week8/Lecture_week_8.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "a7d4e633",
   "metadata": {},
   "source": [
    "# Deep Learning\n",
    "\n",
    "* This week, we'll learn about neural nets and build a model for classifying images of clothes\n",
    "* We will use Tensorflow"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a995b032",
   "metadata": {},
   "source": [
    "## Data: Fashion Classification\n",
    "\n",
    "* https://www.kaggle.com/datasets/zalando-research/fashionmnist"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "00b4b75c",
   "metadata": {},
   "source": [
    "## Tensorflow and Keras\n",
    "\n",
    "* Tensorflow is a framework for deep learning \n",
    "* Keras is a higher level abstraction on top of tensorflow\n",
    "    * Makes it simpler to create and train neural networks\n",
    "* Downloading Tensorflow\n",
    "    * using conda: ```conda install tensorflow```\n",
    "* Loading the images"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "a4a3b459",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras.preprocessing.image import load_img"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "b86ee2a7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'2.4.1'"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tf.__version__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "909b6b03",
   "metadata": {},
   "outputs": [],
   "source": [
    "image_path = \"../data/clothing-dataset-small/train/t-shirt\"\n",
    "name = \"5f0a3fa0-6a3d-4b68-b213-72766a643de7.jpg\"\n",
    "full_name = f'{image_path}/{name}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "baaff277",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=400x533 at 0x7F90A87E86A0>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "load_img(full_name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "6c5c01da",
   "metadata": {},
   "outputs": [],
   "source": [
    "# reduce target size for neural network\n",
    "img = load_img(full_name, target_size=(299, 299))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "1ae6f818",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(299, 299, 3)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# turn image into numpy array\n",
    "x = np.array(img)\n",
    "x.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "75f53ee1",
   "metadata": {},
   "source": [
    "## Pre-trained Convolutional Neural Network\n",
    "\n",
    "* Use a network that was already trained on some data\n",
    "* Imagenet dataset: https://image-net.org\n",
    "* Pre-trained models: https://keras.io/api/applications\n",
    "    * We will use the Xception model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "8c393b5d",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.applications.xception import Xception\n",
    "from tensorflow.keras.applications.xception import preprocess_input\n",
    "from tensorflow.keras.applications.xception import decode_predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "19ca72ee",
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Xception(weights='imagenet', input_shape=(299,299,3))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9b6c774d",
   "metadata": {},
   "source": [
    "* Use this model now to classify the image of the t-shirt above\n",
    "* THis model does not expect a single image, but a couple of images"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "198a9ec5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 299, 299, 3)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = np.array([x])\n",
    "X.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "885ee5f7",
   "metadata": {},
   "source": [
    "* Before applying the model we need to do some preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "235de3c3",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = preprocess_input(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "41b4fe20",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 1000)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred = model.predict(X)\n",
    "pred.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5a78d197",
   "metadata": {},
   "source": [
    "* The output conatins the propabilty of the image of being in a specific class\n",
    "* In total there are 1000 classes\n",
    "* In order to understand the output, we need to now which are the classes\n",
    "* For that we use the decode_predictions function, which gives the best results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "dfe6ae93",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[('n03595614', 'jersey', 0.679245),\n",
       "  ('n02916936', 'bulletproof_vest', 0.03960029),\n",
       "  ('n04370456', 'sweatshirt', 0.035299525),\n",
       "  ('n03710637', 'maillot', 0.010884136),\n",
       "  ('n04525038', 'velvet', 0.0018057585)]]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "decode_predictions(pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9e60f434",
   "metadata": {},
   "source": [
    "* Imagenet does not have a class \"t-shirt\"\n",
    "* To predict the clothing classes we have, we need a different model for that\n",
    "* We can use the Xceptions model and adjust it for our purpose\n",
    "    * We don't have to train it from scratch"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "043a6866",
   "metadata": {},
   "source": [
    "## Convolutional Neural Networks\n",
    "* Mainly consists of convolutional and dense layers\n",
    "    * A convolutional layer consists of filters (of simple shapes) and it is calculated how similar the pixels are to this filter.\n",
    "    * These filters produce feature maps, which go to the next layer\n",
    "    * In each layer the neural net can learn more complicated features\n",
    "* After passing the image through the convolutional filters, we get a vector representation of the image\n",
    "    * This vector representation is passed through a set of dense layers, which then give the prediction\n",
    "    * The prediction is given as a propability\n",
    "    * Depending on the problem, the last layer could be a \n",
    "        * Sigmoid layer for binary classification\n",
    "        * a Softmax layer for multiple classification\n",
    "* REcommended website: https://cs231n.github.io"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5d5913f1",
   "metadata": {},
   "source": [
    "## Transfer Learning\n",
    "* Use a network that was trained on the imagente dataset\n",
    "![transer_learning](Screenshot_1.png)\n",
    "\n",
    "\n",
    "* We will use the convolutioal part of the network that converts the image into a vector representation.\n",
    "* Read the data using ```ImageDataGenerator```\n",
    "* Train ```Xception``` on smaller images (150 x 150)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "46415555",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "1fc874ff",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 3068 images belonging to 10 classes.\n"
     ]
    }
   ],
   "source": [
    "train_gen = ImageDataGenerator(preprocessing_function=preprocess_input)\n",
    "\n",
    "# read images from directory\n",
    "train_ds = train_gen.flow_from_directory('../data/clothing-dataset-small/train', \n",
    "                                         target_size=(150, 150), \n",
    "                                         batch_size=32)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "25773bc8",
   "metadata": {},
   "source": [
    "* The input data size in (32, 150, 150, 3) [(batch_size, width, height, channels)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "d2224a6c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'dress': 0,\n",
       " 'hat': 1,\n",
       " 'longsleeve': 2,\n",
       " 'outwear': 3,\n",
       " 'pants': 4,\n",
       " 'shirt': 5,\n",
       " 'shoes': 6,\n",
       " 'shorts': 7,\n",
       " 'skirt': 8,\n",
       " 't-shirt': 9}"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_ds.class_indices"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "e60fd95e",
   "metadata": {},
   "outputs": [],
   "source": [
    "# look at one image and target\n",
    "X, y = next(train_ds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "30543595",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(32, 150, 150, 3)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "2faf34a2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(32, 10)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "271a0118",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0., 0., 0., 0., 0., 0., 1., 0., 0., 0.],\n",
       "       [1., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
       "       [0., 0., 0., 0., 0., 0., 0., 0., 0., 1.],\n",
       "       [0., 0., 1., 0., 0., 0., 0., 0., 0., 0.],\n",
