A very simple convnet cats-dogs classifier.

dlearn.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "dlearn.ipynb",
      "provenance": [],
      "toc_visible": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nPoBZJu7kBdu"
      },
      "source": [
        "# Setting up data\n",
        "The cats and dogs dataset is from Kaggle, so to use the data in Google Colab, we do the following:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "a-3uswGhSjy4"
      },
      "source": [
        "# do some magic and insert your `kaggle.json` into `~/.kaggle`\r\n",
        "# colab may not be up-to-date on kaggle, maybe this command is necessary\r\n",
        "#!pip install --upgrade --force-reinstall --no-deps kaggle"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ncH7qjrvZHKS"
      },
      "source": [
        "!mkdir -p ~/.kaggle\r\n",
        "!echo '{\"username\":\"utkuboduroglu\",\"key\":\"\"}' > ~/.kaggle/kaggle.json\r\n",
        "!chmod 0600 ~/.kaggle/kaggle.json"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "i5dcSpUHZAOH",
        "outputId": "eb24b079-e182-495f-a1bb-724a199c5660"
      },
      "source": [
        "# apparently you can mount colab to drive and store persistent files there\r\n",
        "# you need to accept the terms & conditions of the challenge first\r\n",
        "!kaggle competitions download -c dogs-vs-cats"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Warning: Looks like you're using an outdated API Version, please consider updating (server 1.5.10 / client 1.5.4)\n",
            "Downloading train.zip to /content\n",
            " 98% 534M/543M [00:10<00:00, 60.9MB/s]\n",
            "100% 543M/543M [00:10<00:00, 55.2MB/s]\n",
            "Downloading test1.zip to /content\n",
            " 97% 264M/271M [00:04<00:00, 81.1MB/s]\n",
            "100% 271M/271M [00:05<00:00, 56.8MB/s]\n",
            "Downloading sampleSubmission.csv to /content\n",
            "  0% 0.00/86.8k [00:00<?, ?B/s]\n",
            "100% 86.8k/86.8k [00:00<00:00, 81.1MB/s]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BAmjVznlZEm5",
        "outputId": "eabd9a49-9eb7-4a63-db9b-99dc8e6fbd24"
      },
      "source": [
        "!pwd\r\n",
        "!ls"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/content\n",
            "sample_data  sampleSubmission.csv  test1.zip  train.zip\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "68xDgYIdb4la"
      },
      "source": [
        "#!unzip -q dogs-vs-cats.zip -d .\r\n",
        "!unzip -q train.zip -d .\r\n",
        "!unzip -q test1.zip -d .\r\n",
        "# you may need to do some shell magic from here to get the folders train/ and test1/"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eXHgkiAFb-Ne",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ca7e2932-ff2a-44e0-ee1d-32bef3f7cdc1"
      },
      "source": [
        "# everytime we run this cell, we need to delete this directory\r\n",
        "# or OS complains about the directory existing\r\n",
        "!rm -rf /content/cats_dogs_small\r\n",
        "\r\n",
        "import os, shutil\r\n",
        "original_dataset_dir = '/content/train'\r\n",
        "\r\n",
        "base_dir = '/content/cats_dogs_small'\r\n",
        "os.mkdir(base_dir)\r\n",
        "\r\n",
        "train_dir = os.path.join(base_dir, 'train')\r\n",
        "os.mkdir(train_dir)\r\n",
        "validation_dir = os.path.join(base_dir, 'validation')\r\n",
        "os.mkdir(validation_dir)\r\n",
        "test_dir = os.path.join(base_dir, 'test')\r\n",
        "os.mkdir(test_dir)\r\n",
        "\r\n",
        "train_cats_dir = os.path.join(train_dir, 'cats')\r\n",
        "os.mkdir(train_cats_dir)\r\n",
        "\r\n",
        "train_dogs_dir = os.path.join(train_dir, 'dogs')\r\n",
        "os.mkdir(train_dogs_dir)\r\n",
        "\r\n",
        "validation_cats_dir = os.path.join(validation_dir, 'cats')\r\n",
        "os.mkdir(validation_cats_dir)\r\n",
        "\r\n",
