NN.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "from IPython.display import SVG\n",
    "\n",
    "import numpy as np\n",
    "\n",
    "import pandas as pd\n",
    "\n",
    "import keras\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Conv2D, Flatten, MaxPool2D, Dropout, BatchNormalization, Activation\n",
    "from keras.callbacks import ModelCheckpoint, EarlyStopping\n",
    "from keras.models import load_model\n",
    "from keras_tqdm import TQDMNotebookCallback\n",
    "from keras.utils import model_to_dot\n",
    "\n",
    "from sklearn import decomposition\n",
    "from sklearn.metrics import confusion_matrix, accuracy_score, average_precision_score, recall_score, precision_score\n",
    "from sklearn.model_selection import train_test_split, KFold, StratifiedKFold"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### PCA Data Transformation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [],
   "source": [
    "def pca_data_transform(X):\n",
    "    X_out = np.zeros_like(X)\n",
    "    for i in range(X_out.shape[0]):\n",
    "        X_sample = X[i]\n",
    "        X_reshape = X_sample.reshape(-1,X_sample.shape[2])\n",
    "        pca = decomposition.PCA(n_components=3)\n",
    "        X_pca = pca.fit_transform(X_reshape)\n",
    "        X_pca = X_pca.reshape(X_sample.shape)\n",
    "        X_out[i] = X_pca\n",
    "    \n",
    "    return X_out"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model Definitions and Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Dropout and Batchnorm Model\n",
    "def train_model(X_train, y_train, X_test, y_test, image_resize, Nber_of_breeds, dropout, fully_connected_nodes, learning_rate, epochs, batch_size):\n",
    "    # Create model\n",
    "    model = Sequential()\n",
    "    model.add(Conv2D(32, kernel_size=5, input_shape=(image_resize,image_resize,3)))\n",
    "    model.add(BatchNormalization())\n",
    "    model.add(Activation(\"relu\"))\n",
    "    model.add(MaxPool2D(pool_size=(2, 2), strides=(2,2), padding='same'))\n",
    "    \n",
    "    model.add(Dropout(dropout))\n",
    "\n",
    "    model.add(Conv2D(64, kernel_size=4))\n",
    "    model.add(BatchNormalization())\n",
    "    model.add(Activation(\"relu\"))\n",
    "    model.add(MaxPool2D(pool_size=(2, 2), strides=(2,2), padding='same'))\n",
    "    \n",
    "    model.add(Dropout(dropout))\n",
    "\n",
    "    model.add(Conv2D(128, kernel_size=3))\n",
    "    model.add(BatchNormalization())\n",
    "    model.add(Activation(\"relu\"))\n",
    "    model.add(MaxPool2D(pool_size=(2, 2), strides=(2,2), padding='same'))\n",
    "    \n",
    "    model.add(Dropout(dropout))\n",
    "\n",
    "    # model.add(Conv2D(256, kernel_size=3, activation='relu'))\n",
    "    # model.add(MaxPool2D(pool_size=(2, 2), strides=(2,2), padding='same'))\n",
    "\n",
    "    model.add(Dropout(dropout))\n",
    "    model.add(Flatten())\n",
    "    \n",
    "    model.add(Dense(fully_connected_nodes))\n",
    "    model.add(BatchNormalization())\n",
    "    model.add(Activation(\"relu\"))\n",
    "    \n",
    "    model.add(Dropout(dropout))\n",
    "\n",
    "    model.add(Dense(Nber_of_breeds, activation='relu'))\n",
    "\n",
    "    # Compile model\n",
    "    model_name = \"model_\" + str(image_resize) + \"x\" + str(image_resize) + \"_\" + str(Nber_of_breeds) + \"breeds_\" + str(dropout) + \"dropout_\" + str(learning_rate) + \"lr_\" + str(fully_connected_nodes) + \"fc_\" + str(batch_size) + \"batch\"\n",
    "    print(model_name)\n",
    "    \n",
    "    checkpointer = ModelCheckpoint(filepath=\"models/\" + model_name + \".hdf5\", verbose=1, save_best_only=True)\n",
    "    early_stopping = EarlyStopping(monitor='val_loss', mode='min', verbose=1, patience=5)\n",
    "    adam = keras.optimizers.Adam(learning_rate=learning_rate)\n",
    "    model.compile(optimizer=adam, loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "\n",
    "    # Train model\n",
    "    history = model.fit(X_train, y_train, validation_data=(X_test, y_test), epochs=epochs, batch_size=batch_size, callbacks=[checkpointer, early_stopping, TQDMNotebookCallback()], verbose=0)\n",
    "    \n",
    "    return model, history"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Optimal model\n",
    "def train_model(X_train, y_train, X_test, y_test, image_resize, Nber_of_breeds, dropout, fully_connected_nodes, learning_rate, epochs, batch_size):\n",
    "    # Create model\n",
    "    model = Sequential()\n",
    "    #add model layers\n",
    "    model.add(Conv2D(32, kernel_size=5, activation='relu', input_shape=(image_resize,image_resize,3)))\n",
    "    model.add(MaxPool2D(pool_size=(2, 2), strides=(2,2), padding='same'))\n",
    "\n",
    "    model.add(Conv2D(64, kernel_size=4, activation='relu'))\n",
    "    model.add(MaxPool2D(pool_size=(2, 2), strides=(2,2), padding='same'))\n",
    "\n",
    "    model.add(Conv2D(128, kernel_size=3, activation='relu'))\n",
    "    model.add(MaxPool2D(pool_size=(2, 2), strides=(2,2), padding='same'))\n",
    "\n",
    "    model.add(Dropout(dropout))\n",
    "    model.add(Flatten())\n",
    "    model.add(Dense(fully_connected_nodes, activation='relu'))\n",
    "    model.add(Dropout(dropout))\n",
    "\n",
    "    model.add(Dense(Nber_of_breeds, activation='softmax'))\n",
    "\n",
    "    # Compile model\n",
    "    model_name = \"model_\" + str(image_resize) + \"x\" + str(image_resize) + \"_\" + str(Nber_of_breeds) + \"breeds_\" + str(dropout) + \"dropout_\" + str(learning_rate) + \"lr_\" + str(fully_connected_nodes) + \"fc_\" + str(batch_size) + \"batch\"\n",
    "    print(model_name)\n",
    "    \n",
    "    checkpointer = ModelCheckpoint(filepath=\"models/\" + model_name + \".hdf5\", verbose=1, save_best_only=True)\n",
    "    early_stopping = EarlyStopping(monitor='val_loss', mode='min', verbose=1, patience=5)\n",
    "    adam = keras.optimizers.Adam(learning_rate=learning_rate)\n",
    "    model.compile(optimizer=adam, loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "\n",
    "    # Train model\n",
    "    history = model.fit(X_train, y_train, validation_data=(X_test, y_test), epochs=epochs, batch_size=batch_size, callbacks=[checkpointer, early_stopping, TQDMNotebookCallback()], verbose=0)\n",
    "    \n",
    "    return model, history"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model Evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [],
   "source": [
    "def evaluate_model(model, X_train, y_train, X_test, y_test, labels_name):\n",
    "    predictions = model.predict(X_test)\n",
    "    \n",
    "    pred_argmax = []\n",
    "    for pred in predictions:\n",
    "        pred_argmax.append(pred.argmax())\n",
    "\n",
    "    true_argmax = []\n",
    "    for pred in y_test:\n",
    "        true_argmax.append(pred.argmax())\n",
    "\n",
    "    accuracy = accuracy_score(true_argmax, pred_argmax)\n",
    "    conf_matrix = confusion_matrix(true_argmax, pred_argmax)\n",
    "    \n",
    "    df = pd.DataFrame(conf_matrix)\n",
    "    df = df.T\n",
    "    df.columns = labels_name\n",
    "    df = df.T\n",
    "    df.columns = labels_name\n",
    "    \n",
    "    y_true = np.array(true_argmax)\n",
    "    y_scores = np.array(pred_argmax)\n",
    "    \n",
    "    precision = precision_score(y_true, y_scores, average='weighted')\n",
    "    \n",
    "    recall = recall_score(y_true, y_scores, average='weighted')\n",
    "    \n",
    "    return accuracy, recall, precision, df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Hyperparameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Data params\n",
    "image_resize = 40\n",
    "Nber_of_breeds = 8\n",
    "use_pca = False\n",
    "\n",
    "# Model params\n",
    "dropout = 0.3\n",
    "fully_connected_nodes = 500\n",
    "learning_rate=0.0005\n",
    "# learning_rates= [0.0001, 0.0005, 0.001]\n",
    "\n",
    "# Training params\n",
    "epochs = 50\n",
    "batch_size = 20"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load data\n",
    "X = pickle.load(open(\"pickles/X_\" + str(image_resize) + \"x\" + str(image_resize) + \"_\" + str(Nber_of_breeds) + \".pkl\", \"rb\"))\n",
    "y = pickle.load(open(\"pickles/y_\" + str(image_resize) + \"x\" + str(image_resize) + \"_\" + str(Nber_of_breeds) + \".pkl\", \"rb\"))\n",
    "labels_name = pickle.load(open(\"pickles/labels_name_\" + str(image_resize) + \"x\" + str(image_resize) + \"_\" + str(Nber_of_breeds) + \".pkl\", \"rb\"))\n",
    "\n",
    "num_validation = 0.30\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=num_validation)\n",
    "\n",
    "# PCA\n",
    "if use_pca:\n",
    "    X_train = pca_data_transform(X_train)\n",
    "    X_test = pca_data_transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7ff16618e908>"
      ]
     },
     "execution_count": 130,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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.imshow(X_train[7])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "354a266106744691b567577e7592bc5c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=645, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.05475, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=645, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 2.05475 to 1.88998, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=645, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 1.88998 to 1.80236, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=645, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss improved from 1.80236 to 1.77288, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=645, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss improved from 1.77288 to 1.71802, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=645, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss did not improve from 1.71802\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=645, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss improved from 1.71802 to 1.68646, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=645, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 1.68646\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=645, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 1.68646\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 9', max=645, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00010: val_loss did not improve from 1.68646\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 10', max=645, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00011: val_loss did not improve from 1.68646\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 11', max=645, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00012: val_loss did not improve from 1.68646\n",
      "Epoch 00012: early stopping\n"
     ]
    },
    {
     "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": [
    "model, history = train_model(X_train, y_train, X_test, y_test, image_resize, Nber_of_breeds, dropout, fully_connected_nodes, learning_rate, epochs, batch_size)\n",
    "\n",
    "plt.plot(history.history[\"val_accuracy\"], label = \"val_accuracy\")\n",
    "plt.plot(history.history[\"accuracy\"], label = \"accuracy\")\n",
    "plt.legend();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "# To train in a loop\n",
    "# for learning_rate in learning_rates:\n",
    "#     train_model(image_resize, Nber_of_breeds, dropout, fully_connected_nodes, learning_rate, epochs, batch_size)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Evaluate Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.4151624548736462\n",
      "Recall: 0.4151624548736462\n",
      "Precision: 0.4193380860348369\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>afghan_hound</th>\n",
       "      <th>bernese_mountain_dog</th>\n",
       "      <th>entlebucher</th>\n",
       "      <th>great_pyrenees</th>\n",
       "      <th>maltese_dog</th>\n",
       "      <th>pomeranian</th>\n",
       "      <th>scottish_deerhound</th>\n",
       "      <th>shih-tzu</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>afghan_hound</th>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bernese_mountain_dog</th>\n",
       "      <td>5</td>\n",
       "      <td>19</td>\n",
       "      <td>7</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>entlebucher</th>\n",
       "      <td>4</td>\n",
       "      <td>12</td>\n",
       "      <td>10</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>great_pyrenees</th>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>10</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>maltese_dog</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>14</td>\n",
       "      <td>6</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>pomeranian</th>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "      <td>17</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>scottish_deerhound</th>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>25</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>shih-tzu</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>6</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      afghan_hound  bernese_mountain_dog  entlebucher  \\\n",
