7-3.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "pqxnslml3tu6"
   },
   "source": [
    "# 신경망 모델 훈련"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "fN7zF6Zy3tvE"
   },
   "source": [
    "<table align=\"left\">\n",
    "  <td>\n",
    "    <a target=\"_blank\" href=\"https://colab.research.google.com/github/rickiepark/hg-mldl/blob/master/7-3.ipynb\"><img src=\"https://www.tensorflow.org/images/colab_logo_32px.png\" />구글 코랩에서 실행하기</a>\n",
    "  </td>\n",
    "</table>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "XGP-X65EmJBg"
   },
   "source": [
    "## 손실 곡선"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "hZsGl9udlqZk",
    "outputId": "dfb7c783-8276-4c75-9a1c-6825310e29d8"
   },
   "outputs": [],
   "source": [
    "from tensorflow import keras\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "(train_input, train_target), (test_input, test_target) = \\\n",
    "    keras.datasets.fashion_mnist.load_data()\n",
    "\n",
    "train_scaled = train_input / 255.0\n",
    "\n",
    "train_scaled, val_scaled, train_target, val_target = train_test_split(\n",
    "    train_scaled, train_target, test_size=0.2, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "id": "iogH7o0Ll6uL"
   },
   "outputs": [],
   "source": [
    "def model_fn(a_layer=None):\n",
    "    model = keras.Sequential()\n",
    "    model.add(keras.layers.Flatten(input_shape=(28, 28)))\n",
    "    model.add(keras.layers.Dense(100, activation='relu'))\n",
    "    if a_layer:\n",
    "        model.add(a_layer)\n",
    "    model.add(keras.layers.Dense(10, activation='softmax'))\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "5Eh6hM4DNdzu",
    "outputId": "0e2c29b8-8fd5-4432-a667-020a01c36fa7"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " flatten (Flatten)           (None, 784)               0         \n",
      "                                                                 \n",
      " dense (Dense)               (None, 100)               78500     \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 10)                1010      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 79,510\n",
      "Trainable params: 79,510\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-01-08 04:42:57.842599: I tensorflow/core/platform/cpu_feature_guard.cc:151] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n"
     ]
    }
   ],
   "source": [
    "model = model_fn()\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "id": "P-UK21N_mCM0"
   },
   "outputs": [],
   "source": [
    "model.compile(loss='sparse_categorical_crossentropy', metrics='accuracy')\n",
    "\n",
    "history = model.fit(train_scaled, train_target, epochs=5, verbose=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "1ix_EA2LpaR0",
    "outputId": "63049ab7-b817-4fcd-9f8a-ab73d73f89cd"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dict_keys(['loss', 'accuracy'])\n"
     ]
    }
   ],
   "source": [
    "print(history.history.keys())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 279
    },
    "id": "uz_TCdfPmG6e",
