legacy_materials/2_dnn_with_mnist.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "e93f4b1e",
   "metadata": {},
   "source": [
    "# DL study second week  \n",
    "### MNIST dataset으로 MLP (DNN) 모델 구현"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "73e43667",
   "metadata": {},
   "outputs": [],
   "source": [
    "# 필요 라이브러리 및 모듈 로드\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import tensorflow as tf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "2197ccf6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# mnist dataset 불러오기\n",
    "mnist = tf.keras.datasets.mnist\n",
    "\n",
    "(X_train, y_train), (X_test, y_test) = mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "7f45cbea",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The dimension of train data : (60000, 28, 28) (60000,)\n",
      "The dimension of test data : (10000, 28, 28) (10000,)\n"
     ]
    }
   ],
   "source": [
    "# dataset의 차원 확인하기\n",
    "print('The dimension of train data :', X_train.shape, y_train.shape)\n",
    "print('The dimension of test data :', X_test.shape, y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9118498c",
   "metadata": {},
   "source": [
    "- 위의 출력을 살펴보자  \n",
    "- train data는 3차원 텐서 (60000, 28, 28), train label data는 벡터 형태 (60000,)의 dimension을 가짐  \n",
    "- 즉, 28x28 사이즈의 흑백 이미지가 60000장이라는 의미!  \n",
    "- label은 각 원소가 0~9의 값을 가지는 길이 60000의 벡터라는 의미!  \n",
    "- test의 경우도 마찬가지  "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "84172a36",
   "metadata": {},
   "source": [
    "### One-Hot Encoding이란?  \n",
    "- 위의 label들을 출력해보면 0~9 사이의 정수 값으로 이루어져 있는 벡터 형태를 가짐  \n",
    "- loss function을 cross entropy loss를 사용하기 위해서 이들을 원-핫 인코딩을 진행해야 함!    \n",
    "- 그림으로 살펴보자  \n",
    "  \n",
    "<img src = \"https://i.ytimg.com/vi/BecEHOVmx9o/maxresdefault.jpg\" width=700>  \n",
    "\n",
    "\n",
    "- 위와 같이 정수를 클래스 갯수 만큼의 n차원 벡터 형태로 바꿔주며 벡터의 원소들은 0 or 1로 구성  \n",
    "- 또한 1은 무조건 한 개만 존재해야 하며 나머지는 0으로 채워야 함!  \n",
    "- tensorflow에는 정수 라벨을 원-핫 인코딩 형태로 변환해주는 함수가 존재함!  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "b4042cb9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "원-핫 인코딩 변환 전 train label 값:\n",
      " [5 0 4]\n",
      "\n",
      "원-핫 인코딩 변환 전 test label 값:\n",
      " [7 2 1]\n",
      "\n",
      "원-핫 인코딩 변환 후 train label 값\n",
      " [[0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n",
      " [1. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]]\n",
      "\n",
      "원-핫 인코딩 변환 후 test label 값\n",
      " [[0. 0. 0. 0. 0. 0. 0. 1. 0. 0.]\n",
      " [0. 0. 1. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]]\n"
     ]
    }
   ],
   "source": [
    "# 원 핫 인코딩 변환 전 값 살펴보기\n",
    "print('원-핫 인코딩 변환 전 train label 값:\\n', y_train[:3])\n",
    "print('\\n원-핫 인코딩 변환 전 test label 값:\\n', y_test[:3])\n",
    "\n",
    "# one-hot encoding 변환 \n",
    "y_train = tf.keras.utils.to_categorical(y_train)\n",
    "y_test = tf.keras.utils.to_categorical(y_test)\n",
    "\n",
    "# 원 핫 인코딩 변환 후 값 살펴보기\n",
    "print('\\n원-핫 인코딩 변환 후 train label 값\\n', y_train[:3,:])\n",
    "print('\\n원-핫 인코딩 변환 후 test label 값\\n', y_test[:3,:])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a2f111a2",
   "metadata": {},
   "source": [
    "### scaling이란?  \n",
    "- 이미지에서 각 픽셀은 0~255의 값을 가짐  \n",