       "       [0., 0., 0., 0., 0., 0., 0., 1., 0., 0.]], dtype=float32)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y[:5] # multiclasss classification"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4773c13b",
   "metadata": {},
   "source": [
    "* Now do the same for the validation dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "518daa1d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 341 images belonging to 10 classes.\n"
     ]
    }
   ],
   "source": [
    "val_gen = ImageDataGenerator(preprocessing_function=preprocess_input)\n",
    "\n",
    "# read images from directory\n",
    "val_ds = val_gen.flow_from_directory('../data/clothing-dataset-small/validation', \n",
    "                                         target_size=(150, 150), \n",
    "                                         batch_size=32,\n",
    "                                         shuffle=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "03235c00",
   "metadata": {},
   "source": [
    "![transfer_learning](Screenshot_2.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9100059b",
   "metadata": {},
   "source": [
    "![nn in keras](Screenshot_3.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "921ba136",
   "metadata": {},
   "source": [
    "* Create the base model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "5cbe8e4d",
   "metadata": {},
   "outputs": [],
   "source": [
    "base_model = Xception(weights='imagenet', \n",
    "                      include_top=False, \n",
    "                      input_shape=(150, 150, 3))\n",
    "\n",
    "base_model.trainable = False"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b2a64224",
   "metadata": {},
   "source": [
    "* Create a new top"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "8c5e7020",
   "metadata": {},
   "outputs": [],
   "source": [
    "inputs = keras.Input(shape=(150, 150, 3))\n",
    "\n",
    "base = base_model(inputs)\n",
    "\n",
    "outputs = base\n",
    "\n",
    "model = keras.Model(inputs, outputs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "801db2d9",
   "metadata": {},
   "outputs": [],
   "source": [
    "preds = model.predict(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "af8a855e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(32, 5, 5, 2048)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "preds.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dad9d1f7",
   "metadata": {},
   "source": [
    "* Turn this output into a vector\n",
    "\n",
    "![pooling](Screenshot_4.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "f9aa47e2",
   "metadata": {},
   "outputs": [],
   "source": [
    "inputs = keras.Input(shape=(150, 150, 3))\n",
    "\n",
    "base = base_model(inputs)\n",
    "pooling = keras.layers.GlobalAveragePooling2D()\n",
    "vectors = pooling(base)\n",
    "\n",
    "outputs = vectors\n",
    "\n",
    "model = keras.Model(inputs, outputs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "632962b3",
   "metadata": {},
   "outputs": [],
   "source": [
    "preds = model.predict(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "1a9319d4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(32, 2048)"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "preds.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "37c96362",
   "metadata": {},
   "source": [
    "* Add the output layer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "26d9d04b",
   "metadata": {},
   "outputs": [],
   "source": [
    "inputs = keras.Input(shape=(150, 150, 3))\n",
    "\n",
    "base = base_model(inputs, training=False)\n",
    "vectors = keras.layers.GlobalAveragePooling2D()(base)\n",
    "\n",
    "outputs = keras.layers.Dense(10)(vectors)\n",
    "\n",
    "model = keras.Model(inputs, outputs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "fa5c2de0",
   "metadata": {},
   "outputs": [],
   "source": [
    "preds = model.predict(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "cdd5c4e1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(32, 10)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "preds.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "afd71497",
   "metadata": {},
   "source": [
    "* This model still has to be trained!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "ff542a73",
   "metadata": {},
   "outputs": [],
   "source": [
    "learning_rate = 0.1\n",
    "optimizer = keras.optimizers.Adam(learning_rate=learning_rate)\n",
    "\n",
    "loss = keras.losses.CategoricalCrossentropy(from_logits=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "f1d14848",
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(optimizer=optimizer, loss=loss, metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "a8c2e4f9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "96/96 [==============================] - 138s 1s/step - loss: 15.3939 - accuracy: 0.5389 - val_loss: 5.3118 - val_accuracy: 0.7185\n",
      "Epoch 2/10\n",
      "96/96 [==============================] - 140s 1s/step - loss: 4.5629 - accuracy: 0.7861 - val_loss: 6.3982 - val_accuracy: 0.7361\n",
      "Epoch 3/10\n",
      "96/96 [==============================] - 148s 2s/step - loss: 2.8703 - accuracy: 0.8440 - val_loss: 6.6066 - val_accuracy: 0.7918\n",
      "Epoch 4/10\n",
      "96/96 [==============================] - 141s 1s/step - loss: 2.5001 - accuracy: 0.8680 - val_loss: 8.5447 - val_accuracy: 0.7771\n",
      "Epoch 5/10\n",
      "96/96 [==============================] - 109s 1s/step - loss: 2.4014 - accuracy: 0.8815 - val_loss: 7.5385 - val_accuracy: 0.7947\n",
      "Epoch 6/10\n",
      "96/96 [==============================] - 119s 1s/step - loss: 1.8285 - accuracy: 0.9050 - val_loss: 8.2347 - val_accuracy: 0.7918\n",
      "Epoch 7/10\n",
      "96/96 [==============================] - 114s 1s/step - loss: 1.5143 - accuracy: 0.9106 - val_loss: 8.9110 - val_accuracy: 0.8123\n",
      "Epoch 8/10\n",
      "96/96 [==============================] - 127s 1s/step - loss: 1.2894 - accuracy: 0.9293 - val_loss: 12.1574 - val_accuracy: 0.7625\n",
      "Epoch 9/10\n",
      "96/96 [==============================] - 105s 1s/step - loss: 1.1891 - accuracy: 0.9378 - val_loss: 10.4875 - val_accuracy: 0.7801\n",
      "Epoch 10/10\n",
      "96/96 [==============================] - 109s 1s/step - loss: 0.7999 - accuracy: 0.9516 - val_loss: 10.3378 - val_accuracy: 0.7683\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(train_ds, epochs=10, validation_data=val_ds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "2c80447a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['accuracy'], label=\"train\")\n",
    "plt.plot(history.history['val_accuracy'], label=\"validation\")\n",
    "plt.legend();"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "66e470fb",
   "metadata": {},
   "source": [
    "## Adjusting the Learning Rate\n",
    "* Too high learning rate may leas to overfitting, too low learning rate may take very long\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "3942cbf3",
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_model(learning_rate=.1):\n",
    "    base_model = Xception(weights='imagenet', \n",
    "                      include_top=False, \n",
    "                      input_shape=(150, 150, 3))\n",
    "\n",
    "    base_model.trainable = False\n",
    "    \n",
    "    ##############################\n",
    "    \n",
    "    inputs = keras.Input(shape=(150, 150, 3))\n",
    "    base = base_model(inputs, training=False)\n",
    "    vectors = keras.layers.GlobalAveragePooling2D()(base)\n",
    "    outputs = keras.layers.Dense(10)(vectors)\n",
    "    model = keras.Model(inputs, outputs)\n",
    "\n",
    "    ##############################\n",
    "    \n",
    "    optimizer = keras.optimizers.Adam(learning_rate=learning_rate)\n",
    "    loss = keras.losses.CategoricalCrossentropy(from_logits=True)\n",
    "    model.compile(optimizer=optimizer, loss=loss, metrics=['accuracy'])\n",
    "    \n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "067716d7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0001\n",
      "Epoch 1/10\n",
      "96/96 [==============================] - 110s 1s/step - loss: 2.1608 - accuracy: 0.2484 - val_loss: 1.5446 - val_accuracy: 0.5132\n",
      "Epoch 2/10\n",
      "96/96 [==============================] - 120s 1s/step - loss: 1.4526 - accuracy: 0.5399 - val_loss: 1.1996 - val_accuracy: 0.6246\n",
      "Epoch 3/10\n",
      "96/96 [==============================] - 105s 1s/step - loss: 1.1589 - accuracy: 0.6444 - val_loss: 1.0249 - val_accuracy: 0.7038\n",
      "Epoch 4/10\n",
      "96/96 [==============================] - 148s 2s/step - loss: 0.9971 - accuracy: 0.6867 - val_loss: 0.9208 - val_accuracy: 0.7331\n",
      "Epoch 5/10\n",