    "outputId": "033f08e3-9dee-4e4a-85b6-dacbfdc8ae38"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEGCAYAAAB/+QKOAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAAi+ElEQVR4nO3deXxddZ3/8dcna9M0SZulW5o2Sfe0sobSslQEgbK0zE9Qq4OCW9WRB/pDR0VlyuLPZZxxfuOIP6wggqLIMKPTArIvZemWQlm6p+mWrmnSJume5fP7456GUG7bW5p7z03yfj4e99Hcc8/NfefAzTvnnO/9HnN3REREjpYSdgAREUlOKggREYlKBSEiIlGpIEREJCoVhIiIRJUWdoCuUlhY6KWlpWHHEBHpVpYuXbrL3YuiPdZjCqK0tJSqqqqwY4iIdCtmtvFYj+kQk4iIRKWCEBGRqFQQIiISlQpCRESiUkGIiEhUKggREYlKBSEiIlH1+oKo33uIO+Ytp/FAS9hRRESSSq8viG2NB/ndaxv492fXhh1FRCSp9PqCmFicx8xzhvPAgg2s3dEcdhwRkaTR6wsC4FuXjSE7I5U75q1AV9gTEYlQQQAF/TK55dIxvFK9i6dX7Ag7johIUlBBBK6fPIIxg/px12MrONjSFnYcEZHQqSACaakpzJ4+gdrdB/jN/Jqw44iIhE4F0cn5owq5YuJg7n6xmq17DoQdR0QkVCqIo3zvyvG4w4//tirsKCIioVJBHKUkvy9f/vBI5r25lUU19WHHEREJjQoiiq9+eCRD8/pw+7wVtLVr2KuI9E4qiCiyMlL5/lUVrNzWxJ8Wbwo7johIKFQQx3DlhwZzblk+//L0avbsPxx2HBGRhFNBHIOZcfuMCTQdaOHnz6wJO46ISMKpII5j/JBcrp88gj8s3Miq7U1hxxERSSgVxAnccukYcrPSuX3ucs3TJCK9igriBPr3zeCbl41lYU0DT7y9Pew4IiIJo4KIwacnDWf8kFx+9MRKDhzWPE0i0juoIGKQmmLcPr2CLXsOcM9L68KOIyKSECqIGJ1bXsDVpw3hnpfWUbt7f9hxRETiTgVxEr535XjM4EdPrAw7iohI3KkgTsLQ/ln8w0WjeOLt7bxWvSvsOCIicaWCOEmzppYzbEAWd8xbQWtbe9hxRETiRgVxkvqkp/KDqypYvaOZhxZpniYR6blUEB/A5RMGccGoQv716dU07NM8TSLSM6kgPgAzY/b0CvYdbuNfnl4ddhwRkbhQQXxAowfl8NkpI/jT4k28s6Ux7DgiIl1OBXEKvvHRMQzom8Ed8zRPk4j0PHEtCDObZmarzazazL4b5fEbzazOzJYFty92euwGM1sb3G6IZ84PKi8rnX+8fCxLNuxm7ptbw44jItKl4lYQZpYK3A1cAVQAnzKziiir/tndzwhu9wbPzQdmA+cCk4DZZjYgXllPxScqS5hYnMuPn1jF/sOtYccREeky8dyDmARUu3uNux8GHgauifG5lwPPuHuDu+8GngGmxSnnKYnM0zSB7U0H+dULmqdJRHqOeBZEMbC50/3aYNnRrjWzt8zsUTMrOcnnJoXK0nz+7oyhzJlfw8b6fWHHERHpEmGfpJ4HlLr7aUT2Eh44mSeb2SwzqzKzqrq6urgEjNV3rxhPWqrxw8c1T5OI9AzxLIgtQEmn+8OCZR3cvd7dDwV37wXOjvW5wfPnuHulu1cWFRV1WfAPYnBeH266eBTPrNjB/DXhlpWISFeIZ0EsAUabWZmZZQAzgbmdVzCzIZ3uzgCO/Pn9FHCZmQ0ITk5fFixLal+4oIwRBX25Y95yWjRPk4h0c3ErCHdvBW4i8ot9JfCIuy83szvNbEaw2s1mttzM3gRuBm4MntsA3EWkZJYAdwbLklpmWiq3XVXBurp9PPDahrDjiIicEuspH/CqrKz0qqqqsGPg7tx4/xJe37ibF/7xIgr7ZYYdSUTkmMxsqbtXRnss7JPUPY6ZcdvVFRxoaeNnT2qeJhHpvlQQcTBqYD8+d34pjyzdzFu1e8KOIyLygagg4uTmS0ZTkJ3J7XOX097eMw7jiUjvooKIk5w+6Xxn2lhe37SHvy573whdEZGkp4KIo2vPGsbpJf358d9WsfeQ5mkSke5FBRFHKSnG7dMrqGs+xC+frw47jojISVFBxNmZwwdw3dnDuO+VGtbv0jxNItJ9qCAS4NvTxpKZlspdj60IO4qISMxUEAkwMKcPN18yiudX7eSFVTvDjiMiEhMVRILceF4Z5YXZ3PnYCg63ap4mEUl+KogEyUhL4bbpFazftY/7X10fdhwRkRNSQSTQR8YO5JJxA/nFc2vZ2XQw7DgiIselgkiw266uoKXN+anmaRKRJKeCSLDSwmw+f0EZ//V6LW9s2h12HBGRY1JBhOCmi0cxMEfzNIlIclNBhKBfZhq3XjmON2sbefT12rDjiIhEpYIIyd+dUcxZw/vzz0+uoulgS9hxRETeRwUREjPj9hkTqN93mF88uzbsOCIi76OCCNFpw/rzycoSfvfaBqp37g07jojIe6ggQvaty8eSlZHKHfOW01OuDy4iPYMKImSF/TL5xkfH8PLaXTy7UvM0iUjyUEEkgc9OGcGogf2467EVHGxpCzuOiAiggkgK6akpzJ5ewaaG/dz3iuZpEpHkoIJIEheOLuKyikHc/UI12xs1T5OIhE8FkUR+cFUFre3OT/62MuwoIiIqiGQyvKAvX55azl+XbaVqQ0PYcUSkl1NBJJmvXjSSIXl9mD13OW2ap0lEQqSCSDJ9M9K49crxLN/axCNVm8OOIyK9mAoiCU0/bQiTSvP52VOradyveZpEJBwqiCRkZsyeUcGe/Yf5t2fXhB1HRHopFUSSmjA0j09NGs7vF25kzY7msOOISC+kgkhi37xsLP0y0zRPk4iEQgWRxPKzM7jl0jG8Wl3PU8u3hx1HRHoZFUSS+/tzhzN2UA4/fHyl5mkSkYRSQSS5tNQUZs+ooHb3AebMrwk7joj0IiqIbuC8kYVc9aEh/OrFarbsORB2HBHpJVQQ3cStV47DHX78hOZpEpHEUEF0E8MG9OWrF43ksbe2sbCmPuw4ItILxLUgzGyama02s2oz++5x1rvWzNzMKoP7pWZ2wMyWBbd74pmzu/jKh0dS3D+L2+cup7WtPew4ItLDxa0gzCwVuBu4AqgAPmVmFVHWywG+Diw66qF17n5GcPtKvHJ2J33SU/n+VeNZtb2ZPy3RPE0iEl/x3IOYBFS7e427HwYeBq6Jst5dwE8BXSUnBldMHMyU8gL+9enV7N53OOw4ItKDxbMgioHOf+bWBss6mNlZQIm7Px7l+WVm9oaZvWRmF8YxZ7dyZJ6m5oOt/PwZzdMkIvET2klqM0sBfg58M8rD24Dh7n4mcAvwRzPLjfI9ZplZlZlV1dXVxTdwEhk3OJfrzx3OQ4s2smJrU9hxRKSHimdBbAFKOt0fFiw7IgeYCLxoZhuAycBcM6t090PuXg/g7kuBdcCYo1/A3ee4e6W7VxYVFcXpx0hO//vSMeRlpWueJhGJm3gWxBJgtJmVmVkGMBOYe+RBd29090J3L3X3UmAhMMPdq8ysKDjJjZmVA6MBfYy4k/59M/jW5WNZtL6Bx9/eFnYcEemB4lYQ7t4K3AQ8BawEHnH35WZ2p5nNOMHTpwJvmdky4FHgK+6uizQfZeY5w6kYksuPHl/JgcOap0lEupb1lMMTlZWVXlVVFXaMhFu8voFP/HoBN18ymlsufd9ROBGR4zKzpe5eGe0xfZK6m5tUls+M04dyz0vr2NywP+w4ItKDqCB6gFuvHEeqGf/ncc3TJCJdRwXRAwzJy+JrHxnJk8u382r1rrDjiEgPoYLoIb54YTkl+VncMW85LZqnSUS6gAqih+iTnsptV1WwZsde/rBwY9hxRKQHUEH0IJdWDOLC0YX82zNrqN97KOw4ItLNqSB6EDNj9vQK9h9u41+e1jxNInJqVBA9zKiBOdxwXikPL9nEO1saw44jIt2YCqIHuvmS0eT3zeD2uZqnSUQ+OBVED5SXlc63p42lauNu5r65New4ItJNxVQQZvZ1M8u1iPvM7HUzuyze4eSD+/jZJZw2LI8fPbGSfYdaw44jIt1QrHsQn3f3JuAyYADwGeAncUslpywlxZg9fQI7mg7xqxerw44jIt1QrAVhwb9XAr939+WdlkmSOnvEAD52ZjG/mb+ejfX7wo4jIt1MrAWx1MyeJlIQT5lZDqCP63YD37liHOmpxl2PaZ4mETk5sRbEF4DvAue4+34gHfhc3FJJlxmU24ebLh7Nsyt38NKa3nNZVhE5dbEWxBRgtbvvMbPrgR8AGmTfTXz+glJKC/pyx7zlHG7Vjp+IxCbWgvh/wH4zOx34JpFrRD8Yt1TSpTLTUvmn6RXU1O3jwQUbwo4jIt1ErAXR6pFPXF0D/NLd7wZy4hdLutrF4wbxkbFF/Puza6lr1jxNInJisRZEs5ndSmR46+NmlkLkPIR0I7ddXcHB1jZ+9tSqsKOISDcQa0F8EjhE5PMQ24FhwM/ilkrioryoH58/v4xHqmpZtnlP2HFEJMnFVBBBKTwE5JnZ1cBBd9c5iG7opotHUZSTye1zl9PernmaROTYYp1q4xPAYuDjwCeARWZ2XTyDSXzk9EnnO9PGsWzzHv7yxpaw44hIEov1ENP3iXwG4gZ3/ywwCbgtfrEknj52ZjFnlPTnJ0+uovlgS9hxRCRJxVoQKe6+s9P9+pN4riSZlBTjjhkTqGs+xC+f1zxNIhJdrL/knzSzp8zsRjO7EXgceCJ+sSTeTi/pz8fPHsZvX11PTd3esOOISBKK9ST1PwJzgNOC2xx3/048g0n8fXvaOPqkpXLXYyvCjiIiSSjmw0Tu/l/ufktw+0s8Q0liFOVk8vWPjuaF1XU8v2pH2HFEJMkctyDMrNnMmqLcms2sKVEhJX4+O6WU8qJs7npsJYda28KOIyJJ5LgF4e457p4b5Zbj7rmJCinxk5GWwj9dXcH6Xfu4/9UNYccRkSSikUjCRWMH8tHxg/iP59ays+lg2HFEJEmoIASA264eT0ub85MnNU+TiESoIASAEQXZfPHCMv779S0s3bg77DgikgRUENLhax8ZxaDcTO6Yp3maREQFIZ1kZ6bxvSvH81ZtI48urQ07joiETAUh7zHj9KFUjhjAT59cReMBzdMk0pupIOQ9zIzbZ0ygYf9hfvHc2rDjiEiIVBDyPhOL85h5znAeeG0D1Tubw44jIiFRQUhU37psDFkZqdwxbwWRy5GLSG8T14Iws2lmttrMqs3su8dZ71ozczOr7LTs1uB5q83s8njmlPcr6JfJLZeO4eW1u3hmheZpEumN4lYQZpYK3A1cAVQAnzKziijr5QBfBxZ1WlYBzAQmANOAXwXfTxLo+skjGDOoHz98fCUHWzRPk0hvE889iElAtbvXuPth4GHgmijr3QX8FOg8x8M1wMPufsjd1wPVwfeTBEpPTWH29AlsatjPfa+sDzuOiCRYPAuiGNjc6X5tsKyDmZ0FlLj74yf7XEmM80cVMm3CYH75fDXbGg+EHUdEEii0k9RmlgL8HPjmKXyPWWZWZWZVdXV1XRdO3uP7V42n3Z0fP6F5mkR6k3gWxBagpNP9YcGyI3KAicCLZrYBmAzMDU5Un+i5ALj7HHevdPfKoqKiLo4vR5Tk9+XLU8uZ++ZWlmxoCDuOiCRIPAtiCTDazMrMLIPISee5Rx5090Z3L3T3UncvBRYCM9y9KlhvppllmlkZMBpYHMescgJfvWgUQ/P6MPt/ltOmeZpEeoW4FYS7twI3AU8BK4FH3H25md1pZjNO8NzlwCPACuBJ4GvurmE0IcrKSOV7V41nxbYmHl6yKew4IpIA1lM+BFVZWelVVVVhx+jR3J2ZcxayZkczT35jKoNy+4QdSUROkZktdffKaI/pk9QSMzNj9vQJ7DvUxtR/foHv/eVtaur2hh1LROJEBSEnpWJoLk9+40I+dtYwHl1ayyU/f4lZD1axdKNOXov0NDrEJB9YXfMhHlywgd8v3Mie/S2cPWIAs6aWc+n4QaSkWNjxRCQGxzvEpIKQU7b/cCuPLNnMva+sp3b3AcoKI5cvvfasYfRJ1wwpIslMBSEJ0drWzpPLtzNnfg1v1TZSkJ3BDeeV8pnJIxiQnRF2PBGJQgUhCeXuLKxpYM78dbywuo6s9FQ+UTmML1xQzvCCvmHHE5FOjlcQaYkOIz2fmTFlZAFTRhawZkczv5lfwx8Xb+L3CzdyxcQhzJpazukl/cOOKSInoD0ISYgdTQe5/9UNPLRoI80HWzm3LJ8vf7ici8YM1AltkRDpEJMkjb2HWnl48SZ++8p6tjYeZPTAfnzpwnKuOXMomWk6oS2SaCoISTotbe08/tY2fj2/hpXbmhiYk8mN55fy9+eOIC8rPex4Ir2GCkKSlrvzSvUu5syv4eW1u8jOSGXmpOF8/oIyivtnhR1PpMdTQUi3sGJrE795uYZ5b27FgatPG8KXLixnYnFe2NFEeiwVhHQrW/cc4P5X1/OnxZvZe6iV80cVMGvqSKaOLsRMJ7RFupIKQrqlxgMt/GnxJu5/dT07mg4xbnAOs6aWM/30oaSnahoxka6ggpBu7XBrO/+zbAu/ebmGNTv2MiSvD58/v4yZk0rI6aMT2iKnQgUhPYK78+KaOua8VMOCmnpyMtP49LnD+dz5ZQzO07UpRD4IFYT0OG/V7mHO/BqeeHsbqSnGjNOLmTW1nLGDc8KOJtKtqCCkx9rcsJ/7XlnPn5ds5kBLGx8eU8SXp5YzZWSBTmiLxEAFIT3env2H+cPCjfzutQ3s2nuYicW5zJo6kisnDiZNJ7RFjkkFIb3GwZY2/vrGFua8XENN3T6GDcjiCxeU8YnKErIzNTelyNFUENLrtLc7z63ayZz561iyYTd5WelcP3k4N5xXysAcndAWOUIFIb3a65t285v5NTy5fDvpKSl87KxivnhhOaMG9gs7mkjoVBAiwIZd+7j3lRr+s6qWQ63tfHT8QGZNHck5pQN0Qlt6LRWESCf1ew/x4IKNPLhgA7v3t3BGSX9mTS3n8gmDSdW1KaSXUUGIRHHgcBuPvl7LvS/XsLF+PyMK+vLFC8q47uwSsjJ0bQrpHVQQIsfR1u48vXw7v55fw7LNe8jPzuAzk0fw2SkjKOiXGXY8kbhSQYjEwN1ZsmE3c+bX8OzKHWSmpXDd2cP40oXllBZmhx1PJC6OVxAaGC4SMDMmleUzqSyf6p17ufflyAntPy7exOUVg5n14XLOGj4g7JgiCaM9CJHj2Nl8kAdf28jvF26k8UAL55QOYNbUkVwybiApOqEtPYAOMYmcon2HWnmkajP3vbKe2t0HKC/K5ksXlvO/ziymT7pOaEv3pYIQ6SKtbe088c525sxfxztbmijsl8mN543g+skj6N83I+x4IidNBSHSxdydBTX1zJlfw4ur68hKT+WT55TwhQvKKMnvG3Y8kZipIETiaPX2ZubMr2Hum1toa3eu/NAQvjx1JB8alhd2NJETUkGIJMD2xoPc/9p6/rhwE82HWplSXsCsqeVcNLZIU3lI0lJBiCRQ88EWHl68md++up5tjQcZM6gf008bypSRBZw2rD8Zabo+hSQPFYRICFra2nnsra387tUNvFnbCEBWeiqVpQOYXF7A5PICThuWR7ouaCQhUkGIhGz3vsMsWt/Awpp6FtbUs2p7MwB9M1KpLM1ncnk+k8sL+FCxCkMSSwUhkmQa9h1mUVAWC2rqWbNjLwDZQWFMGRnZw5g4NFeXTJW4Cq0gzGwa8O9AKnCvu//kqMe/AnwNaAP2ArPcfYWZlQIrgdXBqgvd/SvHey0VhHRnu/YeYlFNQ0dhVO+MFEa/zDTOCQ5JTRlZwISheZqSXLpUKAVhZqnAGuBSoBZYAnzK3Vd0WifX3ZuCr2cA/+Du04KCeMzdJ8b6eioI6Unqmg91HI5aWFPPurp9AORkpjGpLL+jMMYPyVVhyCkJa7K+SUC1u9cEIR4GrgE6CuJIOQSygZ5xvEvkFBXlZDL99KFMP30oADubDrJwfQML1tWzqKae51btBCC3TxqTygqYXB45LDV+cK7miJIuE8+CKAY2d7pfC5x79Epm9jXgFiADuLjTQ2Vm9gbQBPzA3V+OY1aRpDYwtw8zTh/KjKAwtjceZNH6ehasi+xhPLtyBwB5WemcG+xhTC4vYNzgHBWGfGChT/ft7ncDd5vZp4EfADcA24Dh7l5vZmcDfzWzCUftcWBms4BZAMOHD09wcpHwDM7rwzVnFHPNGcUAbGs8EDkcta6BBTX1PL0iUhj9+0YKY0p5AZNHFjBmoApDYhfPcxBTgNvd/fLg/q0A7v7jY6yfAux29/fNT2BmLwLfcvdjnmTQOQiRd23Zc4CF694dJVW7+wAA+dkZkcIIRkmNHthPn/Lu5cI6B7EEGG1mZcAWYCbw6aOCjXb3tcHdq4C1wfIioMHd28ysHBgN1MQxq0iPUtw/i2vPHsa1Zw8DYHPD/uCEd2Sk1N/e2Q5AQXZGcDgqUhoji1QY8q64FYS7t5rZTcBTRIa5/tbdl5vZnUCVu88FbjKzjwItwG4ih5cApgJ3mlkL0A58xd0b4pVVpKcrye9LSX5fPl5ZgrtTu/sAC2rqWbgusofx+NvbACjsl9nxob0pIwsoL8xWYfRi+qCcSC/n7mwK9jAWBIWxo+kQEBlNNbm8IHIOozyfMhVGj6NrUovIMZkZIwqyGVGQzSfPGY67s6H+3cJYWFPPvDe3AjAot3NhFDCioK8KowdTQYjIe5gZZYXZlBVm86lJkcJYv2tf5JBUTQOvVtfzP8sihTEkr8+75zDKCynJz1Jh9CA6xCQiJ8XdWVd3pDAiH9zbtfcwAEPz+jA5GCE1pbxAV9frBjRZn4jEjbtTvXNvR2EsrGmgYV+kMIr7Z3UMqZ0ysoDi/lkhp5WjqSBEJGHa2521O/d2nMNYtL6e3ftbACjJz2JyWUFHaQxVYYROBSEioWlvd1bvaO5UGA00HogUxoiCvu8pjMF5fUJO2/uoIEQkabS3Oyu3N3V8aG9RTT1NB1sBKCvMZnJ5PmeU9Gfc4FzGDMohKyM15MQ9mwpCRJJWW7uzcltTx9Tmi9Y30BwUhhmUFWQzbkgO4wbnMm5wDuOH5FLcP0tzSnURFYSIdBvt7c7m3ftZua2ZVdubWBX8u7FhP0d+XfXLTGPs4BzGDc5h3JBcxg/OYczgHHL7pIcbvhtSQYhIt7fvUCtrdjSzanszq7Y1sTL498jhKYBhA7IipTE4t2Ovo6wwWxdVOg59klpEur3szDTOHD6AM4cP6Fjm7mxrPMiq7U3BHkekNF5YXUdbe+SP38y0FMYMeu/exrghueRnZ4T1o3QbKggR6bbMjKH9sxjaP4uLxw3qWH6wpY3qnXs7CmPV9mZeWL2T/1xa27HOwJzMToUR2dsYWdSPjLSUMH6UpKSCEJEep096KhOL85hY/N7Ly9Q1H2L19uZOexxN3P9qPYfb2gFISzFGDezXsbdx5KT4wJzMXjmFiApCRHqNopxMinIyuWB0YceylrZ2Nuza13FOY9X2Zhavb+CvwXxTAAP6pnec1xgf/Dt6YM8fgquCEJFeLT01hdGDchg9KKfjmt8AjftbIqOoOu1xPLx4Mwda2gBIMSgtzI4URqc9jmEDes6EhSoIEZEo8vqmc255AeeWF3Qsa2+PXDvjSGGs3t7M8q2NHRdcgsgQ3HGdzmuMG5zD2ME55HTDIbga5ioicopiH4Kby/gh7w7DLS0IfwiuhrmKiMTRyQ3B3dkxBLdPeqchuJ3OcQxIkiG4KggRkTg4mSG4z63cySNV7w7BHZSb+b6T4uWFiR+Cq4IQEUmg4w3BPTK1yMrg3wXr3h2Cm55qjCzqx/ghue/50F9RHIfgqiBERJJAZAhuEReOLupYFm0I7qKaev7yxpaOdfKzM7j2rGK+f1VFl2dSQYiIJKlYh+AOyYvPhZdUECIi3Uy0IbjxoElHREQkKhWEiIhEpYIQEZGoVBAiIhKVCkJERKJSQYiISFQqCBERiUoFISIiUfWY6b7NrA7YeArfohDY1UVxupJynRzlOjnKdXJ6Yq4R7l4U7YEeUxCnysyqjjUnepiU6+Qo18lRrpPT23LpEJOIiESlghARkahUEO+aE3aAY1Cuk6NcJ0e5Tk6vyqVzECIiEpX2IEREJCoVhIiIRNWrCsLMppnZajOrNrPvRnk808z+HDy+yMxKkyTXjWZWZ2bLgtsXE5Trt2a208zeOcbjZma/CHK/ZWZnJUmui8yssdP2+qcE5SoxsxfMbIWZLTezr0dZJ+HbLMZcCd9mZtbHzBab2ZtBrjuirJPw92SMucJ6T6aa2Rtm9liUx7p+W7l7r7gBqcA6oBzIAN4EKo5a5x+Ae4KvZwJ/TpJcNwK/DGGbTQXOAt45xuNXAn8DDJgMLEqSXBcBj4WwvYYAZwVf5wBrovy3TPg2izFXwrdZsA36BV+nA4uAyUetE8Z7MpZcYb0nbwH+GO2/VTy2VW/ag5gEVLt7jbsfBh4GrjlqnWuAB4KvHwUuMTNLglyhcPf5QMNxVrkGeNAjFgL9zWxIEuQKhbtvc/fXg6+bgZVA8VGrJXybxZgr4YJtsDe4mx7cjh41k/D3ZIy5Es7MhgFXAfceY5Uu31a9qSCKgc2d7tfy/jdJxzru3go0AvG96GtsuQCuDQ5JPGpmJXHOFKtYs4dhSnCI4G9mNiHRLx7s3p9J5K/PzkLdZsfJBSFss+CQyTJgJ/CMux9zeyXwPRlLLkj8e/L/At8G2o/xeJdvq95UEN3ZPKDU3U8DnuHdvxIkuteJzC9zOvAfwF8T+eJm1g/4L+Ab7t6UyNc+nhPkCmWbuXubu58BDAMmmdnERLzuicSQK6HvSTO7Gtjp7kvj+TpH600FsQXo3PLDgmVR1zGzNCAPqA87l7vXu/uh4O69wNlxzhSrWLZpwrl705FDBO7+BJBuZoWJeG0zSyfyS/ghd//vKKuEss1OlCvMbRa85h7gBWDaUQ+F8Z48Ya4Q3pPnAzPMbAORw9AXm9kfjlqny7dVbyqIJcBoMyszswwiJ3HmHrXOXOCG4OvrgOc9OOMTZq6jjlHPIHIMORnMBT4bjMyZDDS6+7awQ5nZ4CPHXs1sEpH/z+P+SyV4zfuAle7+82OslvBtFkuuMLaZmRWZWf/g6yzgUmDVUasl/D0ZS65Evyfd/VZ3H+bupUR+Rzzv7tcftVqXb6u0U3lyd+LurWZ2E/AUkZFDv3X35WZ2J1Dl7nOJvIl+b2bVRE6CzkySXDeb2QygNch1Y7xzAZjZn4iMbik0s1pgNpETdrj7PcATREblVAP7gc8lSa7rgK+aWStwAJiZgKKHyF95nwHeDo5fA3wPGN4pWxjbLJZcYWyzIcADZpZKpJAecffHwn5PxpgrlPfk0eK9rTTVhoiIRNWbDjGJiMhJUEGIiEhUKggREYlKBSEiIlGpIEREJCoVhEgSsMhsqu+boVMkTCoIERGJSgUhchLM7PrgWgHLzOzXwaRue83s34JrBzxnZkXBumeY2cJgQre/mNmAYPkoM3s2mBjvdTMbGXz7fsHEb6vM7KF4z1oqciIqCJEYmdl44JPA+cFEbm3A3wPZRD7NOgF4icgnuwEeBL4TTOj2dqflDwF3BxPjnQccmWrjTOAbQAWR64OcH+cfSeS4es1UGyJd4BIik7ItCf64zyIyHXQ78OdgnT8A/21meUB/d38pWP4A8J9mlgMUu/tfANz9IEDw/Ra7e21wfxlQCrwS959K5BhUECKxM+ABd7/1PQvNbjtqvQ86f82hTl+3ofenhEyHmERi9xxwnZkNBDCzfDMbQeR9dF2wzqeBV9y9EdhtZhcGyz8DvBRc0a3WzP4u+B6ZZtY3kT+ESKz0F4pIjNx9hZn9AHjazFKAFuBrwD4iF5X5AZFDTp8MnnIDcE9QADW8O3PrZ4BfBzNxtgAfT+CPIRIzzeYqcorMbK+79ws7h0hX0yEmERGJSnsQIiISlfYgREQkKhWEiIhEpYIQEZGoVBAiIhKVCkJERKL6/6HrMPDAztlLAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.plot(history.history['loss'])\n",
    "plt.xlabel('epoch')\n",
    "plt.ylabel('loss')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 283
    },
    "id": "CpmK9lXQcBe9",
    "outputId": "dbbe0523-1ac8-47ab-b655-60ae9de9ea08"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['accuracy'])\n",
    "plt.xlabel('epoch')\n",
    "plt.ylabel('accuracy')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "id": "5cJlWITXqJWr"
   },
   "outputs": [],
   "source": [
    "model = model_fn()\n",
    "model.compile(loss='sparse_categorical_crossentropy', metrics='accuracy')\n",
    "\n",