    "- dataset을 255로 나누게 되면 0~1사이의 값을 가지게 됨!  \n",
    "- 0~1 사이의 값을 가져아 하는 이유는 다음과 같음     \n",
    "    1. 큰 값을 가지는 변수에 대해서 편향되면 안되기 때문에   \n",
    "    2. 학습 과정에서 데이터들의 분포가 서로 다르다면 학습에 어려움이 있음    \n",
    "    \n",
    "- 또 다른 scaling 방법으로 standard scaling 등이 있으나 neural network에서는 일반적으로 normalization을 수행  \n",
    "- neural network에서는 input data의 scaling이 필수적!  \n",
    "- 그러나 label은 스케일링할 필요가 없다는 것을 주의!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "39ddcbb8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The maximum value of X_train : 255 , The minimum value of X_train : 0\n",
      "The maximum value of X_train : 255 , The minimum value of X_test : 0\n"
     ]
    }
   ],
   "source": [
    "print('The maximum value of X_train :', X_train.max(), ', The minimum value of X_train :', X_train.min())\n",
    "print('The maximum value of X_train :', X_test.max(), ', The minimum value of X_test :', X_test.min())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "c08133be",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The maximum value of X_train : 1.0 , The minimum value of X_train : 0.0\n",
      "The maximum value of X_train : 1.0 , The minimum value of X_test : 0.0\n"
     ]
    }
   ],
   "source": [
    "# normalization\n",
    "X_train = X_train/255.0\n",
    "X_test = X_test/255.0\n",
    "\n",
    "print('The maximum value of X_train :', X_train.max(), ', The minimum value of X_train :', X_train.min())\n",
    "print('The maximum value of X_train :', X_test.max(), ', The minimum value of X_test :', X_test.min())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "ab7bac48",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 25 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# check sample images\n",
    "def show_image(nrow, ncol, images):\n",
    "    plt.figure(figsize=(20,20))\n",
    "    for i in range(nrow*ncol):\n",
    "        plt.subplot(nrow, ncol, i+1)\n",
    "        plt.imshow(images[i], cmap='gray')\n",
    "    plt.show()\n",
    "    \n",
    "show_image(5, 5, X_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2728cac4",
   "metadata": {},
   "source": [
    "### image channel이란?  \n",
    "- MNIST handwritten dataset은 grayscale(흑백) 이미지임  \n",
    "- 즉, grayscale은 $m\\times n\\times 1$의 사이즈를 가짐  \n",
    "- 일반적으로 우리가 가지고 있는 컬러 사진들은 RGB 3 channel을 가짐  \n",
    "- 즉, RGB는 $m\\times n\\times 3$의 사이즈를 가짐"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f47d7ec7",
   "metadata": {},
   "source": [
    "### Deep Neural Network Model 구축"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "46b3bc14",
   "metadata": {},
   "source": [
    "#### DNN의 기본 형태\n",
    "<img src = \"https://upload.wikimedia.org/wikipedia/commons/thumb/4/46/Colored_neural_network.svg/800px-Colored_neural_network.svg.png\" width=500>  \n",
    "\n",
    "- 첫 번째 layer를 input layer라고 정의  \n",
    "- 마지막 layer를 output layer라고 정의  \n",
    "- 첫 번째, 마지막 layer를 제외한 나머지 layer를 hidden layer라고 정의  \n",
    "- DNN에서는 첫 번째 layer의 input node 총 갯수는 이미지의 총 픽셀 갯수를 의미  \n",
    "- 즉, MNIST handwritten dataset 기준으로 $28\\times 28 = 784$가 input node 수임\n",
    "- classification task에서 마지막 layer의 node 수는 분류하고자 하는 클래스 갯수를 넣어야 함  \n",
    "- 즉, MNIST handwritten dataset 기준으로 0,1,2,...,7,8,9 총 10개 클래스를 넣어야 함  \n",