       "afghan_hound                    10                     2            0   \n",
       "bernese_mountain_dog             5                    19            7   \n",
       "entlebucher                      4                    12           10   \n",
       "great_pyrenees                   2                     1            0   \n",
       "maltese_dog                      1                     0            0   \n",
       "pomeranian                       2                     2            0   \n",
       "scottish_deerhound               5                     2            2   \n",
       "shih-tzu                         1                     2            2   \n",
       "\n",
       "                      great_pyrenees  maltese_dog  pomeranian  \\\n",
       "afghan_hound                       6            1           3   \n",
       "bernese_mountain_dog               1            1           0   \n",
       "entlebucher                        1            1           0   \n",
       "great_pyrenees                    10            8           2   \n",
       "maltese_dog                        3           14           6   \n",
       "pomeranian                         5            3          17   \n",
       "scottish_deerhound                 1            5           2   \n",
       "shih-tzu                           6            7           2   \n",
       "\n",
       "                      scottish_deerhound  shih-tzu  \n",
       "afghan_hound                           8         2  \n",
       "bernese_mountain_dog                   3         2  \n",
       "entlebucher                            2         5  \n",
       "great_pyrenees                         5         2  \n",
       "maltese_dog                            2         3  \n",
       "pomeranian                             4         3  \n",
       "scottish_deerhound                    25         0  \n",
       "shih-tzu                               5        10  "
      ]
     },
     "execution_count": 120,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy, recall, precision, conf_matrix = evaluate_model(model, X_train, y_train, X_test, y_test, labels_name)\n",
    "print(\"Accuracy:\", accuracy)\n",
    "print(\"Recall:\", recall)\n",
    "print(\"Precision:\", precision)\n",
    "conf_matrix"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### K-Fold Cross Validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training split: 1 of 5\n",
      "model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "87fdc9fdd384428980114f79eba0b75c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 9.20977, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 9.20977 to 6.44615, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 6.44615 to 5.50050, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss did not improve from 5.50050\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss did not improve from 5.50050\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss did not improve from 5.50050\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss did not improve from 5.50050\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 5.50050\n",
      "Epoch 00008: early stopping\n",
      "Evaluating split: 1 of 5\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Recording metrics: 1 of 5\n",
      "\n",
      "\n",
      "Training split: 2 of 5\n",
      "model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e32a209b8d4e438a9323f42ad1ebf02e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 5.36747, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 5.36747 to 5.34259, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss did not improve from 5.34259\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss did not improve from 5.34259\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss did not improve from 5.34259\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss did not improve from 5.34259\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss did not improve from 5.34259\n",
      "Epoch 00007: early stopping\n",
      "Evaluating split: 2 of 5\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Recording metrics: 2 of 5\n",
      "\n",
      "\n",
      "Training split: 3 of 5\n",
      "model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "412414feec9d40b5ae9186bcfcddb62f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 6.57207, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 6.57207 to 4.48007, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 4.48007 to 3.71610, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss improved from 3.71610 to 3.67070, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss improved from 3.67070 to 3.66861, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss improved from 3.66861 to 3.64183, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss did not improve from 3.64183\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 3.64183\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss improved from 3.64183 to 3.38966, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 9', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00010: val_loss improved from 3.38966 to 2.68584, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 10', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00011: val_loss improved from 2.68584 to 2.53080, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 11', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00012: val_loss did not improve from 2.53080\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 12', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00013: val_loss did not improve from 2.53080\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 13', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00014: val_loss improved from 2.53080 to 2.49303, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 14', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00015: val_loss did not improve from 2.49303\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 15', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00016: val_loss did not improve from 2.49303\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 16', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00017: val_loss improved from 2.49303 to 2.48130, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 17', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00018: val_loss improved from 2.48130 to 2.46419, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 18', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00019: val_loss did not improve from 2.46419\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 19', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00020: val_loss did not improve from 2.46419\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 20', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00021: val_loss did not improve from 2.46419\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 21', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00022: val_loss did not improve from 2.46419\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 22', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00023: val_loss did not improve from 2.46419\n",
      "Epoch 00023: early stopping\n",
      "Evaluating split: 3 of 5\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Recording metrics: 3 of 5\n",
      "\n",
      "\n",
      "Training split: 4 of 5\n",
      "model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0dde7a1307744c37931b6916b0febc61",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 6.10793, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 6.10793 to 2.81479, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss did not improve from 2.81479\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss improved from 2.81479 to 2.46329, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss did not improve from 2.46329\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss did not improve from 2.46329\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss did not improve from 2.46329\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 2.46329\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 2.46329\n",
      "Epoch 00009: early stopping\n",
      "Evaluating split: 4 of 5\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Recording metrics: 4 of 5\n",
      "\n",
      "\n",
      "Training split: 5 of 5\n",
      "model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "713a75c9ded14b139928a69c188d2128",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 5.47242, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss did not improve from 5.47242\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 5.47242 to 3.90814, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss improved from 3.90814 to 2.56823, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss improved from 2.56823 to 2.41345, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss improved from 2.41345 to 2.33973, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss improved from 2.33973 to 2.33780, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 2.33780\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 2.33780\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 9', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00010: val_loss did not improve from 2.33780\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 10', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00011: val_loss did not improve from 2.33780\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 11', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00012: val_loss improved from 2.33780 to 2.33738, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 12', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00013: val_loss improved from 2.33738 to 2.32823, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 13', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00014: val_loss improved from 2.32823 to 2.31171, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 14', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00015: val_loss improved from 2.31171 to 2.30321, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 15', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00016: val_loss improved from 2.30321 to 2.29605, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 16', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00017: val_loss improved from 2.29605 to 2.29566, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 17', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00018: val_loss did not improve from 2.29566\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 18', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00019: val_loss did not improve from 2.29566\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 19', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00020: val_loss did not improve from 2.29566\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 20', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00021: val_loss did not improve from 2.29566\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 21', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00022: val_loss did not improve from 2.29566\n",
      "Epoch 00022: early stopping\n",
      "Evaluating split: 5 of 5\n",
      "Recording metrics: 5 of 5\n",
      "\n",
      "\n",
      "Average Mean: 0.1290716803760282\n",
      "Average Recall: 0.1290716803760282\n",
      "Average Precision: 0.028480913656350693\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    }
   ],
   "source": [
    "n_splits = 5\n",
    "\n",
    "\n",
    "accuracies = []\n",
    "recalls = []\n",
    "precisions = []\n",
    "conf_matrices = []\n",
    "\n",
    "# K-Fold cross validation\n",
    "i = 0\n",
    "kf = KFold(n_splits=n_splits)\n",
    "for train, test in kf.split(X, y):\n",
    "    i += 1\n",
    "    # Train model\n",
    "    print(\"Training split:\", i, \"of\", n_splits)\n",
    "#     model, history = train_model(X[train], y[train], X[test], y[test], image_resize, Nber_of_breeds, dropout, fully_connected_nodes, learning_rate, epochs, batch_size)\n",
    "    model, history = train_model(pca_data_transform(X[train]), y[train], pca_data_transform(X[test]), y[test], image_resize, Nber_of_breeds, dropout, fully_connected_nodes, learning_rate, epochs, batch_size)\n",
    "    \n",
    "    # Evaluate model\n",
    "    print(\"Evaluating split:\", i, \"of\", n_splits)\n",
    "    accuracy, recall, precision = evaluate_model(model, X[train], y[train], X[test], y[test], labels_name)\n",
    "    \n",
    "    # Record metrics\n",
    "    print(\"Recording metrics:\", i, \"of\", n_splits)\n",
    "    accuracies.append(accuracy)\n",
    "    recalls.append(recall)\n",
    "    precisions.append(precision)\n",
    "#     conf_matrices.append(conf_matrix)\n",
    "    \n",
    "    print(\"\\n\")\n",
    "    \n",
    "# Print average metrics\n",
    "print(\"Average Mean:\", np.asarray(accuracies).mean())\n",
    "print(\"Average Recall:\", np.asarray(recalls).mean())\n",