        "validation_dogs_dir = os.path.join(validation_dir, 'dogs')\r\n",
        "os.mkdir(validation_dogs_dir)\r\n",
        "\r\n",
        "test_cats_dir = os.path.join(test_dir, 'cats')\r\n",
        "os.mkdir(test_cats_dir)\r\n",
        "\r\n",
        "test_dogs_dir = os.path.join(test_dir, 'dogs')\r\n",
        "os.mkdir(test_dogs_dir)\r\n",
        "\r\n",
        "fnames = ['cat.{}.jpg'.format(i) for i in range(1000)]\r\n",
        "for fname in fnames:\r\n",
        "    src = os.path.join(original_dataset_dir, fname)\r\n",
        "    dst = os.path.join(train_cats_dir, fname)\r\n",
        "    shutil.copyfile(src,dst)\r\n",
        "\r\n",
        "fnames = ['cat.{}.jpg'.format(i) for i in range(1000, 1500)]\r\n",
        "for fname in fnames:\r\n",
        "    src = os.path.join(original_dataset_dir, fname)\r\n",
        "    dst = os.path.join(validation_cats_dir, fname)\r\n",
        "    shutil.copyfile(src,dst)\r\n",
        "\r\n",
        "fnames = ['cat.{}.jpg'.format(i) for i in range(1500, 2000)]\r\n",
        "for fname in fnames:\r\n",
        "    src = os.path.join(original_dataset_dir, fname)\r\n",
        "    dst = os.path.join(test_cats_dir, fname)\r\n",
        "    shutil.copyfile(src,dst)\r\n",
        "\r\n",
        "fnames = ['dog.{}.jpg'.format(i) for i in range(1000)]\r\n",
        "for fname in fnames:\r\n",
        "    src = os.path.join(original_dataset_dir, fname)\r\n",
        "    dst = os.path.join(train_dogs_dir, fname)\r\n",
        "    shutil.copyfile(src,dst)\r\n",
        "\r\n",
        "fnames = ['dog.{}.jpg'.format(i) for i in range(1000, 1500)]\r\n",
        "for fname in fnames:\r\n",
        "    src = os.path.join(original_dataset_dir, fname)\r\n",
        "    dst = os.path.join(validation_dogs_dir, fname)\r\n",
        "    shutil.copyfile(src,dst)\r\n",
        "\r\n",
        "fnames = ['dog.{}.jpg'.format(i) for i in range(1500, 2000)]\r\n",
        "for fname in fnames:\r\n",
        "    src = os.path.join(original_dataset_dir, fname)\r\n",
        "    dst = os.path.join(test_dogs_dir, fname)\r\n",
        "    shutil.copyfile(src,dst)\r\n",
        "\r\n",
        "print('total training cat images:', len(os.listdir(train_cats_dir)))\r\n",
        "print('total training dog images:', len(os.listdir(train_dogs_dir)))\r\n",
        "\r\n",
        "print('total validationing cat images:', len(os.listdir(validation_cats_dir)))\r\n",
        "print('total validationing dog images:', len(os.listdir(validation_dogs_dir)))\r\n",
        "\r\n",
        "print('total testing cat images:', len(os.listdir(test_cats_dir)))\r\n",
        "print('total testing dog images:', len(os.listdir(test_dogs_dir)))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "total training cat images: 1000\n",
            "total training dog images: 1000\n",
            "total validationing cat images: 500\n",
            "total validationing dog images: 500\n",
            "total testing cat images: 500\n",
            "total testing dog images: 500\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UZhtPf7NaTwu",
        "outputId": "6ac671e4-7383-43b7-eb43-9c84d3ec8f2c"
      },
      "source": [
        "!ls /content/cats_dogs_small"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "test  train  validation\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mDOdtF2TkNwV"
      },
      "source": [
        "# Creating a convnet binary classifier\n",
        "The dataset guarantees that we either have cats or dogs in our images. Thus it is enough to make a binary classifier with cat or dog.\n",
        "\n",
        "Our model takes a `150x150` RBG image and outputs a `bool`, signifying whether the image contains a cat or a dog. The model is a `convnet`. We optimize for accuracy, measure loss with binary cross-entropy and trained with backpropogation `RMSprop`."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "b0RNYf6fyCK6"
