Week11_DNN.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Week11-DNN.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyMPOdpcDqaw1P5TtxTWTzEl"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gCSS8fIzalMm",
        "outputId": "08f251eb-0c0e-4582-b305-5aa494fce391"
      },
      "source": [
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "print(tf.__version__)\n",
        "print(keras.__version__)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "2.4.1\n",
            "2.4.0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7JYaswd9a4sP",
        "outputId": "c077ec4e-912a-49e6-f5c6-3e255b2e3899"
      },
      "source": [
        "fashion_mnist = keras.datasets.fashion_mnist\n",
        "(X_train_full,y_train_full), (X_test,y_test)= fashion_mnist.load_data()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-labels-idx1-ubyte.gz\n",
            "32768/29515 [=================================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-images-idx3-ubyte.gz\n",
            "26427392/26421880 [==============================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-labels-idx1-ubyte.gz\n",
            "8192/5148 [===============================================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-images-idx3-ubyte.gz\n",
            "4423680/4422102 [==============================] - 0s 0us/step\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Buq1rpfTbUPo",
        "outputId": "b381fc6a-0533-4362-c7f0-62209fcdf398"
      },
      "source": [
        "print(X_train_full.shape)\n",
        "print(X_test.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(60000, 28, 28)\n",
            "(10000, 28, 28)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uGluHTY9bbPG"
      },
      "source": [
        "# 훈련 세트의 일부분을 검증 세트로 만들고 입력 특성의 스케일을 0~1 사이 범위로 조정하기\n",
        "X_valid, X_train = X_train_full[:5000]/255.0 , X_train_full[5000:]/255.0\n",
        "y_valid, y_train = y_train_full[:5000], y_train_full[5000:]\n",
        "X_test = X_test/255.0"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "otdsrxfsbxn2"
      },
      "source": [
        "# 클래스 이름\n",
        "class_names = ['T-shirt/top','Trouser','Pullover','Dress','Coat','Sandal','Shirt','Sneaker','Bag','Ankle boot']"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "97BEBakNcMBd",
        "outputId": "1e107099-108d-4f7c-c9db-93b839b6fa95"
      },
      "source": [
        "# MLP model 만들기\n",
        "model = keras.models.Sequential([keras.layers.Flatten(input_shape=[28,28], name='flatten'),\n",
        "                                 keras.layers.Dense(300,activation='relu', name='hidden1'),\n",
        "                                 keras.layers.Dense(100, activation='relu', name='hidden2'),\n",
        "                                 keras.layers.Dense(10, activation='softmax', name='output')])\n",
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "flatten (Flatten)            (None, 784)               0         \n",
            "_________________________________________________________________\n",
            "hidden1 (Dense)              (None, 300)               235500    \n",
            "_________________________________________________________________\n",
            "hidden2 (Dense)              (None, 100)               30100     \n",
            "_________________________________________________________________\n",
            "output (Dense)               (None, 10)                1010      \n",
            "=================================================================\n",
            "Total params: 266,610\n",
            "Trainable params: 266,610\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 466
        },
        "id": "7gvAaKKGc2_y",
        "outputId": "9ed5ca3a-c37e-4c09-ff55-5d4a89db7236"
      },
      "source": [
        "# 모델 구조 시각화\n",
        "keras.utils.plot_model(model)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0ud8R0a7c_d3",
        "outputId": "fe8521e0-f78f-4da5-e35f-2aa1684a2ebe"
      },
      "source": [
        "# 모델 컴파일\n",
        "model.compile(loss='sparse_categorical_crossentropy', optimizer='sgd', metrics=['accuracy'])\n",
        "history= model.fit(X_train, y_train, epochs=30, validation_data=(X_valid,y_valid))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.9756 - accuracy: 0.6852 - val_loss: 0.5033 - val_accuracy: 0.8340\n",
            "Epoch 2/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4989 - accuracy: 0.8272 - val_loss: 0.4487 - val_accuracy: 0.8468\n",
            "Epoch 3/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4458 - accuracy: 0.8435 - val_loss: 0.4691 - val_accuracy: 0.8298\n",
            "Epoch 4/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4190 - accuracy: 0.8523 - val_loss: 0.3954 - val_accuracy: 0.8630\n",
            "Epoch 5/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3959 - accuracy: 0.8604 - val_loss: 0.3835 - val_accuracy: 0.8690\n",
            "Epoch 6/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3797 - accuracy: 0.8677 - val_loss: 0.3735 - val_accuracy: 0.8718\n",
            "Epoch 7/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3643 - accuracy: 0.8703 - val_loss: 0.3712 - val_accuracy: 0.8712\n",
            "Epoch 8/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3606 - accuracy: 0.8709 - val_loss: 0.3523 - val_accuracy: 0.8760\n",
