYqJGpbybIej"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "fig = plt.figure(figsize=(6,4))\n",
        "ax = fig.add_axes([0,0,1,1])\n",
        "ax.bar(np.arange(0,10), model_benchmark.predict(one_image_data, verbose=False)[0])\n",
        "ax.set_title('Probability distribution', fontsize=20)\n",
        "ax.set_xticks(np.arange(0,10))\n",
        "ax.set_xticklabels(class_names, rotation=45, fontsize=14)\n",
        "ax.set_xlabel('Class name', fontsize=18)\n",
        "ax.set_ylabel('Probability', fontsize=18)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 435
        },
        "id": "eYlgNxE0dniM",
        "outputId": "c05731b7-4058-45f8-db4f-b49aa29d86e3"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 6.3 Analysis of the dummy classifier\n",
        "\n",
        "Now we turn ourselves to the dummy classifier. We expect that the feature selection in here will be much less tunned than the benchmark classifier. Since the process will be prety much the same as the benchmark, I will be brief with the comments."
      ],
      "metadata": {
        "id": "4E6ayYXIgodI"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "dummy_layers = model_dummy.layers\n",
        "dummy_input = model_dummy.input\n",
        "\n",
        "layer_outputs_dummy = [layer.output for layer in dummy_layers]\n",
        "features_dummy = Model(inputs=dummy_input, outputs=layer_outputs_dummy)"
      ],
      "metadata": {
        "id": "hHosbuVKfFEu"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "extracted_dummy = features_dummy(one_image_data)\n",
        "\n",
        "f1_dummy = extracted_dummy[0]\n",
        "\n",
        "print('\\n f1_dummy shape:', f1_dummy.shape)\n",
        "\n",
        "imgs = f1_dummy[0,...]\n",
        "\n",
        "plt.imshow(imgs[...,0], cmap='gray')\n",
        "plt.axis('off')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 284
        },
        "id": "WywP2d7LXFIt",
        "outputId": "62febdb6-d608-42d2-e761-44c0ae1f87c6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            " f1_dummy shape: (1, 32, 32, 1)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "f2_dummy = extracted_dummy[1]\n",
        "\n",
        "print('\\n f2_dummy shape:', f2_dummy.shape)\n",
        "\n",
        "imgs = f2_dummy[0,...]\n",
        "plt.figure(figsize=(15,15))\n",
        "for n in range(16):\n",
        "    ax = plt.subplot(4, 4, n+1)\n",
        "    plt.imshow(imgs[...,n], cmap='gray')\n",
        "    plt.axis('off')\n",
        "plt.subplots_adjust(wspace=0.01, hspace=0.01)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 882
        },
        "id": "3x8kZpHeXQw0",
        "outputId": "ef2a01fb-9b36-4454-811c-77630f137414"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            " f2_dummy shape: (1, 32, 32, 16)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "f3_dummy = extracted_dummy[9]\n",
        "\n",
        "print('\\n f3_dummy shape:', f3_dummy.shape)\n",
        "\n",
        "imgs = f3_dummy[0,...]\n",