    "history = model.fit(train_scaled, train_target, epochs=20, verbose=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 279
    },
    "id": "YT87Fjo2qKPC",
    "outputId": "1b1831ef-8c0b-4c47-d842-d6e32eb7449c"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['loss'])\n",
    "plt.xlabel('epoch')\n",
    "plt.ylabel('loss')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "bGqf6ceRr3zO"
   },
   "source": [
    "## 검증 손실"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "id": "4_iHvMxwu2D2"
   },
   "outputs": [],
   "source": [
    "model = model_fn()\n",
    "model.compile(loss='sparse_categorical_crossentropy', metrics='accuracy')\n",
    "\n",
    "history = model.fit(train_scaled, train_target, epochs=20, verbose=0, \n",
    "                    validation_data=(val_scaled, val_target))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "nJ5RGEmLu5KI",
    "outputId": "2a59e1ef-9850-4fe1-db3b-a8eaf7d14876"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dict_keys(['loss', 'accuracy', 'val_loss', 'val_accuracy'])\n"
     ]
    }
   ],
   "source": [
    "print(history.history.keys())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 279
    },
    "id": "zcpm7CpXu5vC",
    "outputId": "0aa51ffc-0966-44d4-e513-0d7db84c87ce"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.xlabel('epoch')\n",
    "plt.ylabel('loss')\n",
    "plt.legend(['train', 'val'])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "id": "qC8gDwo3qcJv"
   },
   "outputs": [],
   "source": [
    "model = model_fn()\n",
    "model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', \n",
    "              metrics='accuracy')\n",
    "\n",
    "history = model.fit(train_scaled, train_target, epochs=20, verbose=0, \n",
    "                    validation_data=(val_scaled, val_target))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 279
    },
    "id": "k8wWnyFzsLKb",
    "outputId": "140306b0-3e43-4178-8941-79ec92023510"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.xlabel('epoch')\n",
    "plt.ylabel('loss')\n",
    "plt.legend(['train', 'val'])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "F-dFY8lrYXm3"
   },
   "source": [
    "## 드롭아웃"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "AppFtFKgsk--",
    "outputId": "64f0076d-94a0-4003-c907-e8e149a2458f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_4\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " flatten_4 (Flatten)         (None, 784)               0         \n",
      "                                                                 \n",
      " dense_8 (Dense)             (None, 100)               78500     \n",
      "                                                                 \n",
      " dropout (Dropout)           (None, 100)               0         \n",
      "                                                                 \n",
      " dense_9 (Dense)             (None, 10)                1010      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 79,510\n",
      "Trainable params: 79,510\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = model_fn(keras.layers.Dropout(0.3))\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "id": "_TSe7oM9v1lW"