      },
      "source": [
        "from keras import models, layers\r\n",
        "from keras import optimizers\r\n",
        "\r\n",
        "model = models.Sequential()\r\n",
        "\r\n",
        "model.add(layers.Conv2D(32, (3, 3), activation='relu',\r\n",
        "                        input_shape=(150, 150, 3)))\r\n",
        "model.add(layers.MaxPooling2D((2, 2)))\r\n",
        "model.add(layers.Conv2D(64, (3, 3), activation='relu'))\r\n",
        "model.add(layers.MaxPooling2D((2, 2)))\r\n",
        "model.add(layers.Conv2D(128, (3, 3), activation='relu'))\r\n",
        "model.add(layers.MaxPooling2D((2, 2)))\r\n",
        "model.add(layers.Conv2D(128, (3, 3), activation='relu'))\r\n",
        "model.add(layers.MaxPooling2D((2, 2)))\r\n",
        "model.add(layers.Flatten())\r\n",
        "model.add(layers.Dropout(.5))\r\n",
        "model.add(layers.Dense(512, activation='relu'))\r\n",
        "model.add(layers.Dense(1, activation='sigmoid'))\r\n",
        "\r\n",
        "model.compile(loss='binary_crossentropy',\r\n",
        "              optimizer=optimizers.RMSprop(lr=1e-4),\r\n",
        "              metrics=['acc'])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SwYuwiQklNzf"
      },
      "source": [
        "## Image generation\n",
        "Since the problem is related to computer vision, we can use operations like scaling, translating, rotating etc. to populate our dataset with more images that makes our model transformation-invariant. Another benefit of this is that given a small dataset, we can create significantly more data."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NaxD3FKcyo0O",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "0d8671ff-59b4-4cc7-e1a6-b4e5b69e4f1f"
      },
      "source": [
        "from keras.preprocessing.image import ImageDataGenerator\r\n",
        "\r\n",
        "train_datagen = ImageDataGenerator(\r\n",
        "    rescale=1./255,\r\n",
        "    rotation_range=40,\r\n",
        "    width_shift_range=.2,\r\n",
        "    height_shift_range=.2,\r\n",
        "    shear_range=.2,\r\n",
        "    zoom_range=.2,\r\n",
        "    horizontal_flip=True,\r\n",
        "    fill_mode='nearest'\r\n",
        ")\r\n",
        "\r\n",
        "test_datagen = ImageDataGenerator(rescale=1./255)\r\n",
        "\r\n",
        "train_generator = train_datagen.flow_from_directory(\r\n",
        "    train_dir,\r\n",
        "    target_size = (150, 150),\r\n",
        "    batch_size = 32,\r\n",
        "    class_mode='binary'\r\n",
        ")\r\n",
        "\r\n",
        "validation_generator = test_datagen.flow_from_directory(\r\n",
        "    validation_dir,\r\n",
        "    target_size = (150, 150),\r\n",
        "    batch_size = 32,\r\n",
        "    class_mode='binary'\r\n",
        ")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 2000 images belonging to 2 classes.\n",
            "Found 1000 images belonging to 2 classes.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GJL1YAw4zCcB",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "df9a498e-7f46-4fd2-fdf3-8280cc911384"
      },
      "source": [
        "history = model.fit_generator(\r\n",
        "    train_generator,\r\n",
        "    steps_per_epoch=2000//32,\r\n",
        "    epochs=100,\r\n",
        "    validation_data=validation_generator,\r\n",
        "    validation_steps=1000//32\r\n",
        ")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "62/62 [==============================] - 111s 2s/step - loss: 0.3845 - acc: 0.8222 - val_loss: 0.4568 - val_acc: 0.8135\n",
            "Epoch 2/100\n",
            "62/62 [==============================] - 113s 2s/step - loss: 0.3913 - acc: 0.8211 - val_loss: 0.5180 - val_acc: 0.7712\n",
            "Epoch 3/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3974 - acc: 0.8140 - val_loss: 0.4022 - val_acc: 0.8276\n",