        "plt.figure(figsize=(15,15))\n",
        "for n in range(16):\n",
        "    ax = plt.subplot(4, 4, n+1)\n",
        "    plt.imshow(imgs[...,n], cmap='gray')\n",
        "    plt.axis('off')\n",
        "plt.subplots_adjust(wspace=0.01, hspace=0.01)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 882
        },
        "id": "80zX9J7nXY4N",
        "outputId": "09a30cbd-28d1-45ea-97f4-e6a4fc019708"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            " f3_dummy shape: (1, 8, 8, 64)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 16 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA04AAAM9CAYAAACrFiPuAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3cbcjfZd3H8e9vW3PTvJ9ta8nMyUCJtG2ghDq8STOkThQLH9iDBNHI6IYkKaywQiy6FRYji1Cvmlpqlo4kJe8fTFGYaToFc+bNluXUzWnudz2Li67993Gw4/c/O/d6PfR48Pm17djZu0Pq+r4vAAAARps27g8AAACY7IQTAABAIJwAAAAC4QQAABAIJwAAgGDGjg67rvN/uTfJ7LvvvoPsvPzyy803ptJ/lr7vu1Fn7tHua8mSJc03HnvsseYbVVVHHXVU84177rnHPWIsDjvssEF2nnvuueYbr7766sh7tHz58ub3aP369a0nqqrqqaeeGmSHyWePPfZovvH6669v9x55cQIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAADBjHF/wFQya9as5hvnnXde842qqocffrj5Rtd1zTeqqs4444xBduA/Pfjgg+P+hF3m3nvvHfcnQDPr1q0b9ycM4qijjmq+8de//rX5xlD222+/5htLly5tvlFV9eSTTzbfOProo5tvVFVt3LhxkJ3t8eIEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQdH3fjzw8+OCDRx/uIuvXr289MaXs6Pfrv81dd901yM63vvWt5hurV6/uRp3tu+++zX/TNm3a1HoCmuv7fuQ96rpuyvzlN8Tf43fccUfzjRNPPLH5Bjtvd7lHU8UJJ5wwyM6GDRuab6xdu7b5RlXVrFmzmm9s2bJlu/fIixMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIZuzocP369UN9B29T13WD7Dz77LPNN44//vjmG1VVhx122CA7o8yePbv5xqZNm5pvsPOWLl3afGPNmjXNN6qqHnjggUF2Rjn22GObb5x77rnNN6qqVq1a1Xzj5z//efONJUuWNN+oqnrwwQcH2YFx2LJlyyA7a9euHWRnCK+//vrYtr04AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgKDr+370YdeNPgT+re/7btSZe0RLe+211yA7r732WvONHd2jlStXNr9Hd9xxR+uJqqr61a9+NcgOu6fd5efRxMRE843bb7+9+carr77afKOq6oEHHmi+sWzZsuYbVVVvvfVW841R98iLEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQDBj3B8wlZxyyinNN1avXt18o6pq2rT2Tf3xj3+8+UZV1bXXXjvIDozDa6+9Nu5PGMTGjRubb9x0003NN2Cqmzlz5iA7N9xwQ/ONhQsXNt/YtGlT842qqnPOOaf5xvz585tvVA33Z2x7vDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAYMa4P2Aq+cMf/tB8Y9q0YVp30aJFU2IDmBo2bdrUfGO//fZrvlFVtWXLlkF2YBzeeOONQXa6rmu+MWfOnOYbQ1m7du24P2GXOf7448e27cUJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAg6Pq+H33YdaMPd5GvfvWrrSeqqmrjxo3NN+6///7mGw899FDzDXZe3/fdqLPly5c3v0d33nln6wlobkf3aIifR0xO73nPe5pvrF+/vvnGUNyj/y4HHHDAuD9hl1m8ePEgO8uWLWu+8eMf/3i798iLEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAi6vu/H/Q0AAACTmhcnAACAYMaODruu8xw1yey9996D7LzyyivNN6ZPn958o6rqrbfear7R93036sw9mnwOPPDAQXZefvnl5ht77bVX842qqi1btjTf2Lp1q3vEWBxxxBGD7Lz3ve9tvvG73/1u5D2aOXNm83v05ptvtp4YTNeN/KXcZabSv/k1d+7cQXaWLFnSfOOWW27Z7m++FycAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAgq7v+9GHXTf6kLGYO3fuIDuLFi1qvvHoo48236iq+sc//tF8o+/7btSZe0RLe+655yA7mzdvbr6xu9yjHf3c3VW6buQvJVPcju7R9ddf3/wP3zXXXNN6oqqqbrjhhuYb+++/f/ONf/7zn8032Hmj7pEXJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgGDGuD9gKtl///2bb7z00kvNN6qqDj/88OYb27Zta