    "- 종합적으로 말해, input layer와 output layer의 노드 수는 우리가 수행하고자 하는 task에 맞춰 설정해야함!  \n",
    "- 이외의 hidden layer의 노드 수나 layer 수는 상관없이 엔지니어 편의에 맞춰 설정  \n",
    "- 그럼 이 점을 참고하여 모델을 구축해보자  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "02d02c6a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# DNN model 구축\n",
    "model = tf.keras.models.Sequential([\n",
    "    ################# Input Layer #################\n",
    "    tf.keras.layers.Flatten(input_shape=(28,28)),\n",
    "    ################# First Hidden Layer #################\n",
    "    tf.keras.layers.Dense(100, activation='relu'),\n",
    "    ################# Second hidden layer #################\n",
    "    tf.keras.layers.Dense(100, activation='relu'),\n",
    "    ################# Output Layer #################\n",
    "    tf.keras.layers.Dense(10, activation='softmax')\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "2989747e",
   "metadata": {},
   "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, 100)               10100     \n",
      "                                                                 \n",
      " dense_2 (Dense)             (None, 10)                1010      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 89,610\n",
      "Trainable params: 89,610\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# 모델 구조 살펴보기\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "58f1f04a",
   "metadata": {},
   "source": [
    "### activation function이란?  \n",
    "- activation function은 이전 layer의 output을 다음 layer의 노드로 신호를 전달하는 역할!  \n",
    "- 만약 activation function이 linear하게 구성된다면? 즉, activation이 없다면?  \n",
    "    - activation function이 $h\\left(x\\right)=cx$인 예를 들어보자  \n",
    "    - 그러면 $y=h\\left(h\\left(h\\left(x\\right)\\right)\\right) = c\\times c\\times c\\times x = c^3x$이다  \n",
    "    - 즉, $y=ax$에서 $a=c^3$인 선형 함수를 의미!  \n",
    "    - 이는 hidden layer가 없는 네트워크와 다를 바가 없음!  \n",
    "    - 그래서 activation function은 nonlinear한 함수로 구성해야 함  \n",
    "    - 예를 들면 ReLU, sigmoid, tanh 등이 존재  \n",
    "    - 일반적으로 activation function은 ReLU를 사용함!  \n",
    "    $$ReLU\\,\\, function\\, : ReLU\\left(x\\right) = max\\left(0, x\\right)$$\n",
    "    - why? 가장 빠른 연산 속도를 가지고 있다는 장점이 있음  \n",
    "    \n",
    "- 마지막 layer의 activation function은 무엇을 사용해야 할까?  \n",
    "    - classification task에서는 일반적으로 output layer의 activation function으로 softmax를 사용!  \n",
    "    - softmax function의 식을 살펴보자  \n",
    "    $$softmax\\,\\, function : f_i\\left(x\\right) = \\frac{\\exp\\left(x_i\\right)}{\\sum_j\\exp\\left(x_j\\right)}$$  \n",
    "    - i 번째 class에서 softmax function에서 나오는 모든 값들의 합은 1이어야 함!  \n",
    "    - softmax를 사용할 경우 모든 클래스에 대해 0~1 사이의 확률 값을 출력시켜줌!  \n",
    "    - 즉, 가장 큰 확률의 인덱스가 해당 클래스로 분류 된다는 의미!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "7b6954ed",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "input: 20, output: 20\n",
      "input: -20, output: 0\n"
     ]
    }
   ],
   "source": [
    "# relu 함수 구현\n",
    "def relu(inputs):\n",
    "    out = max(0, inputs)\n",
    "    return out\n",
    "\n",
    "a = 20\n",
    "b = -20\n",
    "print(f'input: {a}, output: {relu(a)}')\n",
    "print(f'input: {b}, output: {relu(b)}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "41358dae",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "softmax 입력하기 전의 값: [10 20 30]\n",