      "96/96 [==============================] - 155s 2s/step - loss: 0.8982 - accuracy: 0.7220 - val_loss: 0.8507 - val_accuracy: 0.7507\n",
      "Epoch 6/10\n",
      "96/96 [==============================] - 148s 2s/step - loss: 0.8465 - accuracy: 0.7307 - val_loss: 0.8041 - val_accuracy: 0.7654\n",
      "Epoch 7/10\n",
      "96/96 [==============================] - 149s 2s/step - loss: 0.7852 - accuracy: 0.7489 - val_loss: 0.7607 - val_accuracy: 0.7683\n",
      "Epoch 8/10\n",
      "96/96 [==============================] - 164s 2s/step - loss: 0.7323 - accuracy: 0.7640 - val_loss: 0.7360 - val_accuracy: 0.7889\n",
      "Epoch 9/10\n",
      "96/96 [==============================] - 137s 1s/step - loss: 0.7226 - accuracy: 0.7635 - val_loss: 0.7102 - val_accuracy: 0.7977\n",
      "Epoch 10/10\n",
      "96/96 [==============================] - 107s 1s/step - loss: 0.6843 - accuracy: 0.7721 - val_loss: 0.6947 - val_accuracy: 0.8006\n",
      "0.001\n",
      "Epoch 1/10\n",
      "96/96 [==============================] - 105s 1s/step - loss: 1.3725 - accuracy: 0.5606 - val_loss: 0.7203 - val_accuracy: 0.7771\n",
      "Epoch 2/10\n",
      "96/96 [==============================] - 162s 2s/step - loss: 0.6661 - accuracy: 0.7731 - val_loss: 0.6128 - val_accuracy: 0.8065\n",
      "Epoch 3/10\n",
      "96/96 [==============================] - 162s 2s/step - loss: 0.5102 - accuracy: 0.8226 - val_loss: 0.5926 - val_accuracy: 0.8123\n",
      "Epoch 4/10\n",
      "96/96 [==============================] - 151s 2s/step - loss: 0.4200 - accuracy: 0.8674 - val_loss: 0.5627 - val_accuracy: 0.8065\n",
      "Epoch 5/10\n",
      "96/96 [==============================] - 164s 2s/step - loss: 0.3701 - accuracy: 0.8891 - val_loss: 0.5473 - val_accuracy: 0.8094\n",
      "Epoch 6/10\n",
      "96/96 [==============================] - 165s 2s/step - loss: 0.3335 - accuracy: 0.8976 - val_loss: 0.5399 - val_accuracy: 0.8211\n",
      "Epoch 7/10\n",
      "96/96 [==============================] - 149s 2s/step - loss: 0.2782 - accuracy: 0.9308 - val_loss: 0.5489 - val_accuracy: 0.8358\n",
      "Epoch 8/10\n",
      "96/96 [==============================] - 108s 1s/step - loss: 0.2506 - accuracy: 0.9305 - val_loss: 0.5358 - val_accuracy: 0.8240\n",
      "Epoch 9/10\n",
      "96/96 [==============================] - 124s 1s/step - loss: 0.2277 - accuracy: 0.9461 - val_loss: 0.5409 - val_accuracy: 0.8299\n",
      "Epoch 10/10\n",
      "96/96 [==============================] - 126s 1s/step - loss: 0.2137 - accuracy: 0.9503 - val_loss: 0.5327 - val_accuracy: 0.8240\n",
      "0.01\n",
      "Epoch 1/10\n",
      "96/96 [==============================] - 143s 1s/step - loss: 1.5319 - accuracy: 0.5898 - val_loss: 0.9636 - val_accuracy: 0.7595\n",
      "Epoch 2/10\n",
      "96/96 [==============================] - 124s 1s/step - loss: 0.6232 - accuracy: 0.8195 - val_loss: 0.9809 - val_accuracy: 0.7654\n",
      "Epoch 3/10\n",
      "96/96 [==============================] - 124s 1s/step - loss: 0.3283 - accuracy: 0.8913 - val_loss: 0.9256 - val_accuracy: 0.7771\n",
      "Epoch 4/10\n",
      "96/96 [==============================] - 122s 1s/step - loss: 0.1667 - accuracy: 0.9446 - val_loss: 0.8457 - val_accuracy: 0.8006\n",
      "Epoch 5/10\n",
      "96/96 [==============================] - 123s 1s/step - loss: 0.1468 - accuracy: 0.9450 - val_loss: 0.7830 - val_accuracy: 0.8123\n",
      "Epoch 6/10\n",
      "96/96 [==============================] - 122s 1s/step - loss: 0.1059 - accuracy: 0.9598 - val_loss: 0.8441 - val_accuracy: 0.8035\n",
      "Epoch 7/10\n",
      "96/96 [==============================] - 122s 1s/step - loss: 0.0935 - accuracy: 0.9729 - val_loss: 0.9329 - val_accuracy: 0.8152\n",
      "Epoch 8/10\n",
      "96/96 [==============================] - 128s 1s/step - loss: 0.0901 - accuracy: 0.9725 - val_loss: 0.8651 - val_accuracy: 0.8065\n",
      "Epoch 9/10\n",
      "96/96 [==============================] - 121s 1s/step - loss: 0.0330 - accuracy: 0.9943 - val_loss: 0.9223 - val_accuracy: 0.8123\n",
      "Epoch 10/10\n",
      "96/96 [==============================] - 123s 1s/step - loss: 0.0393 - accuracy: 0.9867 - val_loss: 0.8856 - val_accuracy: 0.8123\n",
      "0.1\n",
      "Epoch 1/10\n",
      "96/96 [==============================] - 126s 1s/step - loss: 12.9233 - accuracy: 0.5521 - val_loss: 9.9929 - val_accuracy: 0.6921\n",
      "Epoch 2/10\n",
      "96/96 [==============================] - 122s 1s/step - loss: 4.1675 - accuracy: 0.8024 - val_loss: 6.3967 - val_accuracy: 0.7713\n",
      "Epoch 3/10\n",
      "96/96 [==============================] - 132s 1s/step - loss: 3.3863 - accuracy: 0.8264 - val_loss: 7.6026 - val_accuracy: 0.7801\n",
      "Epoch 4/10\n",
      "96/96 [==============================] - 131s 1s/step - loss: 2.9719 - accuracy: 0.8628 - val_loss: 8.8804 - val_accuracy: 0.7595\n",
      "Epoch 5/10\n",
      "96/96 [==============================] - 126s 1s/step - loss: 2.2462 - accuracy: 0.8929 - val_loss: 7.8951 - val_accuracy: 0.7977\n",
      "Epoch 6/10\n",
      "96/96 [==============================] - 127s 1s/step - loss: 1.7516 - accuracy: 0.9039 - val_loss: 9.5018 - val_accuracy: 0.7654\n",
      "Epoch 7/10\n",
      "96/96 [==============================] - 120s 1s/step - loss: 1.0362 - accuracy: 0.9335 - val_loss: 10.8085 - val_accuracy: 0.7625\n",
      "Epoch 8/10\n",
      "96/96 [==============================] - 113s 1s/step - loss: 1.5118 - accuracy: 0.9250 - val_loss: 10.7511 - val_accuracy: 0.7595\n",
      "Epoch 9/10\n",
      "96/96 [==============================] - 115s 1s/step - loss: 1.2811 - accuracy: 0.9315 - val_loss: 10.9381 - val_accuracy: 0.7742\n",
      "Epoch 10/10\n",
      "96/96 [==============================] - 114s 1s/step - loss: 0.7595 - accuracy: 0.9482 - val_loss: 10.4535 - val_accuracy: 0.8035\n"
     ]
    }
   ],
   "source": [
    "scores = {}\n",
    "\n",
    "for lr in [0.0001, 0.001, 0.01, 0.1]:\n",
    "    print(lr)\n",
    "    model = make_model(lr)\n",
    "    history = model.fit(train_ds, epochs=10, validation_data=val_ds)\n",
    "    scores[lr] = history.history"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "f201c388",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for lr, hist in scores.items():\n",
    "    plt.plot(hist['accuracy'], label=\"lr\")\n",
    "    plt.xticks(np.arange(10))\n",
    "plt.legend();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "067d200b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for lr, hist in scores.items():\n",
    "    plt.plot(hist['val_accuracy'], label=\"lr\")\n",
    "    plt.xticks(np.arange(10))\n",
    "plt.legend();"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3a722a86",
   "metadata": {},
   "source": [
    "* Choose the ```learning_rate=0.001```"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1555045b",
   "metadata": {},
   "source": [
    "## Checkpointing\n",
    "\n",
    "* Maybe the last results are not he best ones. Too have the best results we save models during training, when results are better than the previous model\n",
    "* After each epoch ```Calllbacks``` can be invoked"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "0391f426",
   "metadata": {},
   "outputs": [],
   "source": [
    "checkpoint = keras.callbacks.ModelCheckpoint('xception_v1_{epoch:02d}_{val_accuracy:.3f}.h5',\n",
    "                               save_best_only=True,\n",
    "                               monitor='val_accuracy',\n",
    "                               mode='max')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "9a6be034",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "96/96 [==============================] - 132s 1s/step - loss: 1.4994 - accuracy: 0.4937 - val_loss: 0.7027 - val_accuracy: 0.7683\n",
      "Epoch 2/10\n",
      "96/96 [==============================] - 131s 1s/step - loss: 0.6552 - accuracy: 0.7815 - val_loss: 0.6149 - val_accuracy: 0.8094\n",
      "Epoch 3/10\n",
      "96/96 [==============================] - 132s 1s/step - loss: 0.4939 - accuracy: 0.8409 - val_loss: 0.5807 - val_accuracy: 0.7977\n",
      "Epoch 4/10\n",
      "96/96 [==============================] - 131s 1s/step - loss: 0.4225 - accuracy: 0.8661 - val_loss: 0.5623 - val_accuracy: 0.7918\n",
      "Epoch 5/10\n",
      "96/96 [==============================] - 134s 1s/step - loss: 0.3721 - accuracy: 0.8992 - val_loss: 0.5416 - val_accuracy: 0.8123\n",
      "Epoch 6/10\n",
      "96/96 [==============================] - 155s 2s/step - loss: 0.3120 - accuracy: 0.9062 - val_loss: 0.5359 - val_accuracy: 0.8065\n",
      "Epoch 7/10\n",
      "96/96 [==============================] - 151s 2s/step - loss: 0.2969 - accuracy: 0.9153 - val_loss: 0.5312 - val_accuracy: 0.8240\n",
      "Epoch 8/10\n",