77BrjUxMdF8Y8GCBc03PvCBDzTfqKr65S9/2XzjkUceab5RVbV48eJBdkY54IADmm/MmjWr+UZV1Sc/+clBduA/bd68ufnGkiVLmm9UVXVdN8gO/F9enAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAwY9wfMJRjjjmm+cZ9993XfGNiYqL5RlXVTTfdNMjO7uDII49svrFw4cLmG1VVp556avON888/v/nGZZdd1nyjquqPf/zjIDtDeP7558e6v3jx4uYbmzZtar5RVXXVVVcNssPb9853vnOQnenTpw+yM8rNN9/cfOOxxx5rvsHktHnz5kF2vva1rw2ysz1enAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAwY9wfMJQLLrig+cacOXOab/z9739vvlFV9bnPfa75xg9+8IPmG5PBEH8uFi9e3HyjqmrevHnNNw466KDmGxs3bmy+UVV13nnnNd9YuXJl843J4Le//W3zjS9+8YvNN6qq/vznPw+yw9v36quvjvsTBnHZZZc137jllluab1RVffaznx1kZ6qYP39+842h/g5dsWJF843LL798u//cixMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEDQ9X0/+rDrRh/uIkuWLGk9UVVVl1xySfONiYmJ5htTybRpw3T7tm3bmm/0fd+NOpuYmGh+j15//fXWE1VVNW/evOYbv/jFL5pvMDnt6B7dddddze/R8ccf33piMB/96Eebbwzx90FV1cqVKwfZmSp2dI+eeOKJ5vfoIx/5SOuJqqpat25d842zzz67+cbTTz/dfKOq6t577x1kZ6oYdY+8OAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBgxo4OFy5c2PwDXnzxxeYbVVUTExOD7PD2HXTQQYPsvPDCC4PsjPLII48031i3bl3zDSanPfbYY5CdrVu3DrIzyo9+9KOx7v+3+cxnPtN8Y86cOc03qqo2b97cfOPqq69uvjEZXHjhhc03hvp59MEPfrD5xpe+9KXmGwsWLGi+UVX19a9/vfnGihUrmm+MmxcnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAYMaODp9++umhvoPd0AsvvDDuTxjEunXrxv0JjMk73vGO5htbt25tvjEZzJ8/v/nG4sWLm29UVXVd13zjlFNOab5x5JFHNt+oGubXa3cxa9as5huHHHJI842qqieffLL5xpIlS5pvsPMWLFgwtm0vTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACDo+r4f9zcAAABMal6cAAAAghk7Ouy6znMUvA1933ejztwjWpo9e/YgO2+99Vbzja1bt468RwceeGDze3Tttde2nqiqqu9+97vNN4b4/brtttuab1RV7b333s03XnnlleYbQxn3z6MPf/jDrSeqququu+5qvjF9+vTmG5s2bWq+wc4bdY+8OAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQzNjR4bJly5p/wJo1a5pvAExVW7ZsGfcnDOKll15qvrFq1armG1VVq1evHmSntUMPPXSQnfe///3NN2688cbmG5PB7Nmzm2+sXbu2+UZV1WuvvTbITmtz5swZZOfkk09uvvHrX/+6+UZV1ZtvvjnIzvZ4cQIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAi6vu9HH3bd6MP/MmeffXbzjblz5zbfuOiii5pvVFV96lOfar6xePHi5htVVX/605+abzz00EPdqLOpdI9+85vfNN/45S9/2Xzj2muvbb5RVXXvvfc231izZk3zjaqqDRs2NN+49NJLd4t7BC31fb9b3KP3ve99zTcmJiaab1x11VXNN6qqpk+fPsjOEGbOnNl849FHH93uPfLiBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAIIZ4/6A73znO4PsTJvWvhFPOOGE5htLly5tvlFV9dxzzzXfOPzww5tvVFX95Cc/GWRnnK688spBdr785S8333j88cebb3Rd13xjqrn44ovH/QkA/3bhhRc23/j973/ffOOZZ55pvlFVtW3btkF2hrBo0aKxbXtxAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAEHX9/3Iw3nz5o0+3EWef/751hNVVXXhhRc237jiiiuab7Dzuq5rvrFt27aRIyeffHLze/TpT3+69URVVZ155pmD7LB76vt+5D3quq75PYLWFi9e3HzjL3/5y25xjw499NDmG0899VTzDXbezTff3Hzj9NNP3+498uIEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQzNjR4cc+9rHmH/CFL3yh+UZV1RVXXDHIDpPPV77ylbHuL1mypPnGmWee2XyD3duKFSvG/Qnspvq+b77RdV3zjaqqAw44YJCd3cGpp57afMPfe5PT6aefPrZtL04AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgmLGjwz322KP5B/zsZz9rvsHkdNZZZw2y881vfrP5xqWXXjrybOPGjc332X2ddNJJg+xccMEFzTfOP//85hs7MnPmzEF2li9f3nzjtttua74xlK7rxv0Ju8zRRx897k+YMlasWDHuT2BMTjzxxOYbt99++3b/uRcnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAoOv7fvRh140+BP6t7/tu1Jl7NPnstddeg+zMnz+/+ca6deuabwzFPWKq22effZpvvPzyy+4R/8+8efOab/zrX/9qvlFV9cYbbzTfGHWPvDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAoOv7ftzfAAAAMKl5cQIAAAhm7Oiw6zrPUTvhkksuab5x+eWXN9+oqlq4cGHzjcMPP7z5RlXVhg0bmm/cfffd3aiz4447rvk9WrBgQeuJqqpatWpV841ly5Y131izZk3zDXZe3/cj75GfR/D2jPse7bPPPq0nqqrqXe96V/ONdevWNd8YygknnNB844477mi+UVU1e/bs5hubN2/e7j3y4gQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAACCru/70YddN/pwF/n85z/feqKqqqZNa9+I99xzT/ON+++/v/kGO6/v+27U2RD3aGJiovVEVVXdeeedzTe+8Y1vNN/Yb7/9mm9UVT388MPNN4b6O+Hxxx9vvvHCCy+MvEd77rln83u0ZcuW1hPs5mbNmtV8Y8uWLWP9eQRTwaj/XufFCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIOj6vh95eNppp40+3EVuvfXW1hNVVdV13SA77J76vh/5B6zruub3CKaCHd2js846q/k9uv7661tPQHM7ukdz585tfo9efPHF1hPQ3Kh75MUJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABDN2dHjIIYc0/4Cu65pvsPP+53/+p/nGjBk7/OO3y9x6662D7ADtXH/99eP+BP7D9OnTB9m57rrrmm8sW7as+UZV1RlnnDHIzigvvvjiWPeZ2vq+H2Tn+9///iA72+PFCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAARd3/ejD7tu9CHwb33fd6PO3KPJ56KLLhpk5/LLLx9kZ6pwj3adRYsWNd8499xzm29UVZ100knNN2bOnNl8o6rqqKOOGmLGPfovct555w2ys3LlyuYb1113XfONqqrVq1c33/jpT3+63XvkxQkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACDo+r4ffdh1ow/5f5YuXdp844EHHmi+wc7r+74bdeYe7ZxDDz10SmxUVd12223NN4477rjmG1VVd999d/ON3eUeXXzxxc03hvh5dM011zTfqKp64oknmm+sXbu2+UZV1be//e3mGxdffPFucY+gpVE/j7w4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgKDr+370YdeNPtxFpk0bpt2OOOKI5htHH310840rr7yy+cZU8+53v7v5xrPPPtuNOhviHg3lnHPOab7x6KOPNt9Ys2ZN842p5phjjmm+cd999+0W92j27NnNN958883mGxs2bGi+UVV1//33N9847bTTmm8Mpe/73eIefeITn2i+sWrVquYbTE6j7pEXJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEAgnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgGDGjg6/973vNf+AH/7wh803qqqeeeaZ5htr165tvsHO+9vf/jbW/QMPPLD5Rt/3zTeqqm688cbmG8cee2zzDXbehz70oXF/QnMHH3zwIDvLly9vvnH11Vc339h///2bb7BrTZvW/n8v37ZtW/ONqqpVq1YNsgP/lxcnAACAQDgBAAAEwgkAACAQTgAAAIFwAgAACIQTAABAIJwAAAAC4QQAABAIJwAAgEA4AQAABMIJAAAgEE4AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAIBAOAEAAATCCQAAIBBOAAAAgXACAAAIhBMAAEAgnAAAAALhBAAAEHR934/7GwAAACY1L04AAACBcAIAAAiEEwAAQCCcAAAAAuEEAAAQCCcAAPCZnPYAAAAISURBVIDgfwEi795bMkB4KwAAAABJRU5ErkJggg==\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "fig = plt.figure(figsize=(6,4))\n",
        "ax = fig.add_axes([0,0,1,1])\n",
        "ax.bar(np.arange(0,10), model_dummy.predict(one_image_data, verbose=False)[0])\n",
        "ax.set_title('Probability distribution', fontsize=20)\n",
        "ax.set_xticks(np.arange(0,10))\n",
        "ax.set_xticklabels(class_names, rotation=45, fontsize=14)\n",
        "ax.set_xlabel('Class name', fontsize=18)\n",
        "ax.set_ylabel('Probability', fontsize=18)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 435
        },
        "id": "q8vEQGQGXrYl",
        "outputId": "89d9312f-fdab-4009-f820-977bec74d79f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 7. Conclusions and final thoughts\n",
        "\n",
