MNIST_Week2_2.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "MNIST_Week2_2",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/pseudope/Parrot_5th_Level1/blob/main/MNIST_Week2_2.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qWz40uYohBIN",
        "outputId": "af2cd29e-fd69-4fac-f0ae-44d3633c1f8f"
      },
      "source": [
        "#!/usr/bin/env python3\n",
        "!pip install einops\n",
        "#!pip install tensorflow --user\n",
        "\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import tensorflow as tf\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "from tensorflow.data import Dataset\n",
        "from tensorflow.keras.models import Model\n",
        "from tensorflow.keras.utils import to_categorical\n",
        "from tensorflow.keras.optimizers import Adam, RMSprop\n",
        "from tensorflow.keras.layers import Input, Dense, Conv2D, Dropout, Flatten, Activation, MaxPooling2D, BatchNormalization\n",
        "\n",
        "from einops import rearrange\n",
        "from sklearn.model_selection import train_test_split"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting einops\n",
            "  Downloading einops-0.3.2-py3-none-any.whl (25 kB)\n",
            "Installing collected packages: einops\n",
            "Successfully installed einops-0.3.2\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "S3WNN5P4hBIO",
        "outputId": "6526c3e0-c771-4d15-b7e2-ee84b1ef4859"
      },
      "source": [
        "(X_train, y_train), (X_test, y_test) = tf.keras.datasets.mnist.load_data()\n",
        "\n",
        "print(\"X_train shape: {}\".format(X_train.shape))\n",
        "print(\"X_test shape: {}\".format(X_test.shape))"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz\n",
            "11493376/11490434 [==============================] - 0s 0us/step\n",
            "11501568/11490434 [==============================] - 0s 0us/step\n",
            "X_train shape: (60000, 28, 28)\n",
            "X_test shape: (10000, 28, 28)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 657
        },
        "id": "KB832lechBIP",
        "outputId": "6737a873-21de-4a42-ab80-42797c249956"
      },
      "source": [
        "fig = plt.figure(figsize=(9, 9))\n",
        "\n",
        "for i in range(9):\n",
        "    plt.subplot(3, 3, 1 + i)\n",
        "    plt.tight_layout()\n",
        "    plt.imshow(X_train[i], cmap=\"gray\", interpolation=\"none\")\n",
        "    target = y_train[i]\n",
        "    plt.title(\"Ground Truth: {}\".format(int(target)))\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 648x648 with 9 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Nq1waxKDhBIP"
      },
      "source": [
        "target_size = 10   # 0 ~ 9, 총 9개\n",
        "\n",
        "### 주의: 흑백 데이터여도 channel 차원을 살려주어야 정상적으로 작동합니다\n",
        "X_train = rearrange(X_train, \"b h w -> b h w 1\")\n",
        "X_test = rearrange(X_test, \"b h w -> b h w 1\")\n",
        "\n",
        "X_train = X_train.astype(\"float32\") / (2 ** 8 - 1)   # 255로 scaling\n",
        "X_test = X_test.astype(\"float32\") / (2 ** 8 - 1)     # 255로 scaling\n",
        "\n",
        "y_train = to_categorical(y_train, target_size)       # 타겟을 정수로 변환\n",
        "y_test = to_categorical(y_test, target_size)         # 타겟을 정수로 변환"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7xA3xa64hBIP"
      },
      "source": [
        "X_train, X_val, y_train, y_val = train_test_split(X_train, y_train, test_size=0.2)"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SyT8hqophBIQ",
        "outputId": "063503e8-193b-4bce-a2c8-ce2f746f2b3e"
      },
      "source": [
        "print(X_train.shape)\n",
        "print(X_val.shape)\n",
        "print(X_test.shape)\n",
        "print(y_train.shape)\n",
        "print(y_val.shape)\n",
        "print(y_test.shape)"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(48000, 28, 28, 1)\n",
            "(12000, 28, 28, 1)\n",
            "(10000, 28, 28, 1)\n",
            "(48000, 10)\n",
            "(12000, 10)\n",
            "(10000, 10)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "paFz-UgbR85m"
      },
      "source": [
        "Changes Made:\n",
        "*   1st conv layer filter size 256, 128 -> 128 128\n",
        "*   Epochs increased to 20\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "haTxYAXahBIQ",