            "Epoch 4/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3858 - acc: 0.8155 - val_loss: 0.4875 - val_acc: 0.7954\n",
            "Epoch 5/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.4014 - acc: 0.8115 - val_loss: 0.4359 - val_acc: 0.7954\n",
            "Epoch 6/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3831 - acc: 0.8277 - val_loss: 0.4545 - val_acc: 0.7923\n",
            "Epoch 7/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3919 - acc: 0.8252 - val_loss: 0.4052 - val_acc: 0.8196\n",
            "Epoch 8/100\n",
            "62/62 [==============================] - 113s 2s/step - loss: 0.3864 - acc: 0.8181 - val_loss: 0.3867 - val_acc: 0.8276\n",
            "Epoch 9/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3864 - acc: 0.8277 - val_loss: 0.4294 - val_acc: 0.8095\n",
            "Epoch 10/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3877 - acc: 0.8293 - val_loss: 0.4098 - val_acc: 0.8236\n",
            "Epoch 11/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.4036 - acc: 0.8191 - val_loss: 0.3906 - val_acc: 0.8317\n",
            "Epoch 12/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3760 - acc: 0.8338 - val_loss: 0.4001 - val_acc: 0.8306\n",
            "Epoch 13/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3904 - acc: 0.8252 - val_loss: 0.3977 - val_acc: 0.8286\n",
            "Epoch 14/100\n",
            "62/62 [==============================] - 113s 2s/step - loss: 0.3843 - acc: 0.8206 - val_loss: 0.3850 - val_acc: 0.8407\n",
            "Epoch 15/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3821 - acc: 0.8237 - val_loss: 0.4217 - val_acc: 0.8276\n",
            "Epoch 16/100\n",
            "62/62 [==============================] - 113s 2s/step - loss: 0.3912 - acc: 0.8206 - val_loss: 0.3913 - val_acc: 0.8347\n",
            "Epoch 17/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3867 - acc: 0.8242 - val_loss: 0.4043 - val_acc: 0.8216\n",
            "Epoch 18/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3718 - acc: 0.8288 - val_loss: 0.4198 - val_acc: 0.8296\n",
            "Epoch 19/100\n",
            "62/62 [==============================] - 113s 2s/step - loss: 0.3767 - acc: 0.8267 - val_loss: 0.3873 - val_acc: 0.8367\n",
            "Epoch 20/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3712 - acc: 0.8343 - val_loss: 0.3894 - val_acc: 0.8387\n",
            "Epoch 21/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3631 - acc: 0.8333 - val_loss: 0.4119 - val_acc: 0.8317\n",
            "Epoch 22/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3753 - acc: 0.8343 - val_loss: 0.4014 - val_acc: 0.8317\n",
            "Epoch 23/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3650 - acc: 0.8455 - val_loss: 0.3951 - val_acc: 0.8397\n",
            "Epoch 24/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3890 - acc: 0.8308 - val_loss: 0.4294 - val_acc: 0.8175\n",
            "Epoch 25/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3684 - acc: 0.8364 - val_loss: 0.3922 - val_acc: 0.8286\n",
            "Epoch 26/100\n",
            "62/62 [==============================] - 110s 2s/step - loss: 0.3580 - acc: 0.8496 - val_loss: 0.5125 - val_acc: 0.7752\n",
            "Epoch 27/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3570 - acc: 0.8516 - val_loss: 0.4872 - val_acc: 0.7903\n",
            "Epoch 28/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3823 - acc: 0.8267 - val_loss: 0.4215 - val_acc: 0.8337\n",
            "Epoch 29/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3597 - acc: 0.8404 - val_loss: 0.4447 - val_acc: 0.8024\n",
            "Epoch 30/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3566 - acc: 0.8455 - val_loss: 0.4718 - val_acc: 0.8004\n",
            "Epoch 31/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3654 - acc: 0.8364 - val_loss: 0.3877 - val_acc: 0.8407\n",