        "We can see that the processing degrees of the images are very different for the same slice of the network. Qualitative analysis indicates that the benchmark model is more aggressive when it comes to extracting useful information from the input. This is particularly evident from the first convolutional layer comparison: the frog image output is much less distorted and much more recognizable on the dummy than it is on the benchmark model.\n",
        "\n",
        "Also, the plots from the probability distributions show that the benchmark model is accurate when predicting the frog class, while the dummy classifier has a less-peaked distribution profile, and also has its greatest value in the wrong class. After the last convolutional layer, we see more active pixels in the dummy classifier model, while most of the pixels of the benchmark are black. These observations lead to indications of worse feature extraction performance.\n",
        "\n",
        "Further improvements to this analysis would be to use a dataset with larger images, with more pixels. This would allow us to see with more detail the feature extraction process. It wasn't possible to use larger images here, since the training process would take much longer.\n",
        "\n",
        "Another approach would be to stablish a checkpoint routine to save the weights at every epoch of training, and use the best epoch as benchmark, while using earlier epochs as the dummy classifier. Although I believe that the results would be similar, it is something worth trying.\n",
        "\n",
        "Finally, this project had the objective to look at qualitatively aspects of feature extraction, but it is possible to think about metrics that validate and make our reasonings more robust. Two immediate numerical analysis that come to my mind concern the last convolutional layer outputs:\n",
        "\n",
        "1. To calculate the average color of the outputs from the last convolutional layer. We expect the benchmark to have an average value closer to zero (black).\n",
        "2. To look at the correlation between the pixels. The correlation should be higher on the dummy classifier, since it (allegedly) contains more redundant information. \n"
      ],
      "metadata": {
        "id": "l5N2OVfxjKaK"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# References\n",
        "\n",
        "Some references I used through this work:\n",
        "\n",
        "1. Manjunath Jogin, et al. \"Feature Extraction using Convolution Neural Networks (CNN) and Deep Learning\". \n",
        "\n",
        "2. Simonyan, Karen and Zisserman, Andrew. \"Very Deep Convolution Networks for Large-Scale Image Recognition\".\n",
        "\n",
        "3. Zeiler, Matthew and Fergus, Rob. \"Visualizing and Understanding Convolutional Networks\".\n",
        "\n",
        "4. Bodapati, Jyostna and Veeranjaneyulu, N. \"Feature Extraction and Classification Using Deep Convolutional Neural Networks\".\n",
        "\n",
        "This were used as a theoretical background, as well as some implementation insights, specially at the architecture of the NN used. "
      ],
      "metadata": {
        "id": "IiywaveOhsTh"
      }
    }
  ]
}