    "print(\"Average Precision:\", np.asarray(precisions).mean())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Compare Models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [],
   "source": [
    "wout_pca = 0.40457109283196246\n",
    "w_pca = 0.14421856639247946\n",
    "w_batchnorm = 0.11919506462984723\n",
    "w_batchnorm_pca = 0.1290716803760282"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEiCAYAAADjxEWuAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nO3deZwdVZ338c+XRISEIEIiSxIIQgYEDIgxIOASHZSgGFBGtgHBJWYQFR0ZGR+GRVEHZxSHedCADkZ4wLhAMEjYRJZRQNJIDAQTCRFIG5Cwhz0Jv+ePczpUbqq763a6+jad7/v1uq++VXVO1bnnVtfvVp06pxQRmJmZNdqg1QUwM7P+yQHCzMxKOUCYmVkpBwgzMyvlAGFmZqUcIMzMrJQDhJmZlXKAGAAk3S/peUnPSPqbpB9J2qSw/P2Sbpa0XNIySTdJ+lDDOt4tKST9Sx+X/UZJT0h6bV9ut26SpktaKWmbVpelVXIdnFmY3lXSQ5L+ucn13Cjphbx/PyNpYTfpt5b0P3lbyyUtkHSGpKH5/cdL8nxeUlsz5VofOEAMHAdFxCbAnsDbgFMAJB0K/By4EBgFbAmcChzUkP9jwOP5b5+QNAZ4BxDAh7pM3PNtDK5jvd1scyjwEeAp4Kg+3naff94qJO0B3AB8PSK+3YNVnBARm+TXTl1sZ3PgVmBj4O0RMQzYH9gM2AH4MXBMSdaj8zIrigi/XuUv4H7g7wvT/wH8ChDwIHBSN/mHAMuBw4GXgPFdpP0T8MHC9GDgUVJg2gj4f8BjwJPAHGDLLtZ1KvA74DvArwrz9wYeBgYV5h0CzMvvNwBOBu7L2/oZsHleNoYUcD6RP/vNef7P8zqfAm4Gdi2sewvgCuDpXOYzgd8Wlu8MXEcKoAuBj3ZTn8cAS4DPA3c3LBsEfCWXfTlwBzA6L9u1sJ2/AV/J86cDZxbW8W6gveH7/zIwD3gxfycnF7ZxD3BIQzk+lb/LjuV7AicBlzak+2/guz3cL6fnupyQ95FP9nA9N1bNm7d3F7BBJ8tHASuB7Qrz3kTa74e34v+3P79aXgC/euFLLAQIYDQwH/haPrAFsH03+Y8GHsoHryuAc7pIeypwcWH6A8CC/P7TOf+QvK63Apt2sa5FwPE53QoKwSQf3PYvTP8cODm/PxG4Lf+zvxY4D/hJXjYmf+YLgaHAxnn+x4FhOf13gbmFdc/IryHALqSD+2/zsqF5+rh84N0zH+x27eJzXQ98i3S2thLYs7DspHwA24kUwHcnBahh+Tv4Z1KgHQbslfNMp/sAMTd/9x2f9x+AbUjB9DDgWWDrwrK/ks40BewIbAdsndNtltMNBh4B3trD/XI6cC0p4B1dsvxXpB8SZa/iD4YbgWW53n8HvLuLbd4GnNFNua4DTilMfxO4vNX/x/3x1fIC+NULX2I6QDyT/7EeAL5HOsXeNx8sN+om/6/JvxKBI/I/42s6Sbsj6VfnkDx9MXBqfv9x4BZgXIUy70cKCsPz9ALgC4XlZwIX5PfD8oFruzz9J+C9hbRb53UN5pUA8cYutr1ZTvM6UiBbAezUsO2OAHEY8L8N+c8DTutk3dsCLwN75OlrgP8qLF8ITC7JdwRwZyfrnE73AeLj3dT33I7t5jJ9vpN0VwGfyu8/CNyzDvvldNJZ2V9Yh1/nwF68Etw/lve/HTpJey8wtZv1/SOwML/fgHSmeUhPyzeQX26DGDgOjojNImK7iDg+Ip4nXX6BdAAtJWk0MJF0oAf4JekX7AfK0kfEItIB+iBJQ0htB5fkxReRDj4zJC2V9C1Jr+lk0x8Dro2IR/P0JazZ/nEJ8OHceP1h4A8R8UBeth0wU9KTkp7M5VlF+sXeYUnhMw6S9O+S7pP0NOmACjAcGEEKLEvK8uZt7dWxrby9o4CtOvlcRwN/ioi5efpi4MhCPYwmnR016mx+VcUyI+kYSXMLZd6N9Hm729aPSQdQ8t+LyhJJOqrQaHxVF+U6l3TZ7jpJr6/4WdYQEb+PiOUR8WJE/Jh0FnFgJ8kfo4v9PbsM2FrS3qRgOwS4sidlG+gcIAa2haQDx0e6SHM0aT+4QtLDwGJSgChryOvwE9Iv3smkX5iLACJiRUScERG7APuQfoGutR5JGwMfBd4l6eG83S8Au0vaPa/rHtLZ0CTgSF4JQuTPNCkHxI7XRhHx10Ka4jDFR+ay/j3prGFMR1FIZ0srSZerOoxu2NZNDdvaJCL+qZO6OQZ4Y+FzfYd0YJ5UWN8OJfk6mw/p7GlIYbosOK3+vJK2A34AnABsERGbAXeTPm9327ocGCdpN9L3d3FZooi4OF5pNJ5UliZbRQqoDwLXSNq0UM6rCkGm8dVV0InCZ2n0a+AQSZ0e2yLiOeAXpO/qaGBGRLzUxfbWX60+hfFr3V80NFI3LDuU1DB7HLApKRjsB5yfly8ATicddDpeHyI1dm7RyTq3Bp4jNfZ+vjB/IvBm0mWbzYE/AseW5D+CdF1624bt3gx8u5Duy8BvgOcpXKIgBZMbeeWS0wheuXwyhnQAGVxIfzzpEsumpDaF7+U0O+blPyUFoCGkdpsHeeUS0zBSoDoaeE1+vQ14U8nnejsp2Ly54XNdTG78JbVBzAPGkg5y41izDeJE0qWUYhvEp/L3tHle322sfYmpeJPCLsALpHaOQfm7X0lu6CW1QSwhtf2sboMo5P9BLuNv1nG/nE6+NAZsCMwGfgsMbWIdmwHvJ/1oGUwKNs9SuCTYkH7zXB8XFfaPkaRAPa6Q7l2ks42ngbe1+n+4v75aXgC/euFL7CJA5OUHAP9LaqdYlg+uHyDdLfQCMKIkz3zSrYWdrfP6fNDZqjDvCNJZy7Oku3DOoXCgLqS7mkIgKMz/KOlOo8F5uuN6/pUN6TYAvpi3tZx0ueQbedkY1g4Qm5AunS0nHeyPYc0AMYJ0iaHjLqazgOsL+XfKy5flg8pvyG0MDeWaRsNdQHn+BFLA3TwfsE8hXZdfnrc3KqfbLdfrE7keOhrlNyIFsadJB+4v0EWAyPO+TgrCj+aD400U7gQCpub6e4Z0dvGWwrL9cv0ct4775XTWbDvZiPQL/zfkxvQK6xiR62g5qY3tNgo3L3SSZxvgglyHy0nB9TRyu1lOI9LZ8p9a/f/bn1/KlWVmmaSzSIHvY90mHoAkbUs6qG4VEU+3ujzWOm6DsPWepJ0ljVMygdSHYmary9UK+dr9F0nX5R0c1nP9stelWR8bRmp434Z03/+3SZek1iu5B/jfSJfhDmhxcawf8CUmMzMr5UtMZmZWygHCzMxKDag2iOHDh8eYMWNaXQwzs1eNO+6449GIGFG2bEAFiDFjxtDW5iHdzcyqkvRAZ8t8icnMzEo5QJiZWSkHCDMzK+UAYWZmpRwgzMyslAOEmZmVcoAwM7NStQYISQdIWihpkaSTu0j3NkmrJB3abF4zM6tHbR3lJA0iPY92f6AdmCNpVqRHSTamO4v0LOOm8vZqec/o7AmG64c4zYM2mtma6jyDmAAsiojFkZ73OoP0XOBGnwUuJQ2z3GxeMzOrSZ0BYiTpubcd2vO81SSNBA4hPaqxqbyFdUyR1CapbdmyZetcaDMzS+oMEGXXbBqvY3wX+HJErOpB3jQz4vyIGB8R40eMKB1vyszMeqDOwfragdGF6VHA0oY044EZkgCGAwdKWlkxr5mZ1ajOADEHGCtpe+CvwOHAkcUEEbF9x3tJ04FfRcTlkgZ3l9fMzOpVW4CIiJWSTiDdnTQIuCAi5kuampc3tjt0m7euspqZ2dpqfR5ERMwGZjfMKw0MEXFsd3nNzKzvuCe1mZmVcoAwM7NSDhBmZlbKAcLMzEo5QJiZWSkHCDMzK+UAYWZmpRwgzMyslAOEmZmVcoAwM7NSDhBmZlbKAcLMzEo5QJiZWSkHCDMzK+UAYWZmpRwgzMysVK0BQtIBkhZKWiTp5JLlkyXNkzRXUpuk/QrL7pd0V8eyOstpZmZrq+2JcpIGAecC+wPtwBxJsyLinkKy64FZERGSxgE/A3YuLJ8YEY/WVUYzM+tcnWcQE4BFEbE4Il4CZgCTiwki4pmIiDw5FAjMzKxfqDNAjASWFKbb87w1SDpE0gLgSuDjhUUBXCvpDklTaiynmZmVqDNAqGTeWmcIETEzInYGDga+Vli0b0TsCUwCPiPpnaUbkabk9ou2ZcuW9Ua5zcyMegNEOzC6MD0KWNpZ4oi4GdhB0vA8vTT/fQSYSbpkVZbv/IgYHxHjR4wY0VtlNzNb79UZIOYAYyVtL2lD4HBgVjGBpB0lKb/fE9gQeEzSUEnD8vyhwPuAu2ssq5mZNajtLqaIWCnpBOAaYBBwQUTMlzQ1L58GfAQ4RtIK4HngsHxH05bAzBw7BgOXRMTVdZXVzMzWVluAAIiI2cDshnnTCu/PAs4qybcY2L3OspmZWdfck9rMzEo5QJiZWSkHCDMzK+UAYWZmpRwgzMyslAOEmZmVcoAwM7NSDhBmZlbKAcLMzEo5QJiZWanKAULSQZJ+nx8BenydhTIzs9brNEBIahwL6Whgb2BP4J/qLJSZmbVeV4P1HZ+H4j41Ih4mPR3u68DLdPFcBzMzGxg6DRAR8el8FnGepDbg34B9gCGs+eQ3MzMbgLpsg4iIP0bEZGAu6WE/W0fErIh4sU9KZ2ZmLdNVG8RUSXdK+gMwFDgAeL2kayS9o89KaGZmLdHVGcTxEfEWUsP0SRGxMiLOIT069JA+KZ2ZmbVMVwHir5K+BnwDWNAxMyKeiIgvVlm5pAMkLZS0SNLJJcsnS5qXb51tk7Rf1bxmZlavru5imgy8H1gBXNfsiiUNAs4F9gfagTmSZkXEPYVk1wOz8nOoxwE/A3aumNfMzGrU1V1MLwFXrMO6JwCL8vOlkTSDFHRWH+Qj4plC+qFAVM1rZmb1qnOojZGkvhMd2vO8NUg6RNIC4Erg483kNTOz+tQZIFQyL9aaETEzInYGDuaV/hWV8gJImpLbL9qWLVvW48KamdmaKgUISYMkbSNp245XhWztwOjC9Ci66IEdETcDO0ga3kzeiDg/IsZHxPgRI0ZUKJaZmVXRVSM1AJI+C5wG/I00zAakX/Pjusk6BxgraXvgr6TbY49sWPeOwH25kXpPYEPgMeDJ7vKamVm9ug0QwOeBnSLisWZWHBErJZ0AXAMMAi6IiPmSpubl04CPAMdIWgE8DxwWEQGU5m1m+2Zmtm6qBIglwFM9WXlEzAZmN8ybVnh/FnBW1bxmZtZ3qgSIxcCNkq4EVo/BFBHfqa1UZmbWclUCxIP5tWF+mZnZeqDbABERZ/RFQczMrH/pNEBI+m5EnCjpCsr7L3yo1pKZmVlLdXUGcVH++599URAzM+tfuhqL6Y7896a+K46ZmfUXdQ61YWZmr2IOEGZmVqrbACFpt74oiJmZ9S9VziCmSbpd0vGSNqu9RGZm1i90GyAiYj/gKNLoqm2SLpG0f+0lMzOzlqrUBhER9wKnAF8G3gWcI2mBpA/XWTgzM2udKm0Q4ySdDfwJeA9wUES8Kb8/u+bymZlZi1QZi+n/Aj8AvhIRz3fMjIilkk6prWRmZtZSVQLEgcDzEbEKQNIGwEYR8VxEXNR1VjMze7Wq0gbxa2DjwvSQPM/MzAawKgFio4h4pmMivx9SX5HMzKw/qBIgns3PiwZA0ltJjwftlqQDJC2UtEjSySXLj5I0L79ukbR7Ydn9ku6SNFdSW5XtmZlZ76nSBnEi8HNJS/P01sBh3WWSNAg4F9gfaAfmSJoVEfcUkv0FeFdEPCFpEnA+sFdh+cSIeLRCGc3MrJdVeWDQHEk7AzsBAhZExIoK654ALIqIxQCSZgCTgdUBIiJuKaS/DRjVRNnNzKxGVQfr2wnYBXgLcISkYyrkGQksKUy353md+QRwVWE6gGsl3SFpSmeZJE2R1CapbdmyZRWKZWZmVXR7BiHpNODdpAAxG5gE/Ba4sLusJfPWejJd3sZEUoDYrzB739zX4g3AdZIWRMTNa60w4nzSpSnGjx9fun4zM2telTOIQ4H3Ag9HxHHA7sBrK+RrJ43f1GEUsLQxkaRxwA+ByRHxWMf8iFia/z4CzCRdsjIzsz5SJUA8HxEvAyslbQo8AryxQr45wFhJ20vaEDgcmFVMIGlb4DLg6Ij4c2H+UEnDOt4D7wPurvKBzMysd1S5i6ktD/P9A+AO4Bng9u4yRcRKSScA1wCDgAsiYr6kqXn5NOBUYAvge5IAVkbEeGBLYGaeNxi4JCKubvbDmZlZzymi88v2SkfoURGxJE+PATaNiHl9UromjR8/PtraetZlQmeUNZmsP+I0N9+YrY8k3ZF/mK+ly0tMkaLH5YXp+/trcDAzs95VpQ3iNklvq70kZmbWr1Rpg5gIfFrSA8CzpNtXIyLG1VoyMzNrqSoBYlLtpTAzs36nSoBw66WZ2XqoSoC4khQkBGwEbA8sBHatsVxmZtZiVQbre3NxOg/9/enaSmRmZv1C1cH6VouIPwC+q8nMbICrMljfFwuTGwB7Ah421cxsgKvSBjGs8H4lqU3i0nqKY2Zm/UWVNogz+qIgZmbWv3TbBiHpujxYX8f06yVdU2+xzMys1ao0Uo+IiCc7JiLiCeAN9RXJzMz6gyoBYlV+bgMAkrbDnefMzAa8Ko3U/wf4raSb8vQ7gU6fEW1mZgNDlUbqq3PnuL1Jvam/EBGP1l4yMzNrqSqN1IcAKyLiVxFxBenRowdXWbmkAyQtlLRI0skly4+SNC+/bpG0e9W8ZmZWryptEKdFxFMdE7nB+rTuMkkaBJxLGg12F+AISbs0JPsL8K48dPjXgPObyGtmZjWqEiDK0lRpu5gALIqIxRHxEjADmFxMEBG35LuiAG4DRlXNa2Zm9aoSINokfUfSDpLeKOls4I4K+UYCSwrT7XleZz4BXNXDvGZm1suqBIjPAi8BPwV+DrwAHF8hn0rmld4eK2kiKUB8uQd5p0hqk9S2bJmHiDIz6y1V7mJ6FljdSCxpI+AgUrDoSjswujA9CljamEjSOOCHwKSIeKyZvLl855PbLsaPH+/+GWZmvaTScN+SBkmaJOlC4H7gsArZ5gBjJW0vaUPgcGBWw3q3BS4Djo6IPzeT18zM6tXlGYSkdwJHAh8Abgf2Bd4YEc91t+KIWCnpBOAaYBBwQUTMlzQ1L58GnApsAXxPEsDKiBjfWd6efkgzM2tepwFCUjvwIPB94KSIWC7pL1WCQ4eImA3Mbpg3rfD+k8Anq+Y1M7O+09UlpktJdw4dBhwkaSgeg8nMbL3RaYCIiM8DY4DvABOBPwMjJH1U0iZ9UzwzM2uVLhupI/lNRHyKFCyOBA4mNVSbmdkAVqVHNAARsQK4ArhC0sb1FcnMzPqDSre5NoqI53u7IGZm1r/0KECYmdnAVzlA5LuYzMxsPVHleRD7SLoH+FOe3l3S92ovmZmZtVSVM4izgfcDjwFExB9Jjx01M7MBrNIlpohY0jBrVQ1lMTOzfqTKba5LJO0DRB4473Pky01mZjZwVTmDmAp8hjTsRjuwR542M7MBrMrzIB4FjuqDspiZWT/SbYCQdE7J7KeAtoj4Ze8XyczM+oMql5g2Il1Wuje/xgGbA5+Q9N0ay2ZmZi1UpZF6R+A9EbESQNL3gWuB/YG7aiybmZm1UJUziJFAsRf1UGCbiFgFvFhLqczMrOWqBIhvAXMl/UjSdOBO4D/z0Bu/7iqjpAMkLZS0SNLJJct3lnSrpBclfalh2f2S7pI0V1Jb9Y9kZma9ocpdTP8jaTYwARDwlYhYmhef1Fk+SYOAc0mXotqBOZJmRcQ9hWSPk/pVHNzJaibmu6jMzKyPVR2s7wXgIdIBfUdJVYbamAAsiojFEfESMAOYXEwQEY9ExBxgRRNlNjOzPlBlsL5PAjcD1wBn5L+nV1j3SKA4REd7nldVANdKukPSlC7KN0VSm6S2ZcuWNbF6MzPrSpUziM8DbwMeiIiJwFuAKkdilcyLJsq2b0TsCUwCPtPZWUtEnB8R4yNi/IgRI5pYvZmZdaVKgHghIl4AkPTaiFgA7FQhXzswujA9CljaSdq1dLRzRMQjwEzSJSszM+sjVfpBtEvaDLgcuE7SE1Q70M8BxkraHvgrcDhwZJVC5TukNoiI5fn9+4CvVslrraEzyk4Y1x9xWjMnx2avDlXuYjokvz1d0g3A64CrK+RbKekEUpvFIOCCiJgvaWpePk3SVkAbsCnwsqQTgV2A4cBMSR1lvCQiut2mmZn1ni4DhKQNgHkRsRtARNzUzMojYjYwu2HetML7h0mXnho9DezezLbMzKx3ddkGEREvA3+UtG0flcfMzPqJKm0QWwPzJd0OPNsxMyI+VFupzMys5aoEiDNqL4WZmfU7VRqpb5K0HTA2In4taQip0dnMzAawKj2pPwX8AjgvzxpJuuXVzMwGsCqXmD5D6qT2e4CIuFfSG2otlZlZM7R+98Mh6umHU6Un9Yt5sD0AJA2muSEzzMzsVahKgLhJ0leAjSXtD/wcuKLeYpmZWatVCRAnkwbnuwv4NKnj2yl1FsrMzFqvShvEZODCiPhB3YUxM7P+o8oZxIeAP0u6SNIHchuEmfUmaf1+Wb/UbYCIiOOAHUltD0cC90n6Yd0FMzOz1qp0NhARKyRdRbp7aWPSZadP1lkwMzNrrSod5Q6QNB1YBBwK/JA0PpOZmQ1gVc4gjgVmAJ+OiBfrLY6ZmfUXVcZiOrw4LWlf4MiI+ExtpTIzs5archcTkvaQ9C1J9wNnAgsq5jtA0kJJiySdXLJ8Z0m3SnpR0peayWtmZvXq9AxC0t+RniN9BPAY8FNAETGxyoolDQLOBfYH2oE5kmZFxD2FZI8DnwMO7kFeMzOrUVdnEAuA9wIHRcR+EfHfwKom1j0BWBQRi/NYTjNIdz+tFhGPRMQcYEWzec3MrF5dBYiPAA8DN0j6gaT3As30aBkJLClMt+d5dec1M7Ne0GmAiIiZEXEYsDNwI/AFYEtJ35f0vgrrLgsmVUeBrZxX0hRJbZLali1bVnH1ZmbWnSo9qZ+NiIsj4oPAKGAuaQC/7rQDowvTo4ClFctVOW9EnB8R4yNi/IgRIyqu3szMulPpLqYOEfF4RJwXEe+pkHwOMFbS9pI2JDV4z6q4qXXJa2ZmvaC2gfciYqWkE4BrSM+wviAi5kuampdPk7QV0AZsCrws6URgl4h4uixvXWU1M7O11Toya0TMJj0/ojhvWuH9w6TLR5XymplZ32nqEpOZma0/HCDMzKyUA4SZmZVygDAzs1IOEGZmVsoBwszMSjlAmJlZKQcIMzMr5QBhZmalHCDMzKyUA4SZmZVygDAzs1IOEGZmVsoBwszMSjlAmJlZKQcIMzMr5QBhZmalag0Qkg6QtFDSIkknlyyXpHPy8nmS9iwsu1/SXZLmSmqrs5xmZra22h45KmkQcC6wP9AOzJE0KyLuKSSbBIzNr72A7+e/HSZGxKN1ldHMzDpX5xnEBGBRRCyOiJeAGcDkhjSTgQsjuQ3YTNLWNZbJzMwqqjNAjASWFKbb87yqaQK4VtIdkqbUVkozMytV2yUmQCXzook0+0bEUklvAK6TtCAibl5rIyl4TAHYdttt16W8ZmZWUOcZRDswujA9ClhaNU1EdPx9BJhJumS1log4PyLGR8T4ESNG9FLRzcyszgAxBxgraXtJGwKHA7Ma0swCjsl3M+0NPBURD0kaKmkYgKShwPuAu2ssq5mZNajtElNErJR0AnANMAi4ICLmS5qal08DZgMHAouA54DjcvYtgZmSOsp4SURcXVdZzcxsbXW2QRARs0lBoDhvWuF9AJ8pybcY2L3OspmZWdfck9rMzEo5QJiZWSkHCDMzK+UAYWZmpRwgzMyslAOEmZmVcoAwM7NSDhBmZlbKAcLMzEo5QJiZWSkHCDMzK+UAYWZmpRwgzMyslAOEmZmVcoAwM7NSDhBmZlbKAcLMzErVGiAkHSBpoaRFkk4uWS5J5+Tl8yTtWTWvmZnVq7YAIWkQcC4wCdgFOELSLg3JJgFj82sK8P0m8pqZWY3qPIOYACyKiMUR8RIwA5jckGYycGEktwGbSdq6Yl4zM6vR4BrXPRJYUphuB/aqkGZkxbwASJpCOvsAeEbSwnUocysNBx5t1cZ1ulq16d7i+ls3La0/5PpbJ+tWf9t1tqDOAFFW4qiYpkreNDPifOD85orW/0hqi4jxrS7Hq5Xrb924/tbNQK2/OgNEOzC6MD0KWFoxzYYV8pqZWY3qbIOYA4yVtL2kDYHDgVkNaWYBx+S7mfYGnoqIhyrmNTOzGtV2BhERKyWdAFwDDAIuiIj5kqbm5dOA2cCBwCLgOeC4rvLWVdZ+4lV/mazFXH/rxvW3bgZk/Smi9NK+mZmt59yT2szMSjlAmJlZKQeIHpI0StIvJd0r6T5J/5Ub1LvKs5mk4wvT20j6RS+V53RJX+qNdfUF11/XJK2SNFfSHyX9QdI+3aRfo266SHejpAF3O2Yj11/vcIDoAUkCLgMuj4ixwN8BmwBf7ybrZsDqnTAilkbEobUVtJ9y/VXyfETsERG7A/8KfLOb9GvUTV/Lw+P0J66/XuAA0TPvAV6IiB8BRMQq4AvAxyUNkXRs/nV8dR5w8LSc79+BHfIvm/+QNEbS3QA5z+WSrpD0F0knSPqipDsl3SZp85zuU5Lm5F9Gl0oa0oLPv65cf83ZFHgCQNImkq7Pv4rvktQxBM0adZPT/ktO80dJ/15Y3z9Iul3SnyW9I6c9VtJluc7vlfStjsSSjsjruVvSWYX5z0j6qqTfA2+XdL+kb0i6VVKbpD0lXZPPEKfWXEddcf31VET41eQL+Bxwdsn8O4FxwLHAQ8AWwMbA3cB4YAxwdyH96umcZxEwDBgBPAVMzcvOBk7M77co5D8T+Gx+fzrwpVbXjeuv1+poFTAXWJA/y1vz/MHApvn98PyZVVI3k4BbgCF5evP890bg2/n9gcCvC/W3GHgdsBHwAKmz6jbAg7lOBwO/AQ7OeQL4aGGb9wP/VKjzeYXv4xHX36un/jpedfakHshE+dAfxfnXRcRjAJIuA/YDLu9mvTdExHJguaSngCvy/LtIB06A3SSdSTol3oTUV+TVxvXXvclNYNUAAAc+SURBVOcjYg8ASW8HLpS0G6mOviHpncDLpHHLtizJ//fAjyLiOYCIeLyw7LL89w7SgbHD9RHxVN7mPaQxerYAboyIZXn+xcA7Sd/FKuDShu12dGi9C9ik8H28IGmziHiyuWroMddfL/Alpp6ZT/pFu5qkTUm/GO7LsxoPgFU6nLxYeP9yYfplXunUOB04ISLeDJxB+rXyauP6a0JE3Er6tTsCOCr/fWs+AP6N8s/QWRCGV+plFWt2li3WX8eyrkaBeyHS5cGydRfrv2O6JT9IXX895wDRM9cDQyQdA6sbmL4NTO/4xQHsL2lzSRsDBwO/A5aTThnXxTDgIUmvIe3sr0auvyZI2pk0osBjpEsYj0TECkkTeWUkzsa6uZbcppPXsXkPN/974F2Shufv6Qjgph6uqyVcfz3nS0w9EBEh6RDge5L+jRRoZwNfKST7LXARsCNwSUS0AUj6nVLD6lWkhyI1699IO90DpNPQdT1g9jnXXyUbS5qb3wv4WESsypcorpDUxivX2ImIx4p1ExEnSdoDaJP0EmvXbyUR8ZCkfwVuyOWYHRG/XPePVzvXXy/wUBs1kHQsMD4iTmh1WV6NXH9m/YMvMZmZWSmfQZiZWSmfQRRIOlvSiYXpayT9sDD9baXOV6uHeJC0h6QDC2l6bciG3Plmm06WTVfqEDY3d/p5e2HZlyQtyB1z/tjRGJyXjZC0QtKne6OMzejL+s2djjo6OV0raas8fxNJ5+XOR/Ml3Sxpr0K+QyRFbtjsU97/6uX9r3kOEGu6BdgHQNIGpFvjdi0s3wf4Xaw5xMMepA4zdTiW1NGmMyflW/VOBs4DUOpxuT8wISJ2I91zXbzV7h+A20h3U/S1vq7fiZGGWmjjlQbGHwKPA2MjYldSHQ8v5DmC1EB+eA+3uS68/9XL+1+zWtE7r7++SP8M7fn9m4Efk253ez3wWuBJ0uNQx5B6925I6iW5jHRHxGGkHrkXkHpcLgY+V1j/F3O+u3mlZ+8Y1uzB+aW8jkOBZ4CFed0bN5R1OnBofr8R8Fx+/yCwQxef8X+BCaQepCMHUv02bOt+YHh+fwDpLpQdgL8AgzrJswnwV9LYUAu8/3n/W5/2v7KXzyAKImIpsFLStqRfE7eSbol8O6lj17yIeKmQ/iXgVOCnkQYG+2letDPwftI/wmmSXiPpraQn5u0F7A18StJbuijLL0i/PI7K636+i6IfBNwlaRgwLCLuK0skaTSwVUTcDvyMtMP3mTrrt5tNf5B0S+uuwNxYu3NSh4OBqyPiz8DjkvbsyefsKe9/9fL+1zwHiLX9jrTzdOxAtxamb6m4jisj4sWIeBR4hNSVfz9gZkQ8GxHPkLrrv2Mdy/ofSvd6TwE+Qde9PyGdtv4sv59Ba07z66rfMjfk+tmU7kfzhFQfM/L7gVY/3v8S739NcEe5tXVcp3wz6TRzCfDPwNOkU8sqmulyv5I1A3UzQz+clH/prSbpWUlvjIjFJemPALaU1NGDeBtJYyPi3ia2ua7qqt8yE/M/MQCS5gO7S9ogIl4uJpS0BWmU2d0kBannbUj6l8jn/33E+1+9vP81wWcQa/sd6ZTw8YhYFWmQrs1Ip6G3lqSvOvzDzcDBSsNZDwUOIV2P/RvwBklbSHpt3naz6y76JnCu0thGSNpU0hRJOwFDI2JkRIyJiDE5bV83htVVv93Klz7agDMkCUDSWKUhnw8FLoyI7XL9jCZdL96vN7bdBO9/9fL+1wQHiLXdRbqr4LaGeU8Vfw0U3ADskm/36/SaakT8gdSwdzvpuucPI+LOiFgBfDXP+xW56382HZiW171xxfJ/P5dpjtKwATcBz5F+vc1sSHspfX8aW0v9NuGTwFbAIkl3AT8AltJ5/RzZC9tshve/enn/a4I7ypmZWSmfQZiZWSkHCDMzK+UAYWZmpRwgzMyslAOEmZmVcoAwa0IeafOiwvRgScsk/arJ9dwvafi6pjGrkwOEWXOeJfV27egXsD9pgDWzAccBwqx5VwEfyO+PAH7SsUDS5pIulzRP0m2SxuX5W+TnAtwp6TwKQ19I+kdJt+fOWOcpPdzerOUcIMyaNwM4XNJGwDhSL+QOZwB3RsQ40jMALszzTwN+GxFvAWYB2wJIehNpVNN9Iz1bYRVwFGb9gAfrM2tSRMyTNIZ09jC7YfF+wEdyut/kM4fXkR6c8+E8/0pJT+T07wXeShqaAmBj0gihZi3nAGHWM7OA/wTeDWxRmF82amo0/C0S8OOI+NdeLZ1ZL/AlJrOeuQD4akTc1TD/ZvIlIknvBh6NiKcb5k8iPcUM4HrgUElvyMs2l7Rd/cU3657PIMx6ICLagf8qWXQ68CNJ80ijmH4szz8D+ImkP5BGOH0wr+ceSacA1+bnJK8APgM8UO8nMOueR3M1M7NSvsRkZmalHCDMzKyUA4SZmZVygDAzs1IOEGZmVsoBwszMSjlAmJlZKQcIMzMr9f8B5fe+0zauCLEAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "labels = ['Optimal \\nWithout PCA', 'Optimal \\nWith PCA', 'Batchnorm \\nWithout PCA', 'Batchnorm \\nWith PCA']\n",