            "Epoch 32/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3517 - acc: 0.8435 - val_loss: 0.3849 - val_acc: 0.8407\n",
            "Epoch 33/100\n",
            "62/62 [==============================] - 113s 2s/step - loss: 0.3622 - acc: 0.8349 - val_loss: 0.3977 - val_acc: 0.8367\n",
            "Epoch 34/100\n",
            "62/62 [==============================] - 110s 2s/step - loss: 0.3592 - acc: 0.8425 - val_loss: 0.4000 - val_acc: 0.8276\n",
            "Epoch 35/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3580 - acc: 0.8415 - val_loss: 0.4657 - val_acc: 0.7944\n",
            "Epoch 36/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3319 - acc: 0.8613 - val_loss: 0.4249 - val_acc: 0.8236\n",
            "Epoch 37/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3609 - acc: 0.8404 - val_loss: 0.3570 - val_acc: 0.8498\n",
            "Epoch 38/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3319 - acc: 0.8552 - val_loss: 0.4127 - val_acc: 0.8337\n",
            "Epoch 39/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3425 - acc: 0.8496 - val_loss: 0.3650 - val_acc: 0.8558\n",
            "Epoch 40/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3440 - acc: 0.8481 - val_loss: 0.3923 - val_acc: 0.8387\n",
            "Epoch 41/100\n",
            "62/62 [==============================] - 113s 2s/step - loss: 0.3458 - acc: 0.8450 - val_loss: 0.4036 - val_acc: 0.8246\n",
            "Epoch 42/100\n",
            "62/62 [==============================] - 110s 2s/step - loss: 0.3497 - acc: 0.8511 - val_loss: 0.4063 - val_acc: 0.8206\n",
            "Epoch 43/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3521 - acc: 0.8455 - val_loss: 0.4513 - val_acc: 0.8165\n",
            "Epoch 44/100\n",
            "62/62 [==============================] - 113s 2s/step - loss: 0.3324 - acc: 0.8567 - val_loss: 0.4500 - val_acc: 0.8044\n",
            "Epoch 45/100\n",
            "62/62 [==============================] - 113s 2s/step - loss: 0.3542 - acc: 0.8476 - val_loss: 0.3881 - val_acc: 0.8317\n",
            "Epoch 46/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3391 - acc: 0.8506 - val_loss: 0.4179 - val_acc: 0.8357\n",
            "Epoch 47/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3514 - acc: 0.8343 - val_loss: 0.4125 - val_acc: 0.8286\n",
            "Epoch 48/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3303 - acc: 0.8542 - val_loss: 0.3706 - val_acc: 0.8417\n",
            "Epoch 49/100\n",
            "62/62 [==============================] - 113s 2s/step - loss: 0.3552 - acc: 0.8450 - val_loss: 0.3895 - val_acc: 0.8407\n",
            "Epoch 50/100\n",
            "62/62 [==============================] - 113s 2s/step - loss: 0.3389 - acc: 0.8496 - val_loss: 0.4803 - val_acc: 0.8004\n",
            "Epoch 51/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3491 - acc: 0.8455 - val_loss: 0.4223 - val_acc: 0.8246\n",
            "Epoch 52/100\n",
            "62/62 [==============================] - 113s 2s/step - loss: 0.3336 - acc: 0.8582 - val_loss: 0.3671 - val_acc: 0.8468\n",
            "Epoch 53/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3339 - acc: 0.8532 - val_loss: 0.4395 - val_acc: 0.8296\n",
            "Epoch 54/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3258 - acc: 0.8567 - val_loss: 0.4277 - val_acc: 0.8165\n",
            "Epoch 55/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3294 - acc: 0.8526 - val_loss: 0.3912 - val_acc: 0.8438\n",
            "Epoch 56/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3299 - acc: 0.8526 - val_loss: 0.4400 - val_acc: 0.8135\n",
            "Epoch 57/100\n",
            "62/62 [==============================] - 110s 2s/step - loss: 0.3386 - acc: 0.8516 - val_loss: 0.3968 - val_acc: 0.8357\n",
            "Epoch 58/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3232 - acc: 0.8598 - val_loss: 0.3780 - val_acc: 0.8488\n",