      "softmax 입력한 후의 값: [2.06106005e-09 4.53978686e-05 9.99954600e-01]\n",
      "softmax 출력값을 모두 더하면 1인지 확인: 1.0\n"
     ]
    }
   ],
   "source": [
    "# softmax 함수에 입력하기 전과 후의 값을 비교해보자\n",
    "\n",
    "# 아무렇게나 벡터 생성\n",
    "a = np.array([10,20,30])\n",
    "\n",
    "# softmax function 구현\n",
    "def softmax(inputs):\n",
    "    exp = np.exp(inputs)\n",
    "    sum_exp = np.sum(exp)\n",
    "    out = exp/sum_exp\n",
    "    return out\n",
    "\n",
    "print('softmax 입력하기 전의 값:', a)\n",
    "print('softmax 입력한 후의 값:', softmax(a))\n",
    "print('softmax 출력값을 모두 더하면 1인지 확인:', softmax(a).sum())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7db79213",
   "metadata": {},
   "source": [
    "### 활성화 함수 더 쉽게 이해하기!  \n",
    "- google에 tensorflow playground 검색  \n",
    "- 여기에서 다양한 데이터 분포와 다양한 활성화 함수를 적용해 훈련시켜보고   \n",
    "- 활성화 함수마다 어떤 변화가 생기는지, 어떤 활성화 함수가 더 빠르게 데이터에 피팅하는지 살펴보자  "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "89e50d47",
   "metadata": {},
   "source": [
    "### loss function, optimizer, learning rate, metrics가 무엇일까?  \n",
    "- loss function이란?  \n",
    "    - predicted value와 label의 차이를 구하는 기준  \n",
    "    - 즉, 딥러닝 모델 학습에 필수!  \n",
    "    - 손실 함수로는 RMSE, MAE, MSE 등 많은 함수가 존재함  \n",
    "    - 그러나 위의 손실 함수들은 regression task에 적합한 함수들임  \n",
    "    - classificaion task에서는 일반적으로 cross entropy loss라는 손실 함수를 사용  \n",
    "    \n",
    "- optimizer란?  \n",
    "    - optimizer는 가장 작은 loss의 값을 얻을 수 있도록 파라미터들을 업데이트 하는 역할을 함!  \n",
    "    - 즉, predicted가 label과 최대한 비슷하게 출력되도록 해주는 역할임   \n",
    "    - 이도 딥러닝 모델 학습에 필수임!  \n",
    "    - optimizer의 종류로 SGD (Stochastic Gradient Descent), momentum, adagrad, RMSProp, Adam 등이 존재  \n",
    "    \n",
    "- learning rate란?  \n",
    "    - optimizer의 최적화 속도에 영향을 미치는 hyperparameter  \n",
    "    - optimizer에서 learning rate는 gradient에 learning rate를 곱해 다음 지점을 결정하게 됨  \n",
    "    \n",
    "<img src = \"https://blog.kakaocdn.net/dn/cH1K5u/btrorlNWlbS/bIzqxRHPRo4u4rXh56QGaK/img.png\">\n",
    "    \n",
    "- metrics이란?  \n",
    "    - metrics는 model이 train data에 대해 얼마나 잘 피팅되었는지 평가하는 기준  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "6bef215f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# 학습하기 전 모델 compile하기\n",
    "optimizer = tf.keras.optimizers.SGD(learning_rate=0.01)\n",
    "model.compile(optimizer=optimizer, \n",
    "              loss='categorical_crossentropy',\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "3cdb3fcd",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "1875/1875 [==============================] - 2s 795us/step - loss: 0.6547 - accuracy: 0.82590s - los\n",
      "Epoch 2/10\n",
      "1875/1875 [==============================] - 1s 770us/step - loss: 0.3071 - accuracy: 0.9120\n",
      "Epoch 3/10\n",
      "1875/1875 [==============================] - 1s 766us/step - loss: 0.2543 - accuracy: 0.9266\n",
      "Epoch 4/10\n",
      "1875/1875 [==============================] - 1s 767us/step - loss: 0.2185 - accuracy: 0.9369\n",
      "Epoch 5/10\n",
      "1875/1875 [==============================] - 1s 767us/step - loss: 0.1922 - accuracy: 0.9444\n",
      "Epoch 6/10\n",
      "1875/1875 [==============================] - 1s 775us/step - loss: 0.1714 - accuracy: 0.9506\n",