    "x = ['Optimal Model Without PCA', 'Optimal Model With PCA']\n",
    "\n",
    "accuracy = [wout_pca, w_pca]\n",
    "\n",
    "x_pos = [i for i, _ in enumerate(x)]\n",
    "\n",
    "plt.bar(x_pos, accuracy, color='green')\n",
    "\n",
    "x_pos = [i+2 for i, _ in enumerate(x)]\n",
    "x = ['Batchnorm Model Without PCA', 'Batchnorm Model With PCA']\n",
    "accuracy = [w_batchnorm, w_batchnorm_pca]\n",
    "plt.bar(x_pos, accuracy, color='red')\n",
    "\n",
    "plt.xlabel(\"Model\")\n",
    "plt.ylabel(\"Average Accuracy in %\")\n",
    "plt.title(\"PCA vs Average Accuracy - K=5 CV\")\n",
    "\n",
    "x_pos = [i for i, _ in enumerate(x)]\n",
    "plt.xticks([0,1,2,3], labels)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Image Size vs Accuracy - 8 Classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch\n",
      "model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch\n",
      "model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch\n",
      "model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch\n",
      "Training: 20x20\n",
      "Training split: 1 of 5\n",
      "model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0c126269fd0f46858bc3bc8075838290",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.06858, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 2.06858 to 2.01051, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 2.01051 to 1.84055, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss improved from 1.84055 to 1.81315, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss did not improve from 1.81315\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss improved from 1.81315 to 1.81113, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss improved from 1.81113 to 1.80366, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 1.80366\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 1.80366\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 9', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00010: val_loss did not improve from 1.80366\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 10', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00011: val_loss improved from 1.80366 to 1.80296, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 11', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00012: val_loss did not improve from 1.80296\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 12', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00013: val_loss did not improve from 1.80296\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 13', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00014: val_loss did not improve from 1.80296\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 14', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00015: val_loss did not improve from 1.80296\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 15', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00016: val_loss did not improve from 1.80296\n",
      "Epoch 00016: early stopping\n",
      "Evaluating split: 1 of 5\n",
      "Recording metrics: 1 of 5\n",
      "\n",
      "\n",
      "Training split: 2 of 5\n",
      "model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e06a1b1e63d6433f8e68291921e0d561",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.07483, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 2.07483 to 1.98625, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 1.98625 to 1.81645, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss improved from 1.81645 to 1.80495, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss improved from 1.80495 to 1.72185, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss improved from 1.72185 to 1.63639, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss improved from 1.63639 to 1.57058, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 1.57058\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 1.57058\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 9', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00010: val_loss improved from 1.57058 to 1.56990, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 10', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00011: val_loss improved from 1.56990 to 1.51326, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 11', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00012: val_loss did not improve from 1.51326\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 12', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00013: val_loss did not improve from 1.51326\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 13', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00014: val_loss did not improve from 1.51326\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 14', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00015: val_loss did not improve from 1.51326\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 15', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00016: val_loss did not improve from 1.51326\n",
      "Epoch 00016: early stopping\n",
      "Evaluating split: 2 of 5\n",
      "Recording metrics: 2 of 5\n",
      "\n",
      "\n",
      "Training split: 3 of 5\n",
      "model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0bf3ce5ad44b4c72a242953a2d9e820c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.07788, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 2.07788 to 2.02952, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 2.02952 to 1.88208, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss did not improve from 1.88208\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss improved from 1.88208 to 1.80456, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss did not improve from 1.80456\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss did not improve from 1.80456\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 1.80456\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 1.80456\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 9', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00010: val_loss improved from 1.80456 to 1.77816, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 10', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00011: val_loss improved from 1.77816 to 1.76997, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 11', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00012: val_loss did not improve from 1.76997\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 12', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00013: val_loss did not improve from 1.76997\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 13', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00014: val_loss did not improve from 1.76997\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 14', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00015: val_loss did not improve from 1.76997\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 15', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00016: val_loss did not improve from 1.76997\n",
      "Epoch 00016: early stopping\n",
      "Evaluating split: 3 of 5\n",
      "Recording metrics: 3 of 5\n",
      "\n",
      "\n",
      "Training split: 4 of 5\n",
      "model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "03db95fd873a4b5389b1e161a0d58701",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.07448, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 2.07448 to 2.04677, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 2.04677 to 1.93643, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss improved from 1.93643 to 1.75514, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss did not improve from 1.75514\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss improved from 1.75514 to 1.70340, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss did not improve from 1.70340\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 1.70340\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss improved from 1.70340 to 1.68872, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 9', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00010: val_loss did not improve from 1.68872\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 10', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00011: val_loss did not improve from 1.68872\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 11', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00012: val_loss did not improve from 1.68872\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 12', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00013: val_loss did not improve from 1.68872\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 13', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00014: val_loss did not improve from 1.68872\n",
      "Epoch 00014: early stopping\n",
      "Evaluating split: 4 of 5\n",
      "Recording metrics: 4 of 5\n",
      "\n",
      "\n",
      "Training split: 5 of 5\n",
      "model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "cc42e277fcf64d97a558dab6d7842fc2",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.06967, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 2.06967 to 1.99277, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 1.99277 to 1.78889, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss improved from 1.78889 to 1.71996, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss did not improve from 1.71996\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss did not improve from 1.71996\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss did not improve from 1.71996\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss improved from 1.71996 to 1.66790, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 1.66790\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 9', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00010: val_loss improved from 1.66790 to 1.64660, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 10', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00011: val_loss improved from 1.64660 to 1.59955, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 11', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00012: val_loss did not improve from 1.59955\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 12', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00013: val_loss did not improve from 1.59955\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 13', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00014: val_loss did not improve from 1.59955\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 14', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00015: val_loss did not improve from 1.59955\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 15', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00016: val_loss improved from 1.59955 to 1.59561, saving model to models/model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 16', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00017: val_loss did not improve from 1.59561\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 17', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00018: val_loss did not improve from 1.59561\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 18', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00019: val_loss did not improve from 1.59561\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 19', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00020: val_loss did not improve from 1.59561\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 20', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00021: val_loss did not improve from 1.59561\n",
      "Epoch 00021: early stopping\n",
      "Evaluating split: 5 of 5\n",
      "Recording metrics: 5 of 5\n",
      "\n",
      "\n",
      "Average Mean: 0.32538190364277325\n",
      "Average Recall: 0.32538190364277325\n",
      "Average Precision: 0.34817012349915444\n",
      "Training: 40x40\n",
      "Training split: 1 of 5\n",