        "outputId": "49d0d723-b062-464c-801e-f90e4f271ec2"
      },
      "source": [
        "inputs = Input(shape=(28, 28, 1))\n",
        "\n",
        "X = Conv2D(filters=128, kernel_size=(3, 3), padding=\"Same\")(inputs)\n",
        "X = BatchNormalization()(X)\n",
        "X = Activation(\"relu\")(X)\n",
        "X = Conv2D(filters=128, kernel_size=(3, 3), padding=\"Same\", activation=\"relu\")(X)\n",
        "#X = Dropout(0.5)(X)\n",
        "X = MaxPooling2D(pool_size=(2, 2), strides=(2, 2))(X)\n",
        "\n",
        "X = Conv2D(filters=128, kernel_size=(3, 3), padding=\"Same\")(X)\n",
        "X = BatchNormalization()(X)\n",
        "X = Activation(\"relu\")(X)\n",
        "X = Conv2D(filters=64, kernel_size=(3, 3), padding=\"Same\", activation=\"relu\")(X)\n",
        "#X = Dropout(0.3)(X)\n",
        "X = MaxPooling2D(pool_size=(2, 2), strides=(2, 2))(X)\n",
        "\n",
        "X = Flatten()(X)\n",
        "X = Dense(256, activation=\"relu\")(X)\n",
        "X = BatchNormalization()(X)\n",
        "X = Dropout(0.5)(X)\n",
        "X = Dense(128, activation=\"relu\")(X)\n",
        "X = BatchNormalization()(X)\n",
        "X = Dropout(0.4)(X)\n",
        "X = Dense(64, activation=\"relu\")(X)\n",
        "X = BatchNormalization()(X)\n",
        "X = Dropout(0.2)(X)\n",
        "outputs = Dense(10, activation=\"softmax\")(X)\n",
        "\n",
        "model = Model(inputs=inputs, outputs=outputs)\n",
        "model.summary()"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_1 (InputLayer)         [(None, 28, 28, 1)]       0         \n",
            "_________________________________________________________________\n",
            "conv2d (Conv2D)              (None, 28, 28, 128)       1280      \n",
            "_________________________________________________________________\n",
            "batch_normalization (BatchNo (None, 28, 28, 128)       512       \n",
            "_________________________________________________________________\n",
            "activation (Activation)      (None, 28, 28, 128)       0         \n",
            "_________________________________________________________________\n",
            "conv2d_1 (Conv2D)            (None, 28, 28, 128)       147584    \n",
            "_________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D) (None, 14, 14, 128)       0         \n",
            "_________________________________________________________________\n",
            "conv2d_2 (Conv2D)            (None, 14, 14, 128)       147584    \n",
            "_________________________________________________________________\n",
            "batch_normalization_1 (Batch (None, 14, 14, 128)       512       \n",
            "_________________________________________________________________\n",
            "activation_1 (Activation)    (None, 14, 14, 128)       0         \n",
            "_________________________________________________________________\n",
            "conv2d_3 (Conv2D)            (None, 14, 14, 64)        73792     \n",
            "_________________________________________________________________\n",
            "max_pooling2d_1 (MaxPooling2 (None, 7, 7, 64)          0         \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 3136)              0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 256)               803072    \n",
            "_________________________________________________________________\n",
            "batch_normalization_2 (Batch (None, 256)               1024      \n",
            "_________________________________________________________________\n",
            "dropout (Dropout)            (None, 256)               0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 128)               32896     \n",
            "_________________________________________________________________\n",
            "batch_normalization_3 (Batch (None, 128)               512       \n",
            "_________________________________________________________________\n",
            "dropout_1 (Dropout)          (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 64)                8256      \n",
            "_________________________________________________________________\n",
            "batch_normalization_4 (Batch (None, 64)                256       \n",