   },
   "outputs": [],
   "source": [
    "model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', \n",
    "              metrics='accuracy')\n",
    "\n",
    "history = model.fit(train_scaled, train_target, epochs=20, verbose=0, \n",
    "                    validation_data=(val_scaled, val_target))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 279
    },
    "id": "Rj_syB_iv30l",
    "outputId": "71f26e56-f44f-44c9-ce4b-fc6b60b469ce"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.xlabel('epoch')\n",
    "plt.ylabel('loss')\n",
    "plt.legend(['train', 'val'])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "hQi91PCQorHr"
   },
   "source": [
    "## 모델 저장과 복원"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "id": "npsYdX3rv6Oa"
   },
   "outputs": [],
   "source": [
    "model = model_fn(keras.layers.Dropout(0.3))\n",
    "model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', \n",
    "              metrics='accuracy')\n",
    "\n",
    "history = model.fit(train_scaled, train_target, epochs=10, verbose=0, \n",
    "                    validation_data=(val_scaled, val_target))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "id": "hWVYzt0Y2FPm"
   },
   "outputs": [],
   "source": [
    "model.save_weights('model-weights.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "id": "NissHzcq3xbN"
   },
   "outputs": [],
   "source": [
    "model.save('model-whole.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "2vljkGGu3AUl",
    "outputId": "27ddb7a3-99c7-476d-c2b4-ccdb8432a981",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-rw-r--r-- 1 haesun haesun 4049416 Jan  8 04:48 best-cnn-model.h5\r\n",
      "-rw-r--r-- 1 haesun haesun  333448 Jan  8 04:48 model-weights.h5\r\n",
      "-rw-r--r-- 1 haesun haesun  982664 Jan  8 04:48 model-whole.h5\r\n"
     ]
    }
   ],
   "source": [
    "!ls -al *.h5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "id": "7-t6gC5Z3GCM"
   },
   "outputs": [],
   "source": [
    "model = model_fn(keras.layers.Dropout(0.3))\n",
    "\n",
    "model.load_weights('model-weights.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "pHV9tBnzi8St",
    "outputId": "7da9201e-32c9-4802-dad4-9801c4cb50fc"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8804166666666666\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "\n",
    "val_labels = np.argmax(model.predict(val_scaled), axis=-1)\n",
    "print(np.mean(val_labels == val_target))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "sRM3Vpki4QyH",
    "outputId": "69127611-2c5f-4452-b716-fb4a8de59c4c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "375/375 [==============================] - 1s 2ms/step - loss: 0.3281 - accuracy: 0.8804\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.32808488607406616, 0.8804166913032532]"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = keras.models.load_model('model-whole.h5')\n",
    "\n",
    "model.evaluate(val_scaled, val_target)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "4NTCF3YD3EyA"
   },
   "source": [
    "## 콜백"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "L2lKN_934VB4",