      "model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ce6e00b420ba4743aca42a59d5f0d86e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.05328, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 2.05328 to 1.95806, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 1.95806 to 1.86868, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss improved from 1.86868 to 1.83912, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss did not improve from 1.83912\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss improved from 1.83912 to 1.77058, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss improved from 1.77058 to 1.77004, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss improved from 1.77004 to 1.73817, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 1.73817\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 9', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00010: val_loss did not improve from 1.73817\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 10', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00011: val_loss did not improve from 1.73817\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 11', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00012: val_loss did not improve from 1.73817\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 12', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00013: val_loss did not improve from 1.73817\n",
      "Epoch 00013: early stopping\n",
      "Evaluating split: 1 of 5\n",
      "Recording metrics: 1 of 5\n",
      "\n",
      "\n",
      "Training split: 2 of 5\n",
      "model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d82682f9ff7c4fbeb5567e9c3c5257cc",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.05566, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 2.05566 to 1.91404, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 1.91404 to 1.75906, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss improved from 1.75906 to 1.68427, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss improved from 1.68427 to 1.54902, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss did not improve from 1.54902\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss improved from 1.54902 to 1.49100, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 1.49100\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 1.49100\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 9', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00010: val_loss did not improve from 1.49100\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 10', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00011: val_loss did not improve from 1.49100\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 11', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00012: val_loss did not improve from 1.49100\n",
      "Epoch 00012: early stopping\n",
      "Evaluating split: 2 of 5\n",
      "Recording metrics: 2 of 5\n",
      "\n",
      "\n",
      "Training split: 3 of 5\n",
      "model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5f1e35735dc84a8ea3aa6c09de2abb7b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.06279, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 2.06279 to 1.85551, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 1.85551 to 1.84445, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss improved from 1.84445 to 1.83710, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss improved from 1.83710 to 1.75791, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss improved from 1.75791 to 1.72038, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss did not improve from 1.72038\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 1.72038\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss improved from 1.72038 to 1.71914, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 9', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00010: val_loss did not improve from 1.71914\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 10', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00011: val_loss improved from 1.71914 to 1.56311, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 11', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00012: val_loss did not improve from 1.56311\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 12', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00013: val_loss did not improve from 1.56311\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 13', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00014: val_loss did not improve from 1.56311\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 14', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00015: val_loss did not improve from 1.56311\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 15', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00016: val_loss did not improve from 1.56311\n",
      "Epoch 00016: early stopping\n",
      "Evaluating split: 3 of 5\n",
      "Recording metrics: 3 of 5\n",
      "\n",
      "\n",
      "Training split: 4 of 5\n",
      "model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c04c3b9ea1304db9aeb2b887252fe951",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.03791, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 2.03791 to 2.02914, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 2.02914 to 1.76207, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss improved from 1.76207 to 1.74015, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss improved from 1.74015 to 1.69523, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss improved from 1.69523 to 1.67801, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss improved from 1.67801 to 1.63986, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 1.63986\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 1.63986\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 9', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00010: val_loss did not improve from 1.63986\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 10', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00011: val_loss did not improve from 1.63986\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 11', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00012: val_loss improved from 1.63986 to 1.58946, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 12', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00013: val_loss did not improve from 1.58946\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 13', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00014: val_loss did not improve from 1.58946\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 14', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00015: val_loss did not improve from 1.58946\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 15', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00016: val_loss did not improve from 1.58946\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 16', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00017: val_loss did not improve from 1.58946\n",
      "Epoch 00017: early stopping\n",
      "Evaluating split: 4 of 5\n",
      "Recording metrics: 4 of 5\n",
      "\n",
      "\n",
      "Training split: 5 of 5\n",
      "model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ee01052195ce493fadd7bf482a5e5c4c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.00716, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 2.00716 to 1.81591, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 1.81591 to 1.74008, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss improved from 1.74008 to 1.71063, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss improved from 1.71063 to 1.64298, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss improved from 1.64298 to 1.61626, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss improved from 1.61626 to 1.61404, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 1.61404\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss improved from 1.61404 to 1.60330, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 9', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00010: val_loss did not improve from 1.60330\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 10', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00011: val_loss improved from 1.60330 to 1.52006, saving model to models/model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 11', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00012: val_loss did not improve from 1.52006\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 12', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00013: val_loss did not improve from 1.52006\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 13', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00014: val_loss did not improve from 1.52006\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 14', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00015: val_loss did not improve from 1.52006\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 15', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00016: val_loss did not improve from 1.52006\n",
      "Epoch 00016: early stopping\n",
      "Evaluating split: 5 of 5\n",
      "Recording metrics: 5 of 5\n",
      "\n",
      "\n",
      "Average Mean: 0.3828613396004701\n",
      "Average Recall: 0.3828613396004701\n",
      "Average Precision: 0.39754882758774185\n",
      "Training: 60x60\n",
      "Training split: 1 of 5\n",
      "model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "bb4f369478984140a988953b0ac715da",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.04610, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 2.04610 to 1.91345, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 1.91345 to 1.81496, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss improved from 1.81496 to 1.74128, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss did not improve from 1.74128\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss did not improve from 1.74128\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss did not improve from 1.74128\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 1.74128\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 1.74128\n",
      "Epoch 00009: early stopping\n",
      "Evaluating split: 1 of 5\n",
      "Recording metrics: 1 of 5\n",
      "\n",
      "\n",
      "Training split: 2 of 5\n",
      "model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2961a68a8b1b48e5a0cc3c9690109b7a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.00523, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 2.00523 to 1.87006, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 1.87006 to 1.58101, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss did not improve from 1.58101\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss improved from 1.58101 to 1.55165, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss improved from 1.55165 to 1.49740, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss improved from 1.49740 to 1.48140, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss improved from 1.48140 to 1.46502, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 1.46502\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 9', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00010: val_loss did not improve from 1.46502\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 10', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00011: val_loss did not improve from 1.46502\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 11', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00012: val_loss did not improve from 1.46502\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 12', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00013: val_loss did not improve from 1.46502\n",
      "Epoch 00013: early stopping\n",
      "Evaluating split: 2 of 5\n",
      "Recording metrics: 2 of 5\n",
      "\n",
      "\n",
      "Training split: 3 of 5\n",
      "model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "216be3afd3164cb4b051bc01eb48b83f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.04696, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 2.04696 to 1.88020, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 1.88020 to 1.73513, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss did not improve from 1.73513\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss did not improve from 1.73513\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss did not improve from 1.73513\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss improved from 1.73513 to 1.70279, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 1.70279\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 1.70279\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 9', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00010: val_loss did not improve from 1.70279\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 10', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00011: val_loss did not improve from 1.70279\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 11', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00012: val_loss did not improve from 1.70279\n",
      "Epoch 00012: early stopping\n",
      "Evaluating split: 3 of 5\n",
      "Recording metrics: 3 of 5\n",
      "\n",
      "\n",