            "Epoch 59/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3333 - acc: 0.8664 - val_loss: 0.3797 - val_acc: 0.8629\n",
            "Epoch 60/100\n",
            "62/62 [==============================] - 115s 2s/step - loss: 0.3299 - acc: 0.8537 - val_loss: 0.4262 - val_acc: 0.8236\n",
            "Epoch 61/100\n",
            "62/62 [==============================] - 113s 2s/step - loss: 0.3131 - acc: 0.8613 - val_loss: 0.4611 - val_acc: 0.8145\n",
            "Epoch 62/100\n",
            "62/62 [==============================] - 113s 2s/step - loss: 0.3210 - acc: 0.8633 - val_loss: 0.3992 - val_acc: 0.8266\n",
            "Epoch 63/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3259 - acc: 0.8603 - val_loss: 0.4553 - val_acc: 0.7984\n",
            "Epoch 64/100\n",
            "62/62 [==============================] - 113s 2s/step - loss: 0.3259 - acc: 0.8592 - val_loss: 0.4005 - val_acc: 0.8397\n",
            "Epoch 65/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3209 - acc: 0.8618 - val_loss: 0.4404 - val_acc: 0.8337\n",
            "Epoch 66/100\n",
            "62/62 [==============================] - 113s 2s/step - loss: 0.3265 - acc: 0.8598 - val_loss: 0.4996 - val_acc: 0.8145\n",
            "Epoch 67/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3179 - acc: 0.8643 - val_loss: 0.4144 - val_acc: 0.8196\n",
            "Epoch 68/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3320 - acc: 0.8552 - val_loss: 0.3831 - val_acc: 0.8458\n",
            "Epoch 69/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3184 - acc: 0.8592 - val_loss: 0.4764 - val_acc: 0.8196\n",
            "Epoch 70/100\n",
            "62/62 [==============================] - 114s 2s/step - loss: 0.3205 - acc: 0.8577 - val_loss: 0.3805 - val_acc: 0.8448\n",
            "Epoch 71/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3230 - acc: 0.8664 - val_loss: 0.4240 - val_acc: 0.8306\n",
            "Epoch 72/100\n",
            "62/62 [==============================] - 113s 2s/step - loss: 0.3144 - acc: 0.8643 - val_loss: 0.3939 - val_acc: 0.8427\n",
            "Epoch 73/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3132 - acc: 0.8664 - val_loss: 0.4019 - val_acc: 0.8327\n",
            "Epoch 74/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.2956 - acc: 0.8806 - val_loss: 0.4516 - val_acc: 0.8276\n",
            "Epoch 75/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.2902 - acc: 0.8750 - val_loss: 0.3940 - val_acc: 0.8478\n",
            "Epoch 76/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3241 - acc: 0.8577 - val_loss: 0.3683 - val_acc: 0.8468\n",
            "Epoch 77/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.2998 - acc: 0.8649 - val_loss: 0.4372 - val_acc: 0.8347\n",
            "Epoch 78/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.2971 - acc: 0.8740 - val_loss: 0.4054 - val_acc: 0.8478\n",
            "Epoch 79/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.2931 - acc: 0.8730 - val_loss: 0.3819 - val_acc: 0.8508\n",
            "Epoch 80/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3028 - acc: 0.8709 - val_loss: 0.4598 - val_acc: 0.8266\n",
            "Epoch 81/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.2996 - acc: 0.8740 - val_loss: 0.3866 - val_acc: 0.8548\n",
            "Epoch 82/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3033 - acc: 0.8694 - val_loss: 0.4967 - val_acc: 0.7893\n",
            "Epoch 83/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3001 - acc: 0.8714 - val_loss: 0.3707 - val_acc: 0.8468\n",
            "Epoch 84/100\n",
            "62/62 [==============================] - 113s 2s/step - loss: 0.3058 - acc: 0.8643 - val_loss: 0.3823 - val_acc: 0.8518\n",
            "Epoch 85/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.2981 - acc: 0.8633 - val_loss: 0.4105 - val_acc: 0.8306\n",
            "Epoch 86/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3130 - acc: 0.8613 - val_loss: 0.3528 - val_acc: 0.8478\n",