      "Epoch 7/10\n",
      "1875/1875 [==============================] - 1s 778us/step - loss: 0.1547 - accuracy: 0.9546\n",
      "Epoch 8/10\n",
      "1875/1875 [==============================] - 1s 759us/step - loss: 0.1413 - accuracy: 0.9591\n",
      "Epoch 9/10\n",
      "1875/1875 [==============================] - 1s 751us/step - loss: 0.1295 - accuracy: 0.9624\n",
      "Epoch 10/10\n",
      "1875/1875 [==============================] - 1s 747us/step - loss: 0.1198 - accuracy: 0.9657\n"
     ]
    }
   ],
   "source": [
    "# 모델 학습\n",
    "history = model.fit(X_train, y_train, epochs=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "21a6f159",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# loss 및 accuracy 그래프 살펴보기\n",
    "plt.figure(figsize=(10,7))\n",
    "plt.plot(history.history['loss'], label='loss')\n",
    "plt.plot(history.history['accuracy'], label='accuracy')\n",
    "plt.legend(loc='best')\n",
    "plt.show()\n",
    "# loss는 0에 가까울수록, 수렴할수록 잘 훈련되었다는 증거!\n",
    "# accuracy의 경우 1에 가까울수록 잘 훈련되었다는 증거!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "c0612476",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 - 0s - loss: 0.1286 - accuracy: 0.9635 - 257ms/epoch - 820us/step\n",
      "Error: 0.1285957545042038, Accuracy: 96.35%\n"
     ]
    }
   ],
   "source": [
    "# 모델 평가\n",
    "test_loss, test_acc = model.evaluate(X_test, y_test, verbose=2)\n",
    "print(f'Error: {test_loss}, Accuracy: {test_acc*100:.2f}%')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "3a917e0f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# test data에 대해 model prediction 진행\n",
    "prediction = model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "ae6ace2e",
   "metadata": {},
   "outputs": [],
   "source": [
    "# one-hot encoding label -> integer label\n",
    "prediction_label = np.argmax(prediction, axis=1)\n",
    "test_label = np.argmax(y_test, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "f87bef4c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# prediction_label과 test_label이 같지 않은 인덱스들 뽑기\n",
    "false_index = np.where(prediction_label!=test_label)[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "0456ee08",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1800 with 25 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 잘못 분류된 이미지와 라벨과 score 살펴보기\n",
    "plt.figure(figsize=(20,25))\n",
    "for i in range(25):\n",
    "    plt.subplot(5,5,i+1)\n",
    "    plt.imshow(X_test[false_index[i]], cmap='gray')\n",
    "    plt.title(f'score: {prediction[false_index[i]][prediction_label[false_index[i]]]:.2f}'\n",
    "              f'\\nlabel: {test_label[false_index[i]]},'\n",
    "              f' predicted: {prediction_label[false_index[i]]}\\n',\n",
    "              fontsize=15)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "674260a7",
   "metadata": {},
   "source": [
    "# Homework  \n",
    "- fashion mnist dataset으로 DNN model 학습시키고 결과 살펴보기  \n",
    "- 다양한 optimizer와 다양한 learning rate로 model compile한 후 비교해보기 (중요!)  \n",
    "- DNN model의 hidden layer 수와 노드 수를 다양하게 구성하고 모델 학습하기  \n",
    "- 다음 스터디까지 CNN(Convolutional Neural Network)에 대해 간단하게 살펴보기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fe1f9f5a",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}