      "96/96 [==============================] - 144s 2s/step - loss: 0.2599 - accuracy: 0.9337 - val_loss: 0.5234 - val_accuracy: 0.8152\n",
      "Epoch 9/10\n",
      "96/96 [==============================] - 143s 1s/step - loss: 0.2356 - accuracy: 0.9427 - val_loss: 0.5252 - val_accuracy: 0.8270\n",
      "Epoch 10/10\n",
      "96/96 [==============================] - 142s 1s/step - loss: 0.2168 - accuracy: 0.9439 - val_loss: 0.5312 - val_accuracy: 0.8182\n"
     ]
    }
   ],
   "source": [
    "learning_rate = 0.001\n",
    "\n",
    "model = make_model(learning_rate)\n",
    "history = model.fit(train_ds, \n",
    "                    epochs=10, \n",
    "                    validation_data=val_ds,\n",
    "                    callbacks=[checkpoint])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b8b182ea",
   "metadata": {},
   "source": [
    "## Adding more Layers\n",
    "\n",
    "* Add another dense layer, this needs an activation function. Here we use ```ReLU``` activation\n",
    "* To see how much the GPU is used, use ```nvidia-smi``` in the terminal\n",
    "    * e.g. ```watch nvidia-smi``` during training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "1ac09b23",
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_model(learning_rate=.1, size_inner=100):\n",
    "    base_model = Xception(weights='imagenet', \n",
    "                      include_top=False, \n",
    "                      input_shape=(150, 150, 3))\n",
    "\n",
    "    base_model.trainable = False\n",
    "    \n",
    "    ##############################\n",
    "    \n",
    "    inputs = keras.Input(shape=(150, 150, 3))\n",
    "    base = base_model(inputs, training=False)\n",
    "    vectors = keras.layers.GlobalAveragePooling2D()(base)\n",
    "    inner = keras.layers.Dense(size_inner, activation='relu')(vectors)\n",
    "    outputs = keras.layers.Dense(10)(inner)\n",
    "    model = keras.Model(inputs, outputs)\n",
    "\n",
    "    ##############################\n",
    "    \n",
    "    optimizer = keras.optimizers.Adam(learning_rate=learning_rate)\n",
    "    loss = keras.losses.CategoricalCrossentropy(from_logits=True)\n",
    "    model.compile(optimizer=optimizer, loss=loss, metrics=['accuracy'])\n",
    "    \n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "175d05ff",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10\n",
      "Epoch 1/10\n",
      "96/96 [==============================] - 127s 1s/step - loss: 1.7759 - accuracy: 0.4114 - val_loss: 1.1747 - val_accuracy: 0.6686\n",
      "Epoch 2/10\n",
      "96/96 [==============================] - 125s 1s/step - loss: 1.0111 - accuracy: 0.6946 - val_loss: 0.8919 - val_accuracy: 0.7361\n",
      "Epoch 3/10\n",
      "96/96 [==============================] - 132s 1s/step - loss: 0.7266 - accuracy: 0.7677 - val_loss: 0.7750 - val_accuracy: 0.7801\n",
      "Epoch 4/10\n",
      "96/96 [==============================] - 131s 1s/step - loss: 0.6432 - accuracy: 0.7912 - val_loss: 0.7512 - val_accuracy: 0.7654\n",
      "Epoch 5/10\n",
      "96/96 [==============================] - 132s 1s/step - loss: 0.5288 - accuracy: 0.8272 - val_loss: 0.6926 - val_accuracy: 0.7977\n",
      "Epoch 6/10\n",
      "96/96 [==============================] - 131s 1s/step - loss: 0.4632 - accuracy: 0.8631 - val_loss: 0.6664 - val_accuracy: 0.7859\n",
      "Epoch 7/10\n",
      "96/96 [==============================] - 130s 1s/step - loss: 0.4324 - accuracy: 0.8651 - val_loss: 0.6388 - val_accuracy: 0.7918\n",
      "Epoch 8/10\n",
      "96/96 [==============================] - 131s 1s/step - loss: 0.3801 - accuracy: 0.8858 - val_loss: 0.6139 - val_accuracy: 0.8035\n",
      "Epoch 9/10\n",
      "96/96 [==============================] - 135s 1s/step - loss: 0.3305 - accuracy: 0.9073 - val_loss: 0.6031 - val_accuracy: 0.8240\n",
      "Epoch 10/10\n",
      "96/96 [==============================] - 132s 1s/step - loss: 0.2884 - accuracy: 0.9276 - val_loss: 0.6160 - val_accuracy: 0.8123\n",
      "100\n",
      "Epoch 1/10\n",
      "96/96 [==============================] - 134s 1s/step - loss: 1.2741 - accuracy: 0.5690 - val_loss: 0.6428 - val_accuracy: 0.7859\n",
      "Epoch 2/10\n",
      "96/96 [==============================] - 131s 1s/step - loss: 0.4966 - accuracy: 0.8283 - val_loss: 0.5906 - val_accuracy: 0.7977\n",
      "Epoch 3/10\n",
      "96/96 [==============================] - 102s 1s/step - loss: 0.3432 - accuracy: 0.8992 - val_loss: 0.6400 - val_accuracy: 0.7830\n",
      "Epoch 4/10\n",
      "96/96 [==============================] - 108s 1s/step - loss: 0.2364 - accuracy: 0.9358 - val_loss: 0.5644 - val_accuracy: 0.8035\n",
      "Epoch 5/10\n",
      "96/96 [==============================] - 137s 1s/step - loss: 0.1481 - accuracy: 0.9704 - val_loss: 0.5534 - val_accuracy: 0.8211\n",
      "Epoch 6/10\n",
      "96/96 [==============================] - 152s 2s/step - loss: 0.1224 - accuracy: 0.9711 - val_loss: 0.5435 - val_accuracy: 0.8182\n",
      "Epoch 7/10\n",
      "96/96 [==============================] - 152s 2s/step - loss: 0.0866 - accuracy: 0.9848 - val_loss: 0.6314 - val_accuracy: 0.8270\n",
      "Epoch 8/10\n",
      "96/96 [==============================] - 134s 1s/step - loss: 0.0573 - accuracy: 0.9943 - val_loss: 0.5867 - val_accuracy: 0.8182\n",
      "Epoch 9/10\n",
      "96/96 [==============================] - 130s 1s/step - loss: 0.0426 - accuracy: 0.9978 - val_loss: 0.6195 - val_accuracy: 0.8328\n",
      "Epoch 10/10\n",
      "96/96 [==============================] - 132s 1s/step - loss: 0.0313 - accuracy: 0.9990 - val_loss: 0.6119 - val_accuracy: 0.8211\n",
      "1000\n",
      "Epoch 1/10\n",
      "96/96 [==============================] - 156s 2s/step - loss: 1.2974 - accuracy: 0.5698 - val_loss: 0.6160 - val_accuracy: 0.7801\n",
      "Epoch 2/10\n",
      "96/96 [==============================] - 136s 1s/step - loss: 0.4057 - accuracy: 0.8652 - val_loss: 0.6368 - val_accuracy: 0.7918\n",
      "Epoch 3/10\n",
      "96/96 [==============================] - 133s 1s/step - loss: 0.2420 - accuracy: 0.9257 - val_loss: 0.5500 - val_accuracy: 0.8211\n",
      "Epoch 4/10\n",
      "96/96 [==============================] - 131s 1s/step - loss: 0.1498 - accuracy: 0.9635 - val_loss: 0.5428 - val_accuracy: 0.8328\n",
      "Epoch 5/10\n",
      "96/96 [==============================] - 131s 1s/step - loss: 0.0824 - accuracy: 0.9829 - val_loss: 0.6324 - val_accuracy: 0.8182\n",
      "Epoch 6/10\n",
      "96/96 [==============================] - 136s 1s/step - loss: 0.0367 - accuracy: 0.9975 - val_loss: 0.6851 - val_accuracy: 0.8123\n",
      "Epoch 7/10\n",
      "96/96 [==============================] - 119s 1s/step - loss: 0.0281 - accuracy: 0.9963 - val_loss: 0.6290 - val_accuracy: 0.8094\n",
      "Epoch 8/10\n",
      "96/96 [==============================] - 104s 1s/step - loss: 0.0132 - accuracy: 0.9994 - val_loss: 0.6531 - val_accuracy: 0.8152\n",
      "Epoch 9/10\n",
      "96/96 [==============================] - 109s 1s/step - loss: 0.0089 - accuracy: 0.9996 - val_loss: 0.6774 - val_accuracy: 0.8211\n",
      "Epoch 10/10\n",
      "96/96 [==============================] - 114s 1s/step - loss: 0.0121 - accuracy: 0.9981 - val_loss: 0.6985 - val_accuracy: 0.8211\n"
     ]
    }
   ],
   "source": [
    "learning_rate = 0.001\n",
    "scores = {}\n",
    "\n",
    "for size in [10, 100, 1000]:\n",
    "    print(size)\n",
    "    model = make_model(learning_rate=learning_rate, size_inner=size)\n",
    "    history = model.fit(train_ds, epochs=10, validation_data=val_ds)\n",
    "    scores[size] = history.history"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "921d7a1b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for size, hist in scores.items():\n",
    "    plt.plot(hist['accuracy'], label=\"size\")\n",
    "    plt.xticks(np.arange(10))\n",
    "plt.legend();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "45b7b0f4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for size, hist in scores.items():\n",
    "    plt.plot(hist['val_accuracy'], label=\"size\")\n",
    "    plt.xticks(np.arange(10))\n",
    "plt.legend();"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e7992b73",
   "metadata": {},
   "source": [
    "## Regularization and Dropout\n",
    "* Regularization by freezing a part of the network\n",
    "* Adding Dropout to our model\n",
    "* Experimenting with different values\n",
    "\n",
    "![regularization](Screenshot_5.png)\n",
    "![regularization](Screenshot_6.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "4fb14ff0",
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_model(learning_rate=.1, size_inner=100, droprate=0.5):\n",
    "    base_model = Xception(weights='imagenet', \n",
    "                      include_top=False, \n",