            "Epoch 87/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3025 - acc: 0.8628 - val_loss: 0.4115 - val_acc: 0.8317\n",
            "Epoch 88/100\n",
            "62/62 [==============================] - 110s 2s/step - loss: 0.2910 - acc: 0.8720 - val_loss: 0.4211 - val_acc: 0.8347\n",
            "Epoch 89/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.2992 - acc: 0.8653 - val_loss: 0.4343 - val_acc: 0.8347\n",
            "Epoch 90/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.3064 - acc: 0.8740 - val_loss: 0.3759 - val_acc: 0.8468\n",
            "Epoch 91/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3104 - acc: 0.8653 - val_loss: 0.4136 - val_acc: 0.8185\n",
            "Epoch 92/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.2953 - acc: 0.8745 - val_loss: 0.4198 - val_acc: 0.8357\n",
            "Epoch 93/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.2868 - acc: 0.8750 - val_loss: 0.3802 - val_acc: 0.8508\n",
            "Epoch 94/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.2781 - acc: 0.8831 - val_loss: 0.4123 - val_acc: 0.8387\n",
            "Epoch 95/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.3017 - acc: 0.8684 - val_loss: 0.4508 - val_acc: 0.8306\n",
            "Epoch 96/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.2922 - acc: 0.8806 - val_loss: 0.4171 - val_acc: 0.8306\n",
            "Epoch 97/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.2795 - acc: 0.8816 - val_loss: 0.5082 - val_acc: 0.8226\n",
            "Epoch 98/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.2865 - acc: 0.8714 - val_loss: 0.3974 - val_acc: 0.8377\n",
            "Epoch 99/100\n",
            "62/62 [==============================] - 112s 2s/step - loss: 0.2950 - acc: 0.8811 - val_loss: 0.5000 - val_acc: 0.8165\n",
            "Epoch 100/100\n",
            "62/62 [==============================] - 111s 2s/step - loss: 0.2798 - acc: 0.8770 - val_loss: 0.3879 - val_acc: 0.8579\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FuSiQOX6zagK",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6bab21ff-2600-4877-aa2f-78a81e73e830"
      },
      "source": [
        "history.history.keys()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dict_keys(['loss', 'acc', 'val_loss', 'val_acc'])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2yZkwQYvEHWP"
      },
      "source": [
        "#model.save('cats_dogs_small_2.h5')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7mK32dubEN_x"
      },
      "source": [
        "!ls\r\n",
        "!pwd\r\n",
        "from google.colab import files\r\n",
        "files.download('cats_dogs_small_2.h5')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 545
        },
        "id": "jPTdN_mEDsdt",
        "outputId": "26b16401-2b62-4ea8-dd46-8341ce0e365c"
      },
      "source": [
        "import matplotlib.pyplot as plt\r\n",
        "import numpy as np\r\n",
        "acc = history.history['acc']\r\n",
        "val_acc = history.history['val_acc']\r\n",
        "\r\n",
        "loss = history.history['loss']\r\n",
        "val_loss = history.history['val_loss']\r\n",
        "\r\n",
        "epochs = np.array(range(1, len(acc)+1))\r\n",
        "\r\n",
        "plt.plot(epochs, acc, 'ro', label='Training accuracy')\r\n",
        "plt.plot(epochs, val_acc, 'r', label='Validation accuracy')\r\n",
        "plt.title('Training/Validation accuracy')\r\n",
        "plt.legend()\r\n",
        "\r\n",
        "plt.figure()\r\n",
        "\r\n",
        "plt.plot(epochs, loss, 'go', label='Training loss')\r\n",
        "plt.plot(epochs, val_loss, 'g', label='Validation loss')\r\n",
        "plt.title('Training/Validation loss')\r\n",
        "plt.legend()\r\n",
        "\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TvIEzs97-7Fj",
        "outputId": "143ae4d4-c72f-4cb2-e963-64a247fde3af"