      "Training split: 4 of 5\n",
      "model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "fa4b8ac8954b4eeabe57f8b285324b2d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.03780, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 2.03780 to 1.85847, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 1.85847 to 1.71212, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss improved from 1.71212 to 1.68228, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss improved from 1.68228 to 1.67090, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss improved from 1.67090 to 1.61888, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss improved from 1.61888 to 1.59245, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 1.59245\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 1.59245\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 9', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00010: val_loss did not improve from 1.59245\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 10', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00011: val_loss did not improve from 1.59245\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 11', max=738, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00012: val_loss did not improve from 1.59245\n",
      "Epoch 00012: early stopping\n",
      "Evaluating split: 4 of 5\n",
      "Recording metrics: 4 of 5\n",
      "\n",
      "\n",
      "Training split: 5 of 5\n",
      "model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2640491afbfa4730a9ea7d7fa3cb8eb6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.00591, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 2.00591 to 1.75461, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 1.75461 to 1.67403, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss improved from 1.67403 to 1.63653, saving model to models/model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss did not improve from 1.63653\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss did not improve from 1.63653\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss did not improve from 1.63653\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 1.63653\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 1.63653\n",
      "Epoch 00009: early stopping\n",
      "Evaluating split: 5 of 5\n",
      "Recording metrics: 5 of 5\n",
      "\n",
      "\n",
      "Average Mean: 0.4034958871915393\n",
      "Average Recall: 0.4034958871915393\n",
      "Average Precision: 0.4227740301766548\n",
      "Training: 80x80\n",
      "Training split: 1 of 5\n",
      "model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "33df3fde6f124ae386e6f934db2d4c35",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.00182, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 2.00182 to 1.97993, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 1.97993 to 1.88763, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss did not improve from 1.88763\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss improved from 1.88763 to 1.80419, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss did not improve from 1.80419\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss did not improve from 1.80419\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 1.80419\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 1.80419\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 9', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00010: val_loss did not improve from 1.80419\n",
      "Epoch 00010: early stopping\n",
      "Evaluating split: 1 of 5\n",
      "Recording metrics: 1 of 5\n",
      "\n",
      "\n",
      "Training split: 2 of 5\n",
      "model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d65c7de54fd543349b46ef051fe10b1b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.01772, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 2.01772 to 1.80240, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 1.80240 to 1.57176, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss improved from 1.57176 to 1.56994, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss improved from 1.56994 to 1.48906, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss did not improve from 1.48906\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss improved from 1.48906 to 1.48898, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 1.48898\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 1.48898\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 9', max=737, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00010: val_loss did not improve from 1.48898\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 10', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00011: val_loss did not improve from 1.48898\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 11', max=737, style=ProgressStyle(description_width='in…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00012: val_loss did not improve from 1.48898\n",
      "Epoch 00012: early stopping\n",
      "Evaluating split: 2 of 5\n",
      "Recording metrics: 2 of 5\n",
      "\n",
      "\n",
      "Training split: 3 of 5\n",
      "model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9da2c7bba3a24f19974e96e0c49c5b08",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 2.09284, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 2.09284 to 1.80632, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 1.80632 to 1.79692, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss improved from 1.79692 to 1.76525, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss did not improve from 1.76525\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss did not improve from 1.76525\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss did not improve from 1.76525\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 1.76525\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 1.76525\n",
      "Epoch 00009: early stopping\n",
      "Evaluating split: 3 of 5\n",
      "Recording metrics: 3 of 5\n",
      "\n",
      "\n",
      "Training split: 4 of 5\n",
      "model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "80d22e0a67894a65ae9c4a6fc4cc9822",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 1.98504, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 1.98504 to 1.82744, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 1.82744 to 1.60189, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss improved from 1.60189 to 1.59357, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss did not improve from 1.59357\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss did not improve from 1.59357\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss did not improve from 1.59357\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 1.59357\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 1.59357\n",
      "Epoch 00009: early stopping\n",
      "Evaluating split: 4 of 5\n",
      "Recording metrics: 4 of 5\n",
      "\n",
      "\n",
      "Training split: 5 of 5\n",
      "model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b7e1bbec7b524d96b2504f18587398b2",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=50, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 0', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00001: val_loss improved from inf to 1.99249, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 1', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00002: val_loss improved from 1.99249 to 1.72560, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 2', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00003: val_loss improved from 1.72560 to 1.61000, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 3', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00004: val_loss did not improve from 1.61000\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 4', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00005: val_loss improved from 1.61000 to 1.51642, saving model to models/model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch.hdf5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 5', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00006: val_loss did not improve from 1.51642\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 6', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00007: val_loss did not improve from 1.51642\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 7', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00008: val_loss did not improve from 1.51642\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 8', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00009: val_loss did not improve from 1.51642\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Epoch 9', max=738, style=ProgressStyle(description_width='ini…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00010: val_loss did not improve from 1.51642\n",
      "Epoch 00010: early stopping\n",
      "Evaluating split: 5 of 5\n",
      "Recording metrics: 5 of 5\n",
      "\n",
      "\n",
      "Average Mean: 0.4067097532314924\n",
      "Average Recall: 0.4067097532314924\n",
      "Average Precision: 0.4287376645856321\n"
     ]
    }
   ],
   "source": [
    "# Data params\n",
    "image_resizes = [20, 40, 60, 80]\n",
    "Nber_of_breeds = 8\n",
    "\n",
    "# Model params\n",
    "dropout = 0.3\n",
    "fully_connected_nodes = 500\n",
    "learning_rate=0.0005\n",
    "\n",
    "# Training params\n",
    "batch_size = 20\n",
    "\n",
    "models = []\n",
    "for size in image_resizes:\n",
    "    model_name = \"model_\" + str(size) + \"x\" + str(size) + \"_\" + str(Nber_of_breeds) + \"breeds_\" + str(dropout) + \"dropout_\" + str(learning_rate) + \"lr_\" + str(fully_connected_nodes) + \"fc_\" + str(batch_size) + \"batch\"\n",
    "\n",
    "    models.append(model_name)\n",
    "    print(model_name)\n",
    "\n",
    "accuracies = []\n",
    "recalls = []\n",
    "precisions = []\n",
    "for size, model_name in zip(image_resizes, models):\n",
    "    print(f\"Training: {size}x{size}\")\n",
    "    # Load data\n",
    "    X = pickle.load(open(\"pickles/X_\" + str(size) + \"x\" + str(size) + \"_\" + str(Nber_of_breeds) + \".pkl\", \"rb\"))\n",
    "    y = pickle.load(open(\"pickles/y_\" + str(size) + \"x\" + str(size) + \"_\" + str(Nber_of_breeds) + \".pkl\", \"rb\"))\n",
    "    labels_name = pickle.load(open(\"pickles/labels_name_\" + str(size) + \"x\" + str(size) + \"_\" + str(Nber_of_breeds) + \".pkl\", \"rb\"))\n",
    "    \n",
    "    n_splits = 5\n",
    "\n",
    "\n",
    "    sample_accuracies = []\n",
    "    sample_recalls = []\n",
    "    sample_precisions = []\n",
    "\n",
    "    # K-Fold cross validation\n",
    "    i = 0\n",
    "    kf = KFold(n_splits=n_splits)\n",
    "    for train, test in kf.split(X, y):\n",
    "        i += 1\n",
    "        \n",
    "        # Load data\n",
    "        X = pickle.load(open(\"pickles/X_\" + str(size) + \"x\" + str(size) + \"_\" + str(Nber_of_breeds) + \".pkl\", \"rb\"))\n",
    "        y = pickle.load(open(\"pickles/y_\" + str(size) + \"x\" + str(size) + \"_\" + str(Nber_of_breeds) + \".pkl\", \"rb\"))\n",
    "    \n",
    "        # Train model\n",
    "        print(\"Training split:\", i, \"of\", n_splits)\n",
    "        model, history = train_model(X[train], y[train], X[test], y[test], size, Nber_of_breeds, dropout, fully_connected_nodes, learning_rate, epochs, batch_size)\n",
    "\n",
    "        # Evaluate model\n",
    "        print(\"Evaluating split:\", i, \"of\", n_splits)\n",
    "        accuracy, recall, precision = evaluate_model(model, X[train], y[train], X[test], y[test], labels_name)\n",
    "\n",
    "        # Record metrics\n",
    "        print(\"Recording metrics:\", i, \"of\", n_splits)\n",
    "        sample_accuracies.append(accuracy)\n",
    "        sample_recalls.append(recall)\n",
    "        sample_precisions.append(precision)\n",
    "\n",
    "        print(\"\\n\")\n",
    "    \n",
    "    # Print average metrics\n",
    "    avg_acc = np.asarray(sample_accuracies).mean()\n",
    "    print(\"Average Mean:\", avg_acc)\n",
    "    avg_recall = np.asarray(sample_recalls).mean()\n",
    "    print(\"Average Recall:\", avg_recall)\n",
    "    avg_prec = np.asarray(sample_precisions).mean()\n",
    "    print(\"Average Precision:\", avg_prec)\n",
    "    \n",
    "    accuracies.append(avg_acc)\n",
    "    recalls.append(avg_recall)\n",
    "    precisions.append(avg_prec)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "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.scatter(image_resizes, accuracies);\n",
    "plt.xlabel(\"Image Size in Pixels\");\n",
    "plt.ylabel(\"Accuracy in %\");\n",
    "plt.title(f\"Image Size vs Accuracy - {Nber_of_breeds} Classes - K=5 Cross Validation\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "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.scatter(image_resizes, recalls);\n",
    "plt.xlabel(\"Image Size in Pixels\");\n",
    "plt.ylabel(\"Recall\");\n",