    "                      input_shape=(150, 150, 3))\n",
    "\n",
    "    base_model.trainable = False\n",
    "    \n",
    "    ##############################\n",
    "    \n",
    "    inputs = keras.Input(shape=(150, 150, 3))\n",
    "    base = base_model(inputs, training=False)\n",
    "    vectors = keras.layers.GlobalAveragePooling2D()(base)\n",
    "    inner = keras.layers.Dense(size_inner, activation='relu')(vectors)\n",
    "    drop = keras.layers.Dropout(droprate)(inner)\n",
    "    outputs = keras.layers.Dense(10)(drop)\n",
    "    model = keras.Model(inputs, outputs)\n",
    "\n",
    "    ##############################\n",
    "    \n",
    "    optimizer = keras.optimizers.Adam(learning_rate=learning_rate)\n",
    "    loss = keras.losses.CategoricalCrossentropy(from_logits=True)\n",
    "    model.compile(optimizer=optimizer, loss=loss, metrics=['accuracy'])\n",
    "    \n",
    "    return model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f2209a4d",
   "metadata": {},
   "source": [
    "![sketch model](Screenshot_7.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "e1569157",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0\n",
      "Epoch 1/30\n",
      "96/96 [==============================] - 108s 1s/step - loss: 1.2499 - accuracy: 0.5813 - val_loss: 0.6282 - val_accuracy: 0.7859\n",
      "Epoch 2/30\n",
      "96/96 [==============================] - 107s 1s/step - loss: 0.5024 - accuracy: 0.8375 - val_loss: 0.5531 - val_accuracy: 0.8065\n",
      "Epoch 3/30\n",
      "96/96 [==============================] - 107s 1s/step - loss: 0.3602 - accuracy: 0.8820 - val_loss: 0.5879 - val_accuracy: 0.8006\n",
      "Epoch 4/30\n",
      "96/96 [==============================] - 100s 1s/step - loss: 0.2516 - accuracy: 0.9263 - val_loss: 0.5473 - val_accuracy: 0.8211\n",
      "Epoch 5/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.1592 - accuracy: 0.9579 - val_loss: 0.5380 - val_accuracy: 0.8328\n",
      "Epoch 6/30\n",
      "96/96 [==============================] - 97s 1s/step - loss: 0.1261 - accuracy: 0.9710 - val_loss: 0.5605 - val_accuracy: 0.8182\n",
      "Epoch 7/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.0867 - accuracy: 0.9822 - val_loss: 0.5499 - val_accuracy: 0.8211\n",
      "Epoch 8/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.0572 - accuracy: 0.9914 - val_loss: 0.5705 - val_accuracy: 0.8123\n",
      "Epoch 9/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.0429 - accuracy: 0.9971 - val_loss: 0.5962 - val_accuracy: 0.8240\n",
      "Epoch 10/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.0261 - accuracy: 0.9994 - val_loss: 0.6016 - val_accuracy: 0.8240\n",
      "Epoch 11/30\n",
      "96/96 [==============================] - 97s 1s/step - loss: 0.0204 - accuracy: 0.9999 - val_loss: 0.5978 - val_accuracy: 0.8270\n",
      "Epoch 12/30\n",
      "96/96 [==============================] - 97s 1s/step - loss: 0.0168 - accuracy: 0.9997 - val_loss: 0.6104 - val_accuracy: 0.8299\n",
      "Epoch 13/30\n",
      "96/96 [==============================] - 97s 1s/step - loss: 0.0144 - accuracy: 0.9999 - val_loss: 0.6307 - val_accuracy: 0.8240\n",
      "Epoch 14/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.0128 - accuracy: 0.9994 - val_loss: 0.6443 - val_accuracy: 0.8358\n",
      "Epoch 15/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.0105 - accuracy: 0.9995 - val_loss: 0.6544 - val_accuracy: 0.8328\n",
      "Epoch 16/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.0082 - accuracy: 0.9997 - val_loss: 0.6448 - val_accuracy: 0.8270\n",
      "Epoch 17/30\n",
      "96/96 [==============================] - 97s 1s/step - loss: 0.0089 - accuracy: 0.9996 - val_loss: 0.6604 - val_accuracy: 0.8240\n",
      "Epoch 18/30\n",
      "96/96 [==============================] - 97s 1s/step - loss: 0.0055 - accuracy: 1.0000 - val_loss: 0.6711 - val_accuracy: 0.8299\n",
      "Epoch 19/30\n",
      "96/96 [==============================] - 97s 1s/step - loss: 0.0080 - accuracy: 0.9996 - val_loss: 0.6901 - val_accuracy: 0.8211\n",
      "Epoch 20/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.0061 - accuracy: 0.9997 - val_loss: 0.6796 - val_accuracy: 0.8211\n",
      "Epoch 21/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.0047 - accuracy: 1.0000 - val_loss: 0.7780 - val_accuracy: 0.8152\n",
      "Epoch 22/30\n",
      "96/96 [==============================] - 97s 1s/step - loss: 0.0141 - accuracy: 0.9986 - val_loss: 0.6843 - val_accuracy: 0.8182\n",
      "Epoch 23/30\n",
      "96/96 [==============================] - 97s 1s/step - loss: 0.0038 - accuracy: 0.9999 - val_loss: 0.7150 - val_accuracy: 0.8270\n",
      "Epoch 24/30\n",
      "96/96 [==============================] - 97s 1s/step - loss: 0.0085 - accuracy: 0.9992 - val_loss: 0.6907 - val_accuracy: 0.8299\n",
      "Epoch 25/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.0107 - accuracy: 0.9982 - val_loss: 0.7421 - val_accuracy: 0.8240\n",
      "Epoch 26/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.0065 - accuracy: 0.9995 - val_loss: 0.7254 - val_accuracy: 0.8328\n",
      "Epoch 27/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.0058 - accuracy: 0.9991 - val_loss: 0.7426 - val_accuracy: 0.8299\n",
      "Epoch 28/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.0054 - accuracy: 0.9989 - val_loss: 0.7180 - val_accuracy: 0.8328\n",
      "Epoch 29/30\n",
      "96/96 [==============================] - 97s 1s/step - loss: 0.0042 - accuracy: 0.9989 - val_loss: 0.7401 - val_accuracy: 0.8358\n",
      "Epoch 30/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.0058 - accuracy: 0.9988 - val_loss: 0.7647 - val_accuracy: 0.8211\n",
      "0.2\n",
      "Epoch 1/30\n",
      "96/96 [==============================] - 101s 1s/step - loss: 1.3595 - accuracy: 0.5365 - val_loss: 0.6672 - val_accuracy: 0.7977\n",
      "Epoch 2/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.5983 - accuracy: 0.7986 - val_loss: 0.5599 - val_accuracy: 0.8211\n",
      "Epoch 3/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.4577 - accuracy: 0.8335 - val_loss: 0.5866 - val_accuracy: 0.8152\n",
      "Epoch 4/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.3610 - accuracy: 0.8834 - val_loss: 0.5269 - val_accuracy: 0.8240\n",
      "Epoch 5/30\n",
      "96/96 [==============================] - 109s 1s/step - loss: 0.2610 - accuracy: 0.9202 - val_loss: 0.5482 - val_accuracy: 0.8065\n",
      "Epoch 6/30\n",
      "96/96 [==============================] - 106s 1s/step - loss: 0.2487 - accuracy: 0.9241 - val_loss: 0.5991 - val_accuracy: 0.8270\n",
      "Epoch 7/30\n",
      "96/96 [==============================] - 108s 1s/step - loss: 0.1593 - accuracy: 0.9556 - val_loss: 0.5103 - val_accuracy: 0.8358\n",
      "Epoch 8/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.1439 - accuracy: 0.9618 - val_loss: 0.5601 - val_accuracy: 0.8299\n",
      "Epoch 9/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.1093 - accuracy: 0.9720 - val_loss: 0.5576 - val_accuracy: 0.8152\n",
      "Epoch 10/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.0945 - accuracy: 0.9791 - val_loss: 0.5757 - val_accuracy: 0.8211\n",
      "Epoch 11/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.0881 - accuracy: 0.9765 - val_loss: 0.6186 - val_accuracy: 0.8123\n",
      "Epoch 12/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.0737 - accuracy: 0.9819 - val_loss: 0.6095 - val_accuracy: 0.8299\n",
      "Epoch 13/30\n",
      "96/96 [==============================] - 100s 1s/step - loss: 0.0581 - accuracy: 0.9918 - val_loss: 0.5796 - val_accuracy: 0.8270\n",
      "Epoch 14/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.0460 - accuracy: 0.9921 - val_loss: 0.6525 - val_accuracy: 0.8240\n",
      "Epoch 15/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.0408 - accuracy: 0.9914 - val_loss: 0.6447 - val_accuracy: 0.8270\n",
      "Epoch 16/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.0371 - accuracy: 0.9923 - val_loss: 0.6560 - val_accuracy: 0.8211\n",
      "Epoch 17/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.0308 - accuracy: 0.9937 - val_loss: 0.6443 - val_accuracy: 0.8035\n",
      "Epoch 18/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.0313 - accuracy: 0.9944 - val_loss: 0.6445 - val_accuracy: 0.8211\n",
      "Epoch 19/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.0272 - accuracy: 0.9929 - val_loss: 0.6510 - val_accuracy: 0.8152\n",
      "Epoch 20/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.0263 - accuracy: 0.9952 - val_loss: 0.6715 - val_accuracy: 0.8152\n",
      "Epoch 21/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.0239 - accuracy: 0.9953 - val_loss: 0.6959 - val_accuracy: 0.8182\n",
      "Epoch 22/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.0169 - accuracy: 0.9983 - val_loss: 0.6920 - val_accuracy: 0.8152\n",