      },
      "source": [
        "test_generator = test_datagen.flow_from_directory(\n",
        "    test_dir,\n",
        "    target_size = (150, 150),\n",
        "    batch_size = 32,\n",
        "    class_mode='binary'\n",
        ")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 1000 images belonging to 2 classes.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2McilM09AZog"
      },
      "source": [
        "#files.upload()\r\n",
        "model = models.load_model('cats_dogs_small_2(1).h5')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RcUkAWnglxTp"
      },
      "source": [
        "# Manual model testing\n",
        "Although the model is incomplete, we can still test the model on unseen data to see its performance. We see that picking random images from the test set shows that although the model works relatively well on dogs, it misclassifies a lot of cats."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GNj2q06IAjbL",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 287
        },
        "outputId": "149e8c28-109e-4206-cba7-689b7e605d95"
      },
      "source": [
        "# add some image of a cat from the test set and plot it,\r\n",
        "# and later\r\n",
        "from keras.preprocessing import image\r\n",
        "import matplotlib.pyplot as plt\r\n",
        "import os\r\n",
        "import numpy as np\r\n",
        "\r\n",
        "file_name = !ls cats_dogs_small/test/dogs/ -1 | sort -R | head -n1\r\n",
        "file_name = ''.join(file_name)\r\n",
        "\r\n",
        "img = image.load_img(os.path.join(test_dogs_dir, file_name), target_size=(150, 150))\r\n",
        "\r\n",
        "plt.figure()\r\n",
        "plt.imshow(img)\r\n",
        "plt.show()\r\n",
        "\r\n",
        "img_arr = image.img_to_array(img)\r\n",
        "test_arr = np.array([img_arr])\r\n",
        "\r\n",
        "pred = model.predict(test_arr)\r\n",
        "pred[0]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([1.], dtype=float32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 103
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qglHxD6BE7e8",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f01b3f3c-b987-444f-fa65-692056527d9c"
      },
      "source": [
        "test_generator.class_indices"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'cats': 0, 'dogs': 1}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 96
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-1qUW0HSHHSa",
        "outputId": "a68569dc-27a8-42d4-cd02-fa5d4ee126ce"
      },
      "source": [
        "test_hist = model.fit_generator(\r\n",
        "    test_generator\r\n",
        ")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/engine/training.py:1844: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n",
            "  warnings.warn('`Model.fit_generator` is deprecated and '\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "32/32 [==============================] - 46s 1s/step - loss: 0.3747 - acc: 0.8450\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZWGtknTomM03"
      },
      "source": [
        "## Results\n",
        "After fitting the model on the test set for one epoch, we see that the model performs like it did on the validation set, meaning that the data is not overfit; rather, the issue is the loss, which is at 40%~, probably because of misclassification when it comes to cats."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DWdeAWrqH2VA",
        "outputId": "183cb861-0463-42e7-e8e4-ed1da3fe09b9"
      },
      "source": [
        "result = test_hist.history\r\n",
        "print('loss: {}%\\naccu: {}%'.format(100*np.average(result['loss']), 100*np.average(result['acc'])))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "loss: 37.47009336948395%\n",
            "accu: 84.50000286102295%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-DXrT3hdIIdo"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}