    "plt.title(f\"Image Size vs Recall - {Nber_of_breeds} Classes - K=5 Cross Validation\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "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.scatter(image_resizes, precisions);\n",
    "plt.xlabel(\"Image Size in Pixels\");\n",
    "plt.ylabel(\"Precision\");\n",
    "plt.title(f\"Image Size vs Precision - {Nber_of_breeds} Classes - K=5 Cross Validation\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "jupyter": {
     "source_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading model: model_20x20_8breeds_0.3dropout_0.0005lr_500fc_20batch\n",
      "Loading X_test pickle...\n",
      "Loading y_test pickle...\n",
      "Getting predictions...\n",
      "Calculating accuracy...\n",
      "Loading model: model_40x40_8breeds_0.3dropout_0.0005lr_500fc_20batch\n",
      "Loading X_test pickle...\n",
      "Loading y_test pickle...\n",
      "Getting predictions...\n",
      "Calculating accuracy...\n",
      "Loading model: model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch\n",
      "Loading X_test pickle...\n",
      "Loading y_test pickle...\n",
      "Getting predictions...\n",
      "Calculating accuracy...\n",
      "Loading model: model_80x80_8breeds_0.3dropout_0.0005lr_500fc_20batch\n",
      "Loading X_test pickle...\n",
      "Loading y_test pickle...\n",
      "Getting predictions...\n",
      "Calculating accuracy...\n"
     ]
    },
    {
     "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": [
    "# # Data params\n",
    "# image_resizes = [20, 40, 60, 80]\n",
    "# Nber_of_breeds = 8\n",
    "\n",
    "# # Model params\n",
    "# dropout = 0.3\n",
    "# fully_connected_nodes = 500\n",
    "# learning_rate=0.0005\n",
    "\n",
    "# # Training params\n",
    "# batch_size = 20\n",
    "\n",
    "# models = []\n",
    "# for size in image_resizes:\n",
    "#     model_name = \"model_\" + str(size) + \"x\" + str(size) + \"_\" + str(Nber_of_breeds) + \"breeds_\" + str(dropout) + \"dropout_\" + str(learning_rate) + \"lr_\" + str(fully_connected_nodes) + \"fc_\" + str(batch_size) + \"batch\"\n",
    "\n",
    "#     models.append(model_name)\n",
    "\n",
    "# accuracies = []\n",
    "# for size, model_name in zip(image_resizes, models):\n",
    "#     print(\"Loading model:\", model_name)\n",
    "#     model = load_model('models/' + model_name + '.hdf5')\n",
    "    \n",
    "#     print(\"Loading X_test pickle...\")\n",
    "#     X_test = pickle.load(open(\"pickles/X_test_\" + str(size) + \"x\" + str(size) + \"_\" + str(Nber_of_breeds) + \".pkl\", \"rb\"))\n",
    "    \n",
    "#     print(\"Loading y_test pickle...\")\n",
    "#     y_test = pickle.load(open(\"pickles/y_test_\" + str(size) + \"x\" + str(size) + \"_\" + str(Nber_of_breeds) + \".pkl\", \"rb\"))\n",
    "    \n",
    "#     print(\"Getting predictions...\")\n",
    "#     predictions = model.predict(X_test)\n",
    "\n",
    "    \n",
    "#     pred_argmax = []\n",
    "#     for pred in predictions:\n",
    "#         pred_argmax.append(pred.argmax())\n",
    "\n",
    "#     true_argmax = []\n",
    "#     for pred in y_test:\n",
    "#         true_argmax.append(pred.argmax())\n",
    "\n",
    "#     print(\"Calculating accuracy...\")\n",
    "#     accuracies.append(accuracy_score(pred_argmax, true_argmax))\n",
    "    \n",
    "# plt.scatter(image_resizes, accuracies);\n",
    "# plt.xlabel(\"Image Size in Pixels\");\n",
    "# plt.ylabel(\"Accuracy in %\");\n",
    "# plt.title(f\"Image Size vs Accuracy - {Nber_of_breeds} Classes\");"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Number of Breeds vs Accuracy - 40x40 Images"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Data params\n",
    "size = 40\n",
    "num_breeds = [4, 8, 16, 32, 64]\n",
    "\n",
    "# Model params\n",
    "dropout = 0.3\n",
    "fully_connected_nodes = 500\n",
    "learning_rate=0.0005\n",
    "\n",
    "# Training params\n",
    "batch_size = 20\n",
    "\n",
    "models = []\n",
    "for breed in num_breeds:\n",
    "    model_name = \"model_\" + str(size) + \"x\" + str(size) + \"_\" + str(breed) + \"breeds_\" + str(dropout) + \"dropout_\" + str(learning_rate) + \"lr_\" + str(fully_connected_nodes) + \"fc_\" + str(batch_size) + \"batch\"\n",
    "\n",
    "    models.append(model_name)\n",
    "print(models)\n",
    "accuracies = []\n",
    "recalls = []\n",
    "precisions = []\n",
    "for breed, model_name in zip(num_breeds, models):\n",
    "    print(f\"Training: {size}x{size}\")\n",
    "    # Load data\n",
    "    X = pickle.load(open(\"pickles/X_\" + str(size) + \"x\" + str(size) + \"_\" + str(breed) + \".pkl\", \"rb\"))\n",
    "    y = pickle.load(open(\"pickles/y_\" + str(size) + \"x\" + str(size) + \"_\" + str(breed) + \".pkl\", \"rb\"))\n",
    "    labels_name = pickle.load(open(\"pickles/labels_name_\" + str(size) + \"x\" + str(size) + \"_\" + str(breed) + \".pkl\", \"rb\"))\n",
    "    \n",
    "    n_splits = 5\n",
    "\n",
    "\n",
    "    sample_accuracies = []\n",
    "    sample_recalls = []\n",
    "    sample_precisions = []\n",
    "\n",
    "    # K-Fold cross validation\n",
    "    i = 0\n",
    "    kf = KFold(n_splits=n_splits)\n",
    "    for train, test in kf.split(X, y):\n",
    "        i += 1\n",
    "        \n",
    "        # Load data\n",
    "        X = pickle.load(open(\"pickles/X_\" + str(size) + \"x\" + str(size) + \"_\" + str(breed) + \".pkl\", \"rb\"))\n",
    "        y = pickle.load(open(\"pickles/y_\" + str(size) + \"x\" + str(size) + \"_\" + str(breed) + \".pkl\", \"rb\"))\n",
    "    \n",
    "        # Train model\n",
    "        print(\"Training split:\", i, \"of\", n_splits)\n",
    "        model, history = train_model(X[train], y[train], X[test], y[test], size, breed, dropout, fully_connected_nodes, learning_rate, epochs, batch_size)\n",
    "\n",
    "        # Evaluate model\n",
    "        print(\"Evaluating split:\", i, \"of\", n_splits)\n",
    "        accuracy, recall, precision = evaluate_model(model, X[train], y[train], X[test], y[test], labels_name)\n",
    "\n",
    "        # Record metrics\n",
    "        print(\"Recording metrics:\", i, \"of\", n_splits)\n",
    "        sample_accuracies.append(accuracy)\n",
    "        sample_recalls.append(recall)\n",
    "        sample_precisions.append(precision)\n",
    "\n",
    "        print(\"\\n\")\n",
    "    \n",
    "    # Print average metrics\n",
    "    avg_acc = np.asarray(sample_accuracies).mean()\n",
    "    print(\"Average Accuracy:\", avg_acc)\n",
    "    avg_recall = np.asarray(sample_recalls).mean()\n",
    "    print(\"Average Recall:\", avg_recall)\n",
    "    avg_prec = np.asarray(sample_precisions).mean()\n",
    "    print(\"Average Precision:\", avg_prec)\n",
    "    \n",
    "    accuracies.append(avg_acc)\n",
    "    recalls.append(avg_recall)\n",
    "    precisions.append(avg_prec)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "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.scatter(num_breeds, accuracies);\n",
    "plt.xlabel(\"Number of Breeds\");\n",
    "plt.ylabel(\"Accuracy in %\");\n",
    "plt.title(f\"Num Breeds vs Accuracy - {size}x{size} Images - K=5 Cross Validation\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "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.scatter(num_breeds, precisions);\n",
    "plt.xlabel(\"Number of Breeds\");\n",
    "plt.ylabel(\"Precision\");\n",
    "plt.title(f\"Num Breeds vs Precision - {size}x{size} Images - K=5 Cross Validation\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZMAAAEWCAYAAACjYXoKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nO3de5wcVZn/8c+XSZAAwQiJSC4QQIiiBJCRi7IaUAiuKEHlJ4oiushGF5FdjRJXuXhZdbPuegFFZBEQECPGkEU0sEjAFdRMDBBuwRAuuYCEQLg5QhKe3x/nNFTanpme1FRmevi+X695Tdep06eeU11VT926WhGBmZlZGZv1dwBmZtb6nEzMzKw0JxMzMyvNycTMzEpzMjEzs9KcTMzMrDQnk01M0hmSLu7vODYlSfdJemt+/aLrv1lXJIWkV+bX50j6QjN1N2I6x0q6emPjbMYmTyZ5w/JnSVsVyk6QNG8TTX+epL9KekrS45JukLTnpph2f8sb8rW572sk3SjpwP6Oa2NIenNeub5cV/5+SfdLelrSbEnb9rLdzSXdJWl5Xfl4SddJ+kse/9Zu2hjUCVPSpOL8yfNslqTfStqmF+0cL2l9Xh5rf5O6qS9JJ0u6LX++yyX9tL/WX0nfl3RRg/KJkp7p7bIXEVMj4kt9ENf4vG4MKbR9SUQcVrbt7vTXkckQ4JP9NG2AkyJia2A7YB7wo64qFj+QQeInue8jgeuAn/ZzPL0maSjwLeD3deWvAb4PfBDYHvgL8N1eNj8NeLhB+Y+BhaRl5l+ByyWN6mXbg46klwCzgBHAYRHxRC+buCkiti78zeum7rdI242TgW2B3YHZwNu7iK2tl7H01gXAu4o7xtlxwJUR8WjF0x9Q+iuZzAA+LWlE/YhGWTUfTZyQXx+f94D+K+9dL5X0hly+TNLDkj7UTBARsQ64DNijMK0zJF0u6WJJTwDHS9pM0qmS7pG0WtLM4l6HpAPyXv4aSbcU964k7SzpeklPSrqGtBGvjdsiT2d1fu98Sds3mCenSrq8ruxbkr5dmCdL8zTulXRsk32/BBhT3ChKOkLSzYUjl4mFcePyHuiqHPNZuXxXSb/OZY9IuqTRZ9uHPgVcDdxVV34s8D8RcUNEPAV8gbSyD5e0bd6TfUeOeWtJSyQdV+jfzsAHgK8WG5W0O/A64PSI6IyInwGLgHc3E2xenj8u6U/5M/pSnmc3SXoiL0+b57ovk3RlnseP5ddjizEqHU0/Kel/JZ2twlFQD8tir5eTHvq1JfA/wFDg7RHxdJn2epjWbsA/Ae+LiF9HxDMR8Ze8x/21XOcCSd+TdJWkp4GDJb1U0kV5ft4v6fOSNsv1X5nXzcfzcvuTXK68fXk4j7tV0mvrY4qIm4AVFJYDpQT2fuDCPLxf/pzXSHpQ0lm1z7pBHy9Q4Uhb0rT8npWSPlJX9+2SFublZ5mkMwqjb8j/1ygd7R2YP/v/K7z/DXl783j+/4bCuHl5Gf1tXlauljSSnkTEJv0D7gPeStqb+XIuOwGYl1+PBwIYUnjPPOCE/Pp4YB3wYaAN+DLwAHA28BLgMOBJYOsupl9sa3PgK8ANhfFnAGuBKaRkOww4BfgdMDZP4/vAj3P9McBq4O9z/UPz8Kg8/ibgP/P73pRjuziP+0fSyrhl7su+wDYNYt6JtJe9TR5uAx4EDgC2Ap4AJuRxOwCv6aLvZxSmvTnwNeCR2rwmbTAfBvbP0/hQ/rxekodvAf4rT3ML4KD8vlfmfr8EGEVamL9Z/5nXx7CRy89OwN3A1qQ9wy8Xxl0BfLau/lPAvvn1YcBDwMuBHwCX19W9EjgKmAQsL5QfBdxZV/cs4Ds9zec8HMAcYBvgNcAzwLXALsBLgTuAD+W625E2TlsCw0lHjrMLbd0E/Ef+/A7Kn33tM+1yWezNctLEZzAJWAVcn/v1krrx7wfWdPO3Y2Fdfjovg3eTkv+QLqY5Fbi/h7guAB4H3pj7vwVwUV4uhpO2LXcD/5Dr/5h0lFmrW1ueJwMLSEdbAl4N7NDFNP8V+N/C8OQ8b4bm4X1J6+mQPP07gVPqlo1XFuKvbRMPB/4MvDZ/dpfW1Z0E7Jljn5jrTulmG3o88H/59bbAY6Qj+CHA+/LwdoVt5D2kI79hefhrPS4XG7tSl9gY3EdKJq/NH/woep9M/lQYt2euv32hbDWwdxfTn0faMK8Bns0xvKVuQ3BD3XvurKuzAynhDAE+C/yorv5c0oZ4R1Li26ow7lJeWPk/AtwITGxivv0fcFx+fShwT369Ve7Lu4FhPbRxRu7zGmB9nk+TCuO/B3yp7j2LgTcDB5JWkoYre917pgAL6z/zQgxlkskVwHvrV748fC0wta7+iro+fod0VLGSvPLk8qOAXxVW1GIy+SDwu7p2vwJc0M18rk8mbywML6CQ9IBvUEi+dW3tDTyWX9eWpy0L4y8uLE/dLYtNLydNfAaTgL/mZendJdrZBdiZtEHck5RUp3dR91/rP4MGdS4ALioMt5ES9x6Fsn/khW3NRcC5wNi6dg4hJZ0DgM16mOaOpG3B2Dx8CfCtbuqfAvy8btlolEzOp7ABJ23Yn6/boN1vAv+VX4+n+2TyQeAPde+/CTg+v54HfL4w7uO1daO7v367mysibiPtCZ66EW//c+F1Z26vvmzrbt5/ckSMIO2NHEE6/z2xMH5ZXf2dgJ/nQ9U1pOSynnRefifg6Nq4PP4gUsIZTdoQFA//7y+8/hFpZb8sH8r+u9L1gEYuJe1BQNrzuzT3+2ngvaQ9twcl/ULSq7rp+8zc9+2B20h7TsV+fqquL+NyP8aR9gzX1Tco6eWSLpO0QunU4MUUTuc1S9Iv9cKF2L85BZNPUQ2PiJ900cRTpL3/om1IR4M155J2ZH4YEatzu1sB/w58okS7PalfPhsur5K2VLqwe3+elzcAI/Lpk9HAoxHxl8J7i8tql8tib5YTbXhBfMcu+vMIcAxwoaTJTc6DDUTE0oi4NyKei4hFwBeB93RRfTVpnepJcX6MJB3BFde5+0lHcACfIR15/EHS7bVTSRHxa9KR59nAnyWdqy5uLIiIB0if0QckbU3akbqwNl7S7vlU5UP58/w3mls3Rtf1pdgHJO2vdEPIKkmPkz7XZte50fXtseF8gXQEX/MXut+eAv1/a/DpwEfZsBO1De+WhbJXVDHxvBD/BlhCOgXy/Ki6qsuAt0XEiMLfFhGxIo/7Ud24rSKdx30QeJk2vED3/MoZEWsj4syI2AN4AymxHUdjPwUmKZ0/P4qcTHI7cyPiUNLKdhfpFE5PfX+EtJd2hqTaSroM+EpdX7aMiB/ncTuq8Q0JX83zbGJEbEO67qCeYmgQ09vihQuxlzSo8hagPa+YD5E2jqdIuiKPvx3Yq1ZZ0i6kU2935+E20inKi4CP6YXbLHcj7c39Jrc7C9ghT2d8bncXScMLseyVy/vap4AJwP55Xr6p1h3S8rSt0rWKmnGF190ti00vJ7HhBfEHugo0ImaR1t/LJR1cK1e6DfWpbv66SlBB18vNtcBYSe1dxVNoo+YR0lHDToWyHUlHq0TEQxHx0YgYTVoXvltbJiLi2xGxL+m05O6kGzO6ciFpvX03cG9E/LEw7nukeb1b/jw/100fix5kw8+2fp5dSjrFOC4iXgqcU2i3fvtVbyUbzpNa+yuaiKtL/ZpMImIJ8BPS3Rm1slWkTn1AUlveW9i1qhiUbo3dg+43DOcAX5G0U37PKElH5nEXA++QNDnHu4XSrZNjI+J+oAM4U+n2yYOAdxSmfbCkPfNG7gnSgr++UQB5vswDfkhaYO/MbWwv6Z05YT1D2otu2EaDNu8iHRl9Jhf9AJia93okaat8oW848AfSAv61XL6FpDfm9w3P010jaQzdr3hlfIG0Yu+d/+bkmD+cx19C+iz+Ls+PLwKzIqJ2BPG5/P8jpOsOF+V5fxtpxa21ewLpyGFvYFlE3A3cDJye+30U6Tz1zyro43DSkcoapZs8Tq+NKCxPZ+Tl6UAKyxPdLItllpPu5B2Nk4ArastDpIviW3fz9wCApLcp33CSj5K+QDqN2Wg6fyLdmffj3KfNc/+OkdTw7EZErAdmktbd4Xn9/Zc8n5B0tF64ueEx0kZ4vaTX53VgKGnn9q89zKufkZafMykclWTDSev2U7mPH+umnaKZpJt/9sg7D6fXjR9OOkr9q6T9SGcralYBz5FOIzZyFbC70m30QyS9l7QNvLLJ2Brq7yMTSCt8/a11HyVtkFaT9gxu7ONpnlXbSyKdavp8RPyym/rfIm24rpb0JOli/P4AEbEMOJK0oVpF2jucxgvz9v257qOkBaJ4X/orgMtJC9udpAua3X0/4VLS9aZLC2WbkfZmV+ZpvJl0jrNZM4ATJb08IjpI8/4s0sq1hHSutbZivoN0sf0BYDnpyADSSvQ60vWnX5D27PtcRDyZ9yYfioiHSBvdpyPfghkRt5MO9y8h3UgwnDwvJO1L2pAcl/vyddLG49SIWFfX7qPAc3m4thE5BmjP8+VrwHtygu9r3yRd9HyEtJz9qm78saTrV6tJN5/8hJQceloWyy4nXYqIC3Pbv8gbtma9BbhV6c6rq0jLzb91U/9kXjj9tIZ0kfgo0k0sXfkEKSEsJV13vJR0PQLg9cDv83ZgDvDJiLiXdArzB6TP+n7SvP6PriaQTyHWEkr9EfWnSduAJ3ObXZ2irW/zl6Rl4dek9fDXdVU+Dnwxb49OIyWf2nv/Qrqm91ul050H1LW9mnQW5FO5b58BjshnKzaa8gUWM2tBSrez3hUR9XuuZpvUQDgyMbMm5VMwuyp99+lw0pHI7P6Oy2ywfbvbbLB7Bel00HakU40fi4iF/RuSmU9zmZlZH/BpLjMzK63lTnONHDkyxo8f399hmJm1lAULFjwSEZU9nLTlksn48ePp6Ojo7zDMzFqKpPpvvfcpn+YyM7PSnEzMzKw0JxMzMyvNycTMzEpzMjEzs9KcTMzMrDQnEzMzK83JxMzMSnMyMTOz0pxMzMysNCcTMzMrreWezbUxZi9cwYy5i1m5ppPRI4YxbfIEpuwzpr/DMjMbNAZ9Mpm9cAXTZy2ic236Ke8VazqZPmsRgBOKmVkfGfSnuWbMXfx8IqnpXLueGXMX91NEZmaDz6BPJivXdPaq3MzMem/QJ5PRI4b1qtzMzHpv0CeTaZMnMGxo2wZlw4a2MW3yhH6KyMxs8Bn0F+BrF9l9N5eZWXUGfTKBlFCcPMzMqjPoT3OZmVn1Kk0mkg6XtFjSEkmndlFnkqSbJd0u6foq4zEzs2pUdppLUhtwNnAosByYL2lORNxRqDMC+C5weEQ8IOnlVcVjZmbVqfLIZD9gSUQsjYhngcuAI+vqvB+YFREPAETEwxXGY2ZmFakymYwBlhWGl+eyot2Bl0maJ2mBpOMaNSTpREkdkjpWrVpVUbhmZraxqkwmalAWdcNDgH2BtwOTgS9I2v1v3hRxbkS0R0T7qFGj+j5SMzMrpcpbg5cD4wrDY4GVDeo8EhFPA09LugHYC7i7wrjMzKyPVXlkMh/YTdLOkjYHjgHm1NW5Avg7SUMkbQnsD9xZYUxmZlaByo5MImKdpJOAuUAbcH5E3C5pah5/TkTcKelXwK3Ac8B5EXFbVTGZmVk1FFF/GWNga29vj46Ojv4Ow8yspUhaEBHtVbXvb8CbmVlpTiZmZlaak4mZmZXmZGJmZqU5mZiZWWlOJmZmVpqTiZmZleZkYmZmpTmZmJlZaU4mZmZWmpOJmZmV5mRiZmalOZmYmVlpTiZmZlaak4mZmZXmZGJmZqU5mZiZWWlOJmZmVpqTiZmZleZkYmZmpTmZmJlZaU4mZmZWmpOJmZmV5mRiZmalOZmYmVlpTiZmZlaak4mZmZXmZGJmZqU5mZiZWWmVJhNJh0taLGmJpFMbjJ8k6XFJN+e/06qMx8zMqjGkqoYltQFnA4cCy4H5kuZExB11VX8TEUdUFYeZmVWvyiOT/YAlEbE0Ip4FLgOOrHB6ZmbWT6pMJmOAZYXh5bms3oGSbpH0S0mvadSQpBMldUjqWLVqVRWxmplZCVUmEzUoi7rhPwI7RcRewHeA2Y0aiohzI6I9ItpHjRrVx2GamVlZVSaT5cC4wvBYYGWxQkQ8ERFP5ddXAUMljawwJjMzq0CVyWQ+sJuknSVtDhwDzClWkPQKScqv98vxrK4wJjMzq0Bld3NFxDpJJwFzgTbg/Ii4XdLUPP4c4D3AxyStAzqBYyKi/lSYmZkNcGq1bXd7e3t0dHT0dxhmZi1F0oKIaK+qfX8D3szMSnMyMTOz0pxMzMysNCcTMzMrzcnEzMxKczIxM7PSnEzMzKw0JxMzMyvNycTMzEpzMjEzs9KcTMzMrDQnEzMzK83JxMzMSnMyMTOz0pxMzMysNCcTMzMrzcnEzMxKczIxM7PSnEzMzKw0JxMzMyvNycTMzEpzMjEzs9KcTMzMrDQnEzMzK83JxMzMShvS3wG0qtkLVzBj7mJWrulk9IhhTJs8gSn7jOnvsMzM+oWTyUaYvXAF02ctonPtegBWrOlk+qxFAE4oZvai5NNcG2HG3MXPJ5KazrXrmTF3cT9FZGbWv7o9MpH0L92Nj4j/7NtwWsPKNZ29KjczG+x6OjIZ3sNftyQdLmmxpCWSTu2m3uslrZf0nuZD7z+jRwzrVbmZ2WDX7ZFJRJy5sQ1LagPOBg4FlgPzJc2JiDsa1Ps6MHdjp7WpTZs8YYNrJgDDhrYxbfKEfozKzKz/9HSa69vdjY+Ik7sZvR+wJCKW5rYuA44E7qir9wngZ8Dre4x2gKhdZPfdXGZmSU93cy0o0fYYYFlheDmwf7GCpDHAUcAhdJNMJJ0InAiw4447lgip70zZZ4yTh5lZ1tNprgtLtK1GTdYNfxP4bESslxpVfz6Oc4FzAdrb2+vbMDOzftbU90wkjQI+C+wBbFErj4hDunnbcmBcYXgssLKuTjtwWU4kI4G/l7QuImY3E5eZmQ0MzX7P5BLgTmBn4EzgPmB+D++ZD+wmaWdJmwPHAHOKFSJi54gYHxHjgcuBjzuRmJm1nmaTyXYR8d/A2oi4PiI+AhzQ3RsiYh1wEukurTuBmRFxu6SpkqaWitrMzAaUZh+nsjb/f1DS20mnq8b29KaIuAq4qq7snC7qHt9kLGZmNsA0m0y+LOmlwKeA7wDbAP9cWVRmZtZSmkomEXFlfvk4cHB14ZiZWStq6pqJpAsljSgMv0zS+dWFZWZmraTZC/ATI2JNbSAiHgP2qSYkMzNrNc0mk80kvaw2IGlb/FsoZmaWNZsQvgHcKOly0rfY/x/wlcqiMjOzltLsBfiLJHWQnqEl4F31T/81M7MXr9780uK2wNMR8R1glaSdK4rJzMxaTLN3c51OejbX9Fw0FLi4qqDMzKy1NHtkchTwTuBpgIhYSRO/tGhmZi8OzSaTZyMiyI+Ql7RVdSGZmVmraTaZzJT0fWCEpI8C/wucV11YZmbWSpq9m+s/JB0KPAFMAE6LiGsqjczMzFpG0188zMnjGgBJbZKOjYhLKovMzMxaRrenuSRtI2m6pLMkHabkJGAp6YuLZmZmPR6Z/Ah4DLgJOAGYBmwOHBkRN1ccm5mZtYiekskuEbEngKTzgEeAHSPiycojMzOzltHT3Vy1X1gkItYD9zqRmJlZvZ6OTPaS9ER+LWBYHhYQEbFNpdGZmVlL6DaZRETbpgrEzMxaV28e9GhmZtaQk4mZmZXmZGJmZqU5mZiZWWlOJmZmVpqTiZmZleZkYmZmpTmZmJlZaU4mZmZWWqXJRNLhkhZLWiLp1Abjj5R0q6SbJXVIOqjKeMzMrBpN/zhWb0lqA84GDgWWA/MlzYmIOwrVrgXmRERImgjMBF5VVUxmZlaNKo9M9gOWRMTSiHgWuAw4slghIp6KiMiDWwGBmZm1nCqTyRhgWWF4eS7bgKSjJN0F/AL4SIXxmJlZRapMJmpQ9jdHHhHx84h4FTAF+FLDhqQT8zWVjlWrVvVxmGZmVlaVyWQ5MK4wPBZY2VXliLgB2FXSyAbjzo2I9ohoHzVqVN9HamZmpVSZTOYDu0naWdLmwDHAnGIFSa+UpPz6daTfl19dYUxmZlaByu7mioh1kk4C5gJtwPkRcbukqXn8OcC7geMkrQU6gfcWLsibmVmLUKttu9vb26Ojo6O/wzAzaymSFkREe1Xt+xvwZmZWWmWnuax/zV64ghlzF7NyTSejRwxj2uQJTNnnb+7MNjPrE04mg9DshSuYPmsRnWvXA7BiTSfTZy0CcEIxs0r4NNcgNGPu4ucTSU3n2vXMmLu4nyIys8HOyWQQWrmms1flZmZlOZkMQqNHDOtVuZlZWU4mg9C0yRMYNrRtg7JhQ9uYNnlCP0VkZoOdL8APQrWL7L6by8w2FSeTQWrKPmOcPMxsk/FpLjMzK83JxMzMSnMyMTOz0pxMzMysNCcTMzMrzcnEzMxKczIxM7PSnEzMzKw0JxMzMyvNycTMzEpzMjEzs9KcTMzMrDQnEzMzK83JxMzMSnMyMTOz0pxMzMysNCcTMzMrzcnEzMxKczIxM7PSnEzMzKw0JxMzMyut0mQi6XBJiyUtkXRqg/HHSro1/90oaa8q4zEzs2pUlkwktQFnA28D9gDeJ2mPumr3Am+OiInAl4Bzq4rHzMyqU+WRyX7AkohYGhHPApcBRxYrRMSNEfFYHvwdMLbCeMzMrCJVJpMxwLLC8PJc1pV/AH7ZaISkEyV1SOpYtWpVH4ZoZmZ9ocpkogZl0bCidDApmXy20fiIODci2iOifdSoUX0YopmZ9YUhFba9HBhXGB4LrKyvJGkicB7wtohYXWE8ZmZWkSqPTOYDu0naWdLmwDHAnGIFSTsCs4APRsTdFcZiZmYVquzIJCLWSToJmAu0AedHxO2Spubx5wCnAdsB35UEsC4i2quKyaw3Zi9cwYy5i1m5ppPRI4YxbfIEpuzT3WU/sxcvRTS8jDFgtbe3R0dHR3+HYYPc7IUrmD5rEZ1r1z9fNmxoG199155OKNaSJC2ocmfd34A3a2DG3MUbJBKAzrXrmTF3cT9FZDawOZmYNbByTWevys1e7JxMzBoYPWJYr8rNXuycTMwamDZ5AsOGtm1QNmxoG9MmT+iniMwGtiq/Z2LWsmoX2X03l1lznEzMujBlnzFOHmZN8mkuMzMrzcnEzMxKczIxM7PSnEzMzKw0JxMzMyvNycTMzEpzMjEzs9KcTMzMrDQnEzMzK83JxMzMSvPjVMzMBrhW+NVPJxMzswGs/lc/V6zpZPqsRQADKqH4NJeZ2QDWKr/66WRiZjaAtcqvfjqZmJkNYK3yq59OJmZmA1ir/OqnL8CbmQ1grfKrn04mZmYDXCv86qdPc5mZWWlOJmZmVpqTiZmZleZkYmZmpTmZmJlZaZUmE0mHS1osaYmkUxuMf5WkmyQ9I+nTVcZiZmbVqezWYEltwNnAocByYL6kORFxR6Hao8DJwJSq4jAzs+pVeWSyH7AkIpZGxLPAZcCRxQoR8XBEzAfWVhiHmZlVrMpkMgZYVhhenst6TdKJkjokdaxatapPgjMzs75TZTJRg7LYmIYi4tyIaI+I9lGjRpUMy8zM+lqVj1NZDowrDI8FVpZtdMGCBY9Iur9sO/1gJPBIfwfRRwZLXwZLP8B9GYgGWj92qrLxKpPJfGA3STsDK4BjgPeXbTQiWvLQRFJHRLT3dxx9YbD0ZbD0A9yXgWiw9KNZlSWTiFgn6SRgLtAGnB8Rt0uamsefI+kVQAewDfCcpFOAPSLiiariMjOzvlfpU4Mj4irgqrqycwqvHyKd/jIzsxbmb8BvOuf2dwB9aLD0ZbD0A9yXgWiw9KMpitioG6zMzMye5yMTMzMrzcnEzMxKczKpgKTzJT0s6bZC2baSrpH0p/z/Zf0ZYzMkjZN0naQ7Jd0u6ZO5vBX7soWkP0i6JfflzFzecn2B9Ow7SQslXZmHW7Uf90laJOlmSR25rFX7MkLS5ZLuyuvMga3al43hZFKNC4DD68pOBa6NiN2Aa/PwQLcO+FREvBo4APgnSXvQmn15BjgkIvYC9gYOl3QArdkXgE8CdxaGW7UfAAdHxN6F72S0al++BfwqIl4F7EX6fFq1L70XEf6r4A8YD9xWGF4M7JBf7wAs7u8YN6JPV5CeAt3SfQG2BP4I7N+KfSHdTn8tcAhwZS5ruX7kWO8DRtaVtVxfSN+Vu5d8U1Mr92Vj/3xksulsHxEPAuT/L+/neHpF0nhgH+D3tGhf8qmhm4GHgWsiolX78k3gM8BzhbJW7Aek5/VdLWmBpBNzWSv2ZRdgFfDDfPrxPElb0Zp92ShOJtYjSVsDPwNOiRZ+OkFErI+IvUl79vtJem1/x9Rbko4AHo6IBf0dSx95Y0S8Dngb6TTqm/o7oI00BHgd8L2I2Ad4msF8SqsBJ5NN58+SdgDI/x/u53iaImkoKZFcEhGzcnFL9qUmItYA80jXtVqtL28E3inpPtJvBB0i6WJarx8ARMTK/P9h4Oek30Fqxb4sB5bno12Ay0nJpRX7slGcTDadOcCH8usPka4/DGiSBPw3cGdE/GdhVCv2ZZSkEfn1MOCtwF20WF8iYnpEjI2I8aSHp/46Ij5Ai/UDQNJWkobXXgOHAbfRgn2J9GioZZIm5KK3AHfQgn3ZWP4GfAUk/RiYRHoE9Z+B04HZwExgR+AB4OiIeLS/YmyGpIOA3wCLeOH8/OdI101arS8TgQtJDx3dDJgZEV+UtB0t1pcaSZOAT0fEEa3YD0m7kI5GIJ0mujQivtKKfQGQtDdwHrA5sBT4MHlZo8X6sjGcTMzMrDSf5jIzs9KcTMzMrDQnEzMzK83JxMzMSnMyMTOz0pxMbMCSFJK+URj+tKQz+qjtCyS9py/a6mE6R+cnyF5XVz5eUmd+Wu4tkm4sfEehr2M4XtJZVbRtVuNkYgPZM8C7JI3s70CKJLX1ovo/AB+PiIMbjLsn0tNy9yJ9B+ZzDaY1ZCPDNNuknExsIFtH+h3tf64fUX9kIemp/H+SpOslzZR0t6SvSTo2/5bJIkm7Fpp5q6Tf5HpH5Pe3SZohab6kWyX9Y6Hd6yRdSvoSZ308745CTZwAAAMASURBVMvt3ybp67nsNOAg4BxJM3ro6zbAY/l9x0v6qaT/Aa7OZdMKMZ1ZmO4Hct9ulvT9WqKT9OHcr+tJj2Cp1T86x3iLpBt6iMmsad7rsYHubOBWSf/ei/fsBbwaeJT0TeTzImI/pR/3+gRwSq43HngzsCtwnaRXAscBj0fE6yW9BPitpKtz/f2A10bEvcWJSRoNfB3Yl5QQrpY0JX/D/hDSt9Q7GsS5a36K8XDSY/H3L4w7EJgYEY9KOgzYLU9fwJz8QMRVwHtJD0tcK+m7wLGSrgHOzPE8DlwHLMztngZMjogVtcfLmPUFJxMb0CLiCUkXAScDnU2+bX7tsd+S7iHv3ZOOKIqnm2ZGxHPAnyQtBV5Fej7UxMJRz0tJG/JngT/UJ5Ls9cC8iFiVp3kJ8CbSI3S6c09+ijGS3ks6Cqv9qNo1hcduHJb/aglh6xzTRFLCmJ8eo8Yw0oME96+L5yfA7vm9vwUukDQTqD2406w0JxNrBd8k/ZjVDwtl68inafMDKTcvjHum8Pq5wvBzbLjM1z9LKEh7/p+IiLnFEfk5WE93EZ967EHP5rBh/4rTEvDViPh+XUyfAC6MiOl15VP4274BEBFTJe0PvB24WdLeEbG6D+K3FzlfM7EBL++hzyRdzK65j7RXDnAkMHQjmj5a0mb5OsoupF/Fmwt8TOnR+0jaPT/Rtju/B94saWS+ZvE+4PpexnIQcE8X4+YCH1H6XRkkjZH0ctKvLb4nv679dvpOOZ5JkrbL/Ti61pCkXSPi9xFxGvAIMK6XcZo15CMTaxXfAE4qDP8AuELSH0gb1a6OGrqzmLTR3x6YGhF/lXQe6VrKH/MRzypgSneNRMSDkqaTrk0IuCoimnnUeO2aiUin0U7oov2rJb0auCmfznoK+EBE3CHp86RrNJsBa4F/iojf5VuobwIeJB3V1e5AmyFptzzNa4FbmojTrEd+arCZmZXm01xmZlaak4mZmZXmZGJmZqU5mZiZWWlOJmZmVpqTiZmZleZkYmZmpf1/qUv+gxNRd5kAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(num_breeds, recalls);\n",
    "plt.xlabel(\"Number of Breeds\");\n",
    "plt.ylabel(\"Recall\");\n",
    "plt.title(f\"Num Breeds vs Recall - {size}x{size} Images - K=5 Cross Validation\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "jupyter": {
     "source_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading model: model_60x60_4breeds_0.3dropout_0.0001lr_500fc_20batch\n",
      "Loading X_test pickle...\n",
      "Loading y_test pickle...\n",
      "Getting predictions...\n",
      "Calculating accuracy...\n",
      "Loading model: model_60x60_8breeds_0.3dropout_0.0001lr_500fc_20batch\n",
      "Loading X_test pickle...\n",
      "Loading y_test pickle...\n",
      "Getting predictions...\n",
      "Calculating accuracy...\n",
      "Loading model: model_60x60_16breeds_0.3dropout_0.0001lr_500fc_20batch\n",
      "Loading X_test pickle...\n",
      "Loading y_test pickle...\n",
      "Getting predictions...\n",
      "Calculating accuracy...\n",
      "Loading model: model_60x60_32breeds_0.3dropout_0.0001lr_500fc_20batch\n",
      "Loading X_test pickle...\n",
      "Loading y_test pickle...\n",
      "Getting predictions...\n",
      "Calculating accuracy...\n",
      "Loading model: model_60x60_64breeds_0.3dropout_0.0001lr_500fc_20batch\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/nschenon/.local/lib/python3.7/site-packages/keras/engine/saving.py:341: UserWarning: No training configuration found in save file: the model was *not* compiled. Compile it manually.\n",
      "  warnings.warn('No training configuration found in save file: '\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading X_test pickle...\n",
      "Loading y_test pickle...\n",
      "Getting predictions...\n",
      "Calculating accuracy...\n"
     ]
    },
    {
     "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": [
    "# # Data params\n",
    "# size = 60\n",
    "# num_breeds = [4, 8, 16, 32, 64]\n",
    "\n",
    "# # Model params\n",
    "# dropout = 0.3\n",
    "# fully_connected_nodes = 500\n",
    "# learning_rate=0.0001\n",
    "\n",
    "# # Training params\n",
    "# batch_size = 20\n",
    "\n",
    "# models = []\n",
    "# for breed in num_breeds:\n",
    "#     model_name = \"model_\" + str(size) + \"x\" + str(size) + \"_\" + str(breed) + \"breeds_\" + str(dropout) + \"dropout_\" + str(learning_rate) + \"lr_\" + str(fully_connected_nodes) + \"fc_\" + str(batch_size) + \"batch\"\n",
    "\n",
    "#     models.append(model_name)\n",
    "\n",
    "# accuracies = []\n",
    "# for breed, model_name in zip(num_breeds, models):\n",
    "#     print(\"Loading model:\", model_name)\n",
    "#     model = load_model('models/' + model_name + '.hdf5')\n",
    "    \n",
    "#     print(\"Loading X_test pickle...\")\n",
    "#     X_test = pickle.load(open(\"pickles/X_test_\" + str(size) + \"x\" + str(size) + \"_\" + str(breed) + \".pkl\", \"rb\"))\n",
    "    \n",
    "#     print(\"Loading y_test pickle...\")\n",
    "#     y_test = pickle.load(open(\"pickles/y_test_\" + str(size) + \"x\" + str(size) + \"_\" + str(breed) + \".pkl\", \"rb\"))\n",
    "    \n",
    "#     print(\"Getting predictions...\")\n",
    "#     predictions = model.predict(X_test)\n",
    "\n",
    "    \n",
    "#     pred_argmax = []\n",
    "#     for pred in predictions:\n",
    "#         pred_argmax.append(pred.argmax())\n",
    "\n",
    "#     true_argmax = []\n",
    "#     for pred in y_test:\n",
    "#         true_argmax.append(pred.argmax())\n",
    "\n",
    "#     print(\"Calculating accuracy...\")\n",
    "#     accuracies.append(accuracy_score(pred_argmax, true_argmax))\n",
    "    \n",
    "# plt.scatter(num_breeds, accuracies);\n",
    "# plt.xlabel(\"Number of Breeds\");\n",
    "# plt.ylabel(\"Accuracy in %\");\n",
    "# plt.title(f\"Number of Breeds vs Accuracy - {size}x{size} Images\");"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Learning Rate vs Accuracy - 60x60 Images"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading model: model_60x60_8breeds_0.3dropout_0.0001lr_500fc_20batch\n",
      "Loading X_test pickle...\n",
      "Loading y_test pickle...\n",
      "Getting predictions...\n",
      "Calculating accuracy...\n",
      "Loading model: model_60x60_8breeds_0.3dropout_0.0005lr_500fc_20batch\n",
      "Loading X_test pickle...\n",
      "Loading y_test pickle...\n",
      "Getting predictions...\n",
      "Calculating accuracy...\n",
      "Loading model: model_60x60_8breeds_0.3dropout_0.001lr_500fc_20batch\n",
      "Loading X_test pickle...\n",
      "Loading y_test pickle...\n",
      "Getting predictions...\n",
      "Calculating accuracy...\n"
     ]
    },
    {
     "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": [
    "# Data params\n",
    "size = 60\n",
    "breed = 8\n",
    "\n",
    "# Model params\n",
    "dropout = 0.3\n",
    "fully_connected_nodes = 500\n",
    "learning_rates= [0.0001, 0.0005, 0.001]\n",
    "\n",
    "# Training params\n",
    "batch_size = 20\n",
    "\n",
    "models = []\n",
    "for learning_rate in learning_rates:\n",
    "    model_name = \"model_\" + str(size) + \"x\" + str(size) + \"_\" + str(breed) + \"breeds_\" + str(dropout) + \"dropout_\" + str(learning_rate) + \"lr_\" + str(fully_connected_nodes) + \"fc_\" + str(batch_size) + \"batch\"\n",
    "\n",
    "    models.append(model_name)\n",
    "\n",
    "accuracies = []\n",
    "for breed, model_name in zip(num_breeds, models):\n",
    "    print(\"Loading model:\", model_name)\n",
    "    model = load_model('models/' + model_name + '.hdf5')\n",
    "    \n",
    "    print(\"Loading X_test pickle...\")\n",
    "    X_test = pickle.load(open(\"pickles/X_test_\" + str(size) + \"x\" + str(size) + \"_\" + str(breed) + \".pkl\", \"rb\"))\n",
    "    \n",
    "    print(\"Loading y_test pickle...\")\n",
    "    y_test = pickle.load(open(\"pickles/y_test_\" + str(size) + \"x\" + str(size) + \"_\" + str(breed) + \".pkl\", \"rb\"))\n",
    "    \n",
    "    print(\"Getting predictions...\")\n",
    "    predictions = model.predict(X_test)\n",
    "\n",
    "    \n",
    "    pred_argmax = []\n",
    "    for pred in predictions:\n",
    "        pred_argmax.append(pred.argmax())\n",
    "\n",
    "    true_argmax = []\n",
    "    for pred in y_test:\n",
    "        true_argmax.append(pred.argmax())\n",
    "\n",
    "    print(\"Calculating accuracy...\")\n",
    "    accuracies.append(accuracy_score(pred_argmax, true_argmax))\n",
    "    \n",
    "plt.scatter(learning_rates, accuracies);\n",
    "plt.xlabel(\"Learning Rate\");\n",
    "plt.ylabel(\"Accuracy in %\");\n",
    "plt.title(f\"Learning Rate vs Accuracy - {size}x{size} Images\");"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "cogs118b",
   "language": "python",
   "name": "cogs118b"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}