    "outputId": "29d7e4a2-5363-4e8d-b700-a531bb4759df"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7f5290f975c0>"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = model_fn(keras.layers.Dropout(0.3))\n",
    "model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', \n",
    "              metrics='accuracy')\n",
    "\n",
    "checkpoint_cb = keras.callbacks.ModelCheckpoint('best-model.h5', \n",
    "                                                save_best_only=True)\n",
    "\n",
    "model.fit(train_scaled, train_target, epochs=20, verbose=0, \n",
    "          validation_data=(val_scaled, val_target),\n",
    "          callbacks=[checkpoint_cb])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "qYISeH6U5oh9",
    "outputId": "1f4391a4-6dab-403f-aa44-a00cb8b6dba8"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "375/375 [==============================] - 1s 3ms/step - loss: 0.3149 - accuracy: 0.8888\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.31489625573158264, 0.8888333439826965]"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = keras.models.load_model('best-model.h5')\n",
    "\n",
    "model.evaluate(val_scaled, val_target)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "id": "HLLlkR0s5Nd8"
   },
   "outputs": [],
   "source": [
    "model = model_fn(keras.layers.Dropout(0.3))\n",
    "model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', \n",
    "              metrics='accuracy')\n",
    "\n",
    "checkpoint_cb = keras.callbacks.ModelCheckpoint('best-model.h5', \n",
    "                                                save_best_only=True)\n",
    "early_stopping_cb = keras.callbacks.EarlyStopping(patience=2,\n",
    "                                                  restore_best_weights=True)\n",
    "\n",
    "history = model.fit(train_scaled, train_target, epochs=20, verbose=0, \n",
    "                    validation_data=(val_scaled, val_target),\n",
    "                    callbacks=[checkpoint_cb, early_stopping_cb])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "b6TazMcDxqXD",
    "outputId": "3f559a9c-3ba2-478a-b607-d8ac6ea4d55a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8\n"
     ]
    }
   ],
   "source": [
    "print(early_stopping_cb.stopped_epoch)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 279
    },
    "id": "QrUNYGPB6Kq7",
    "outputId": "13ccaee5-3d8b-421b-8636-90cf23dbacba"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.xlabel('epoch')\n",
    "plt.ylabel('loss')\n",
    "plt.legend(['train', 'val'])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "H-7y1qlg5yqO",
    "outputId": "f374810c-71af-4690-a90d-7571e734e62e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "375/375 [==============================] - 1s 2ms/step - loss: 0.3310 - accuracy: 0.8803\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.3309899568557739, 0.8803333044052124]"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate(val_scaled, val_target)"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "7-3 신경망 모델 훈련.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "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": 1
}