      "Epoch 23/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.0246 - accuracy: 0.9968 - val_loss: 0.7011 - val_accuracy: 0.8123\n",
      "Epoch 24/30\n",
      "96/96 [==============================] - 100s 1s/step - loss: 0.0149 - accuracy: 0.9977 - val_loss: 0.7004 - val_accuracy: 0.8299\n",
      "Epoch 25/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.0205 - accuracy: 0.9954 - val_loss: 0.7464 - val_accuracy: 0.8270\n",
      "Epoch 26/30\n",
      "96/96 [==============================] - 108s 1s/step - loss: 0.0156 - accuracy: 0.9975 - val_loss: 0.6794 - val_accuracy: 0.8152\n",
      "Epoch 27/30\n",
      "96/96 [==============================] - 107s 1s/step - loss: 0.0151 - accuracy: 0.9992 - val_loss: 0.7787 - val_accuracy: 0.8299\n",
      "Epoch 28/30\n",
      "96/96 [==============================] - 106s 1s/step - loss: 0.0320 - accuracy: 0.9904 - val_loss: 0.7263 - val_accuracy: 0.8240\n",
      "Epoch 29/30\n",
      "96/96 [==============================] - 101s 1s/step - loss: 0.0253 - accuracy: 0.9941 - val_loss: 0.8182 - val_accuracy: 0.8240\n",
      "Epoch 30/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.0457 - accuracy: 0.9857 - val_loss: 0.8459 - val_accuracy: 0.8094\n",
      "0.5\n",
      "Epoch 1/30\n",
      "96/96 [==============================] - 102s 1s/step - loss: 1.5829 - accuracy: 0.4879 - val_loss: 0.7706 - val_accuracy: 0.7449\n",
      "Epoch 2/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.8141 - accuracy: 0.7344 - val_loss: 0.6446 - val_accuracy: 0.7830\n",
      "Epoch 3/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.6748 - accuracy: 0.7717 - val_loss: 0.6278 - val_accuracy: 0.7889\n",
      "Epoch 4/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.5946 - accuracy: 0.7984 - val_loss: 0.5842 - val_accuracy: 0.7830\n",
      "Epoch 5/30\n",
      "96/96 [==============================] - 97s 1s/step - loss: 0.5170 - accuracy: 0.8214 - val_loss: 0.5529 - val_accuracy: 0.8035\n",
      "Epoch 6/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.4188 - accuracy: 0.8647 - val_loss: 0.5228 - val_accuracy: 0.8152\n",
      "Epoch 7/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.3861 - accuracy: 0.8704 - val_loss: 0.5413 - val_accuracy: 0.8065\n",
      "Epoch 8/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.3847 - accuracy: 0.8653 - val_loss: 0.5089 - val_accuracy: 0.8211\n",
      "Epoch 9/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.3175 - accuracy: 0.8904 - val_loss: 0.5201 - val_accuracy: 0.8182\n",
      "Epoch 10/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.2755 - accuracy: 0.9011 - val_loss: 0.5090 - val_accuracy: 0.8299\n",
      "Epoch 11/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.2582 - accuracy: 0.9178 - val_loss: 0.5138 - val_accuracy: 0.8358\n",
      "Epoch 12/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.2399 - accuracy: 0.9194 - val_loss: 0.5195 - val_accuracy: 0.8387\n",
      "Epoch 13/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.2279 - accuracy: 0.9166 - val_loss: 0.5430 - val_accuracy: 0.8240\n",
      "Epoch 14/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.1957 - accuracy: 0.9325 - val_loss: 0.5276 - val_accuracy: 0.8328\n",
      "Epoch 15/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.1831 - accuracy: 0.9404 - val_loss: 0.5636 - val_accuracy: 0.8240\n",
      "Epoch 16/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.1878 - accuracy: 0.9355 - val_loss: 0.5834 - val_accuracy: 0.8358\n",
      "Epoch 17/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.1359 - accuracy: 0.9594 - val_loss: 0.5741 - val_accuracy: 0.8152\n",
      "Epoch 18/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.1261 - accuracy: 0.9693 - val_loss: 0.5703 - val_accuracy: 0.8328\n",
      "Epoch 19/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.1273 - accuracy: 0.9580 - val_loss: 0.5203 - val_accuracy: 0.8416\n",
      "Epoch 20/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.1132 - accuracy: 0.9657 - val_loss: 0.5209 - val_accuracy: 0.8416\n",
      "Epoch 21/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.1061 - accuracy: 0.9674 - val_loss: 0.5589 - val_accuracy: 0.8211\n",
      "Epoch 22/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.1226 - accuracy: 0.9601 - val_loss: 0.5889 - val_accuracy: 0.8416\n",
      "Epoch 23/30\n",
      "96/96 [==============================] - 97s 1s/step - loss: 0.1094 - accuracy: 0.9661 - val_loss: 0.5981 - val_accuracy: 0.8387\n",
      "Epoch 24/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.1047 - accuracy: 0.9652 - val_loss: 0.5917 - val_accuracy: 0.8270\n",
      "Epoch 25/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.1126 - accuracy: 0.9635 - val_loss: 0.6312 - val_accuracy: 0.8299\n",
      "Epoch 26/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.1035 - accuracy: 0.9633 - val_loss: 0.5947 - val_accuracy: 0.8152\n",
      "Epoch 27/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.0917 - accuracy: 0.9691 - val_loss: 0.6568 - val_accuracy: 0.8211\n",
      "Epoch 28/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.0901 - accuracy: 0.9747 - val_loss: 0.6542 - val_accuracy: 0.8416\n",
      "Epoch 29/30\n",
      "96/96 [==============================] - 100s 1s/step - loss: 0.0779 - accuracy: 0.9735 - val_loss: 0.6532 - val_accuracy: 0.8211\n",
      "Epoch 30/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.0927 - accuracy: 0.9692 - val_loss: 0.6773 - val_accuracy: 0.8387\n",
      "0.8\n",
      "Epoch 1/30\n",
      "96/96 [==============================] - 101s 1s/step - loss: 2.0724 - accuracy: 0.3138 - val_loss: 1.1485 - val_accuracy: 0.6393\n",
      "Epoch 2/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 1.4339 - accuracy: 0.5076 - val_loss: 0.9459 - val_accuracy: 0.7009\n",
      "Epoch 3/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 1.2817 - accuracy: 0.5386 - val_loss: 0.8088 - val_accuracy: 0.7478\n",
      "Epoch 4/30\n",
      "96/96 [==============================] - 100s 1s/step - loss: 1.2049 - accuracy: 0.5536 - val_loss: 0.7426 - val_accuracy: 0.7537\n",
      "Epoch 5/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 1.1232 - accuracy: 0.5758 - val_loss: 0.7646 - val_accuracy: 0.7478\n",
      "Epoch 6/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 1.0436 - accuracy: 0.5903 - val_loss: 0.6942 - val_accuracy: 0.7889\n",
      "Epoch 7/30\n",
      "96/96 [==============================] - 97s 1s/step - loss: 1.0512 - accuracy: 0.6160 - val_loss: 0.6881 - val_accuracy: 0.7713\n",
      "Epoch 8/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.9919 - accuracy: 0.6241 - val_loss: 0.6691 - val_accuracy: 0.7771\n",
      "Epoch 9/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.9503 - accuracy: 0.6555 - val_loss: 0.6470 - val_accuracy: 0.7654\n",
      "Epoch 10/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.9158 - accuracy: 0.6582 - val_loss: 0.6241 - val_accuracy: 0.7859\n",
      "Epoch 11/30\n",
      "96/96 [==============================] - 97s 1s/step - loss: 0.8883 - accuracy: 0.6661 - val_loss: 0.6118 - val_accuracy: 0.8094\n",
      "Epoch 12/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.8541 - accuracy: 0.6700 - val_loss: 0.5982 - val_accuracy: 0.7889\n",
      "Epoch 13/30\n",
      "96/96 [==============================] - 97s 1s/step - loss: 0.8851 - accuracy: 0.6522 - val_loss: 0.5951 - val_accuracy: 0.7977\n",
      "Epoch 14/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.8190 - accuracy: 0.6967 - val_loss: 0.5900 - val_accuracy: 0.8065\n",
      "Epoch 15/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.8020 - accuracy: 0.6968 - val_loss: 0.5668 - val_accuracy: 0.8123\n",
      "Epoch 16/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.7859 - accuracy: 0.7093 - val_loss: 0.5718 - val_accuracy: 0.8065\n",
      "Epoch 17/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.7881 - accuracy: 0.6959 - val_loss: 0.5658 - val_accuracy: 0.8094\n",
      "Epoch 18/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.7607 - accuracy: 0.7062 - val_loss: 0.5652 - val_accuracy: 0.8006\n",
      "Epoch 19/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.7480 - accuracy: 0.7020 - val_loss: 0.6005 - val_accuracy: 0.7889\n",
      "Epoch 20/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.7148 - accuracy: 0.7143 - val_loss: 0.5872 - val_accuracy: 0.8006\n",
      "Epoch 21/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.7057 - accuracy: 0.7400 - val_loss: 0.5572 - val_accuracy: 0.8328\n",
      "Epoch 22/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.7166 - accuracy: 0.7190 - val_loss: 0.5482 - val_accuracy: 0.8240\n",
      "Epoch 23/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.6939 - accuracy: 0.7428 - val_loss: 0.5919 - val_accuracy: 0.8182\n",
      "Epoch 24/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.6510 - accuracy: 0.7423 - val_loss: 0.5532 - val_accuracy: 0.8094\n",
      "Epoch 25/30\n",
      "96/96 [==============================] - 97s 1s/step - loss: 0.6734 - accuracy: 0.7267 - val_loss: 0.5861 - val_accuracy: 0.8006\n",
      "Epoch 26/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.6825 - accuracy: 0.7255 - val_loss: 0.5916 - val_accuracy: 0.8035\n",
      "Epoch 27/30\n",
      "96/96 [==============================] - 99s 1s/step - loss: 0.6379 - accuracy: 0.7523 - val_loss: 0.5569 - val_accuracy: 0.8006\n",
      "Epoch 28/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.6198 - accuracy: 0.7558 - val_loss: 0.5562 - val_accuracy: 0.7889\n",
      "Epoch 29/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.6478 - accuracy: 0.7401 - val_loss: 0.5585 - val_accuracy: 0.7977\n",
      "Epoch 30/30\n",
      "96/96 [==============================] - 98s 1s/step - loss: 0.6250 - accuracy: 0.7541 - val_loss: 0.5640 - val_accuracy: 0.8065\n"
     ]
    }
   ],
   "source": [
    "learning_rate = 0.001\n",
    "size = 100\n",
    "scores = {}\n",
    "\n",
    "for droprate in [0.0, 0.2, 0.5, 0.8]:\n",
    "    print(droprate)\n",
    "    model = make_model(learning_rate=learning_rate, size_inner=size, droprate=droprate)\n",
    "    history = model.fit(train_ds, epochs=30, validation_data=val_ds)\n",
    "    scores[lr] = history.history"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "e2fb3c0e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for droprate, hist in scores.items():\n",
    "    plt.plot(hist['accuracy'], label=\"droprate\")\n",
    "    plt.xticks(np.arange(30))\n",
    "plt.legend();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "ad942674",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for droprate, hist in scores.items():\n",
    "    plt.plot(hist['val_accuracy'], label=\"droprate\")\n",
    "    plt.xticks(np.arange(30))\n",
    "plt.legend();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "e5b4dd1b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# look at specific droprates for training and validation\n",
    "hist = scores[0.1]\n",
    "plt.plot(hist['accuracy'], label=\"droprate\")\n",
    "plt.plot(hist['val_accuracy'], label=\"droprate\");"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ce7d998f-9f93-4953-b29a-eb0eb0918ec3",
   "metadata": {},
   "source": [
    "## Data Augmentation\n",
    "\n",
    "* Create more data from the existing data\n",
    "* Possible image transformations\n",
    "    * flip (horiontally, vertically)\n",
    "    * rotation\n",
    "    * shift, width-shift\n",
    "    * shear \n",
    "    * zoom\n",
    "    * brightness/ contrast\n",
    "    * black patch\n",
    "* All these transformations can be combined\n",
    "* In keras this can be done in the ```ImageDataGenerator```\n",
    "* Augmentation is only applied to the training dataset, not to the validation set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "272b6d50-8bf8-45bc-ad4a-aec4b3084e6a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 3068 images belonging to 10 classes.\n",
      "Found 341 images belonging to 10 classes.\n"
     ]
    }
   ],
   "source": [
    "train_gen = ImageDataGenerator(\n",
    "                preprocessing_function=preprocess_input,\n",
    "                shear_range=10.0,\n",
    "                zoom_range=0.1,\n",
    "                vertical_flip=True\n",
    "                )\n",
    "\n",
    "# read images from directory\n",
    "train_ds = train_gen.flow_from_directory('../data/clothing-dataset-small/train', \n",
    "                                         target_size=(150, 150), \n",
    "                                         batch_size=32)\n",
    "\n",
    "val_gen = ImageDataGenerator(preprocessing_function=preprocess_input)\n",
    "\n",
    "# read images from directory\n",
    "val_ds = val_gen.flow_from_directory('../data/clothing-dataset-small/validation', \n",
    "                                         target_size=(150, 150), \n",
    "                                         batch_size=32)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c3090756-4c03-4e1b-9e19-298c61908539",
   "metadata": {},
   "source": [
    "### How to choose augmentation:\n",
    "\n",
    "* Use your own judgement (Does it make sense?)\n",
    "* Look at the images, what kind of variations are there?\n",
    "    * E.g. are objects always centeres?\n",
    "* Tune as a hyperparameter\n",
    "    * Train for a certain amount of epochs -> Is it better?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "220b450e-7df5-4f6e-9d13-5a20b759ff9d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "96/96 [==============================] - 112s 1s/step - loss: 1.6114 - accuracy: 0.4490 - val_loss: 0.8044 - val_accuracy: 0.7331\n",
      "Epoch 2/10\n",
      "96/96 [==============================] - 143s 1s/step - loss: 0.9001 - accuracy: 0.7079 - val_loss: 0.6676 - val_accuracy: 0.7947\n",
      "Epoch 3/10\n",
      "96/96 [==============================] - 123s 1s/step - loss: 0.8034 - accuracy: 0.7246 - val_loss: 0.6446 - val_accuracy: 0.7801\n",
      "Epoch 4/10\n",
      "96/96 [==============================] - 113s 1s/step - loss: 0.7502 - accuracy: 0.7343 - val_loss: 0.6332 - val_accuracy: 0.7947\n",
      "Epoch 5/10\n",
      "96/96 [==============================] - 112s 1s/step - loss: 0.6272 - accuracy: 0.7828 - val_loss: 0.5863 - val_accuracy: 0.8006\n",
      "Epoch 6/10\n",
      "96/96 [==============================] - 112s 1s/step - loss: 0.6340 - accuracy: 0.7740 - val_loss: 0.5930 - val_accuracy: 0.8211\n",
      "Epoch 7/10\n",
      "96/96 [==============================] - 110s 1s/step - loss: 0.5851 - accuracy: 0.7980 - val_loss: 0.5900 - val_accuracy: 0.7918\n",
      "Epoch 8/10\n",
      "96/96 [==============================] - 111s 1s/step - loss: 0.5063 - accuracy: 0.8225 - val_loss: 0.5789 - val_accuracy: 0.8152\n",
      "Epoch 9/10\n",
      "96/96 [==============================] - 109s 1s/step - loss: 0.4986 - accuracy: 0.8280 - val_loss: 0.6019 - val_accuracy: 0.7977\n",
      "Epoch 10/10\n",
      "96/96 [==============================] - 111s 1s/step - loss: 0.4958 - accuracy: 0.8328 - val_loss: 0.5895 - val_accuracy: 0.7889\n"
     ]
    }
   ],
   "source": [
    "learning_rate = 0.001\n",
    "size = 100\n",
    "droprate = 0.2\n",
    "\n",
    "model = make_model(\n",
    "        learning_rate=learning_rate,\n",
    "        size_inner=size,\n",
    "        droprate=droprate)\n",
    "\n",
    "history = model.fit(train_ds, epochs=10, validation_data=val_ds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "cc990bc0-c742-4467-b19f-3750f26bbb5e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "hist = history\n",
    "plt.plot(hist.history['accuracy'], label=\"train\")\n",
    "plt.plot(hist.history['val_accuracy'], label=\"valid\");"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3e7529e0-5908-481e-8864-037215b3c896",
   "metadata": {},
   "source": [
    "## Train a larger Model\n",
    "* train on images 299x299 "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "bde6000b-b99a-40d7-bf0c-6593b2e660e4",
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_model(input_size=150, \n",
    "               learning_rate=.1, \n",
    "               size_inner=100, \n",
    "               droprate=0.5):\n",
    "    base_model = Xception(weights='imagenet', \n",
    "                      include_top=False, \n",
    "                      input_shape=(input_size, input_size, 3))\n",
    "\n",
    "    base_model.trainable = False\n",
    "    \n",
    "    ##############################\n",
    "    \n",
    "    inputs = keras.Input(shape=(input_size, input_size, 3))\n",
    "    base = base_model(inputs, training=False)\n",
    "    vectors = keras.layers.GlobalAveragePooling2D()(base)\n",
    "    inner = keras.layers.Dense(size_inner, activation='relu')(vectors)\n",
    "    drop = keras.layers.Dropout(droprate)(inner)\n",
    "    outputs = keras.layers.Dense(10)(drop)\n",
    "    model = keras.Model(inputs, outputs)\n",
    "\n",
    "    ##############################\n",
    "    \n",
    "    optimizer = keras.optimizers.Adam(learning_rate=learning_rate)\n",
    "    loss = keras.losses.CategoricalCrossentropy(from_logits=True)\n",
    "    model.compile(optimizer=optimizer, loss=loss, metrics=['accuracy'])\n",
    "    \n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "b9e242dd-817d-463d-a335-0d82ba62f74f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 3068 images belonging to 10 classes.\n",
      "Found 341 images belonging to 10 classes.\n"