            "Epoch 9/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3408 - accuracy: 0.8805 - val_loss: 0.3497 - val_accuracy: 0.8776\n",
            "Epoch 10/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3391 - accuracy: 0.8790 - val_loss: 0.3452 - val_accuracy: 0.8802\n",
            "Epoch 11/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3274 - accuracy: 0.8826 - val_loss: 0.3407 - val_accuracy: 0.8804\n",
            "Epoch 12/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3288 - accuracy: 0.8797 - val_loss: 0.3397 - val_accuracy: 0.8802\n",
            "Epoch 13/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3093 - accuracy: 0.8899 - val_loss: 0.3396 - val_accuracy: 0.8778\n",
            "Epoch 14/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3039 - accuracy: 0.8907 - val_loss: 0.3373 - val_accuracy: 0.8776\n",
            "Epoch 15/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3035 - accuracy: 0.8902 - val_loss: 0.3275 - val_accuracy: 0.8796\n",
            "Epoch 16/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.2938 - accuracy: 0.8944 - val_loss: 0.3520 - val_accuracy: 0.8750\n",
            "Epoch 17/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.2869 - accuracy: 0.8970 - val_loss: 0.3411 - val_accuracy: 0.8800\n",
            "Epoch 18/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.2808 - accuracy: 0.8977 - val_loss: 0.3162 - val_accuracy: 0.8816\n",
            "Epoch 19/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.2800 - accuracy: 0.8996 - val_loss: 0.3094 - val_accuracy: 0.8886\n",
            "Epoch 20/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.2768 - accuracy: 0.8998 - val_loss: 0.3104 - val_accuracy: 0.8890\n",
            "Epoch 21/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.2685 - accuracy: 0.9042 - val_loss: 0.3112 - val_accuracy: 0.8892\n",
            "Epoch 22/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.2603 - accuracy: 0.9048 - val_loss: 0.3116 - val_accuracy: 0.8902\n",
            "Epoch 23/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.2566 - accuracy: 0.9072 - val_loss: 0.3116 - val_accuracy: 0.8926\n",
            "Epoch 24/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.2508 - accuracy: 0.9095 - val_loss: 0.3318 - val_accuracy: 0.8818\n",
            "Epoch 25/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.2543 - accuracy: 0.9082 - val_loss: 0.3068 - val_accuracy: 0.8916\n",
            "Epoch 26/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.2450 - accuracy: 0.9115 - val_loss: 0.3039 - val_accuracy: 0.8960\n",
            "Epoch 27/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.2397 - accuracy: 0.9136 - val_loss: 0.3092 - val_accuracy: 0.8910\n",
            "Epoch 28/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.2351 - accuracy: 0.9155 - val_loss: 0.3306 - val_accuracy: 0.8832\n",
            "Epoch 29/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.2293 - accuracy: 0.9175 - val_loss: 0.3008 - val_accuracy: 0.8926\n",
            "Epoch 30/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.2268 - accuracy: 0.9202 - val_loss: 0.3060 - val_accuracy: 0.8914\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 324
        },
        "id": "75el5A76dfDT",
        "outputId": "cbe8400c-d1d6-4e95-8d98-091ce51a8179"
      },
      "source": [
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "pd.DataFrame(history.history).plot(figsize=(8,5))\n",
        "plt.grid(True)\n",
        "plt.gca().set_ylim(0,1)\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EUNm_iFseZC6",
        "outputId": "d757dbee-9afd-4423-9826-066d270448ae"
      },
      "source": [
        "# 검증이 끝나고, 테스트 세트로 모델을 평가하여 일반화 오차를 추정한다. \n",
        "model.evaluate(X_test,y_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "313/313 [==============================] - 0s 1ms/step - loss: 0.3309 - accuracy: 0.8836\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.3309383988380432, 0.8835999965667725]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bWn0y_kSedYJ",
        "outputId": "f96c490e-9f21-4102-8f4f-8b49b32c573e"
      },
      "source": [
        "# predict()를 이용해 새로운 샘플에 대해 예측을 만든다.\n",
        "# test 세트 처음 3개를 새로운 샘플로 간주\n",
        "import numpy as np\n",
        "X_new = X_test[:3]\n",
        "y_prob = model.predict(X_new)\n",
        "print(y_prob.round(2))\n",
        "y_pred = np.argmax(model.predict(X_new), axis=-1)\n",
        "np.array(class_names)[y_pred]\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[0.   0.   0.   0.   0.   0.01 0.   0.01 0.   0.99]\n",
            " [0.   0.   1.   0.   0.   0.   0.   0.   0.   0.  ]\n",
            " [0.   1.   0.   0.   0.   0.   0.   0.   0.   0.  ]]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array(['Ankle boot', 'Pullover', 'Trouser'], dtype='<U11')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 227
        },
        "id": "LUYy6ER1fQBW",
        "outputId": "089928e3-1893-45f0-c0b1-0449263d18d3"
      },
      "source": [
        "# 천유정 1813128\n",
        "plt.figure(figsize=(10,10))\n",
        "for img in range(3):\n",
        "  plt.subplot(1,3,img+1)\n",
        "  plt.grid(False)\n",
        "  plt.imshow(X_new[img])\n",
        "  plt.xlabel(class_names[y_pred[img]])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 3 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E5pTIYgnfiFH"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}