     ]
    }
   ],
   "source": [
    "train_gen = ImageDataGenerator(\n",
    "                preprocessing_function=preprocess_input,\n",
    "                shear_range=10.0,\n",
    "                zoom_range=0.1,\n",
    "                vertical_flip=True\n",
    "                )\n",
    "\n",
    "# read images from directory\n",
    "train_ds = train_gen.flow_from_directory('../data/clothing-dataset-small/train', \n",
    "                                         target_size=(299, 299), \n",
    "                                         batch_size=32)\n",
    "\n",
    "val_gen = ImageDataGenerator(preprocessing_function=preprocess_input)\n",
    "\n",
    "# read images from directory\n",
    "val_ds = val_gen.flow_from_directory('../data/clothing-dataset-small/validation', \n",
    "                                         target_size=(299, 299), \n",
    "                                         batch_size=32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "5356742c-309d-4618-aa1c-07f07c5717f9",
   "metadata": {},
   "outputs": [],
   "source": [
    "checkpoint = keras.callbacks.ModelCheckpoint(\n",
    "                    'xception_v4_{epoch:02d}_{val_accuracy:.3f}.h5',\n",
    "                    save_best_only=True,\n",
    "                    monitor='val_accuracy',\n",
    "                    mode='max')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "ac9ddc58-61c1-4775-b543-8daa5cb33f63",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "96/96 [==============================] - 447s 5s/step - loss: 1.4414 - accuracy: 0.5220 - val_loss: 0.5670 - val_accuracy: 0.8152\n",
      "Epoch 2/5\n",
      "96/96 [==============================] - 423s 4s/step - loss: 0.7087 - accuracy: 0.7538 - val_loss: 0.4926 - val_accuracy: 0.8152\n",
      "Epoch 3/5\n",
      "96/96 [==============================] - 402s 4s/step - loss: 0.5693 - accuracy: 0.8063 - val_loss: 0.4612 - val_accuracy: 0.8211\n",
      "Epoch 4/5\n",
      "96/96 [==============================] - 420s 4s/step - loss: 0.4935 - accuracy: 0.8339 - val_loss: 0.4405 - val_accuracy: 0.8534\n",
      "Epoch 5/5\n",
      "96/96 [==============================] - 445s 5s/step - loss: 0.4677 - accuracy: 0.8431 - val_loss: 0.4391 - val_accuracy: 0.8475\n"
     ]
    }
   ],
   "source": [
    "learning_rate = 0.001\n",
    "size = 100\n",
    "droprate = 0.2\n",
    "input_size = 299\n",
    "\n",
    "model = make_model(\n",
    "        input_size=input_size,\n",
    "        learning_rate=learning_rate,\n",
    "        size_inner=size,\n",
    "        droprate=droprate)\n",
    "\n",
    "history = model.fit(train_ds, \n",
    "                    epochs=5, \n",
    "                    validation_data=val_ds,\n",
    "                    callbacks=[checkpoint])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8b3c50d9-9c2d-4ccf-9578-8411c4e6d540",
   "metadata": {},
   "source": [
    "## Using the Model\n",
    "\n",
    "* Loading the Model\n",
    "* Evaluating the Model\n",
    "* Getting Predictions\n",
    "\n",
    "**Pretend the model has been trained and we now start a new Notebook**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "372d54f0-2eb0-4bb9-8df5-ee7097d175a4",
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "from tensorflow.keras.preprocessing.image import load_img\n",
    "from tensorflow.keras.applications.xception import preprocess_input\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "90007640-8870-46fb-aefe-cd16aa2c24c6",
   "metadata": {},
   "outputs": [],
   "source": [
    "model = keras.models.load_model('xception_v4_05_0.850.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "862343a1-83af-4be4-a288-526b4e5951a1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 341 images belonging to 10 classes.\n"
     ]
    }
   ],
   "source": [
    "# read test dataset in the same way as validation dataset\n",
    "test_gen = ImageDataGenerator(preprocessing_function=preprocess_input)\n",
    "\n",
    "# read images from directory\n",
    "test_ds = test_gen.flow_from_directory('../data/clothing-dataset-small/validation', \n",
    "                                         target_size=(299, 299), \n",
    "                                         batch_size=32,\n",
    "                                         shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "cc479f65-383a-4541-a329-5b1ab9621a21",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "11/11 [==============================] - 35s 3s/step - loss: 0.4433 - accuracy: 0.8504\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.443279892206192, 0.8504399061203003]"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# gives loss and accuracy on test set\n",
    "model.evaluate(test_ds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "34072e44-5c03-41f9-9a13-170d6aa8c6b0",
   "metadata": {},
   "outputs": [],
   "source": [
    "# apply the model to an image\n",
    "path = '../data/clothing-dataset-small/test/pants/c8d21106-bbdb-4e8d-83e4-bf3d14e54c16.jpg'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "367070ec-789e-4204-a359-9a8add861617",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<PIL.Image.Image image mode=RGB size=299x299 at 0x7FA29691CF70>"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "img = load_img(path, target_size=(299,299))\n",
    "img"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "a84cc160-b10b-403e-9296-bd14dbe90616",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 299, 299, 3)"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# turn image into a numpy array\n",
    "x = np.array(img)\n",
    "X = np.array([x])\n",
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "e1ed544d-331a-4288-9525-e4625aa2c828",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = preprocess_input(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "1cddad41-a6d4-42f6-9add-6557cca72edf",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:5 out of the last 5 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fa296925a60> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n"
     ]
    }
   ],
   "source": [
    "pred = model.predict(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "e4f0d98c-8725-4517-9ece-933005e5a065",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([-1.7340311 , -3.5461495 ,  0.06269813,  1.0456045 ,  8.627468  ,\n",
       "       -0.29695672, -4.8185697 ,  3.4118295 , -0.6146142 , -1.990872  ],\n",
       "      dtype=float32)"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "1523e0e4-bbea-4caa-af24-e8b7865512fb",
   "metadata": {},
   "outputs": [],
   "source": [
    "classes = ['dress',\n",
    " 'hat',\n",
    " 'longsleeve',\n",
    " 'outwear',\n",
    " 'pants',\n",
    " 'shirt',\n",
    " 'shoes',\n",
    " 'shorts',\n",
    " 'skirt',\n",
    " 't-shirt']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "884c9524-9371-40d1-9fa1-5a03867cac3b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'dress': -1.7340311,\n",
       " 'hat': -3.5461495,\n",
       " 'longsleeve': 0.062698126,\n",
       " 'outwear': 1.0456045,\n",
       " 'pants': 8.627468,\n",
       " 'shirt': -0.29695672,\n",
       " 'shoes': -4.8185697,\n",
       " 'shorts': 3.4118295,\n",
       " 'skirt': -0.6146142,\n",
       " 't-shirt': -1.990872}"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dict(zip(classes, pred[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "538b89b5-c23f-4d67-a8c5-d0c93d3dc2ea",
   "metadata": {},
   "source": [
    "**Note**\n",
    "\n",
    "The predictions here are not propabilities, but logits\n",
    "* could be transformed into propabilities (using softmax), but can also be treated as relative likelihoods. I.e. the largest number is the predicted class "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3660b3a8-c479-4691-8ddc-ac6b4d35e640",
   "metadata": {},
   "source": [
    "## Summary\n",
    "* We can use transfer-learning for general image classification\n",
    "* Convolutional layers let us turn images into a vector\n",
    "* Dense layers use the vector to make predictions\n",
    "* Instead of training a model from scratch, we can use transfer learning and re-use already trained convolutional layers\n",
    "* First, train a small model (150x150), before training a big one (299x299)\n",
    "* Learning rate - how fast the model trains. Fast learners aren't always the best ones\n",
    "* We can save the best model using callbacks and checkpointing\n",
    "* To avoid overfitting, use dropout and augmentation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6f219ce7-c3ac-4d5c-84fb-035ad76e90ee",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "ml-zoomcamp",
   "language": "python",
   "name": "ml-zoomcamp"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}