            "_________________________________________________________________\n",
            "dropout_2 (Dropout)          (None, 64)                0         \n",
            "_________________________________________________________________\n",
            "dense_3 (Dense)              (None, 10)                650       \n",
            "=================================================================\n",
            "Total params: 1,217,930\n",
            "Trainable params: 1,216,522\n",
            "Non-trainable params: 1,408\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MqkQ72d1hBIS",
        "outputId": "f9c143ae-3235-47a1-caaf-dee52bd9697f"
      },
      "source": [
        "def scheduler(epoch, lr):\n",
        "  if epoch < 8:\n",
        "    return lr\n",
        "  else:\n",
        "    return lr * tf.math.exp(-0.1)\n",
        "\n",
        "callback = [\n",
        "            tf.keras.callbacks.LearningRateScheduler(scheduler),\n",
        "            tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=5)\n",
        "]\n",
        "\n",
        "model.compile(optimizer='adam', loss=\"categorical_crossentropy\", metrics=[\"accuracy\"])\n",
        "history = model.fit(X_train, y_train, batch_size=256, epochs=20, callbacks = callback, validation_data=(X_val, y_val), verbose=1)"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "188/188 [==============================] - 54s 115ms/step - loss: 0.3551 - accuracy: 0.8975 - val_loss: 6.9666 - val_accuracy: 0.1101\n",
            "Epoch 2/20\n",
            "188/188 [==============================] - 20s 105ms/step - loss: 0.0839 - accuracy: 0.9767 - val_loss: 3.4366 - val_accuracy: 0.1840\n",
            "Epoch 3/20\n",
            "188/188 [==============================] - 20s 106ms/step - loss: 0.0582 - accuracy: 0.9841 - val_loss: 0.0826 - val_accuracy: 0.9749\n",
            "Epoch 4/20\n",
            "188/188 [==============================] - 21s 113ms/step - loss: 0.0449 - accuracy: 0.9869 - val_loss: 0.0363 - val_accuracy: 0.9901\n",
            "Epoch 5/20\n",
            "188/188 [==============================] - 20s 107ms/step - loss: 0.0392 - accuracy: 0.9882 - val_loss: 0.0462 - val_accuracy: 0.9878\n",
            "Epoch 6/20\n",
            "188/188 [==============================] - 20s 107ms/step - loss: 0.0302 - accuracy: 0.9914 - val_loss: 0.0516 - val_accuracy: 0.9838\n",
            "Epoch 7/20\n",
            "188/188 [==============================] - 20s 108ms/step - loss: 0.0279 - accuracy: 0.9915 - val_loss: 0.0273 - val_accuracy: 0.9927\n",
            "Epoch 8/20\n",
            "188/188 [==============================] - 20s 107ms/step - loss: 0.0241 - accuracy: 0.9926 - val_loss: 0.0381 - val_accuracy: 0.9893\n",
            "Epoch 9/20\n",
            "188/188 [==============================] - 20s 107ms/step - loss: 0.0203 - accuracy: 0.9942 - val_loss: 0.1021 - val_accuracy: 0.9728\n",
            "Epoch 10/20\n",
            "188/188 [==============================] - 20s 107ms/step - loss: 0.0165 - accuracy: 0.9945 - val_loss: 0.0488 - val_accuracy: 0.9883\n",
            "Epoch 11/20\n",
            "188/188 [==============================] - 20s 107ms/step - loss: 0.0133 - accuracy: 0.9961 - val_loss: 0.0306 - val_accuracy: 0.9923\n",
            "Epoch 12/20\n",
            "188/188 [==============================] - 20s 107ms/step - loss: 0.0115 - accuracy: 0.9964 - val_loss: 0.0328 - val_accuracy: 0.9919\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yPRsfabXhBIS",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8ed6e26c-99af-440e-ac5e-1d50c8591364"
      },
      "source": [
        "model.evaluate(X_test, y_test, batch_size=256)"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "40/40 [==============================] - 2s 37ms/step - loss: 0.0234 - accuracy: 0.9943\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.02337127923965454, 0.9943000078201294]"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mZ49DBA3hBIT"
      },
      "source": [
        "def acc_loss_plot(hist):\n",
        "    \n",
        "    fit, loss_ax = plt.subplots()\n",
        "    acc_ax = loss_ax.twinx()\n",
        "    \n",
        "    loss_ax.plot(hist.history[\"loss\"], \"y\", label=\"train_loss\")\n",
        "    loss_ax.plot(hist.history[\"val_loss\"], \"r\", label=\"val_loss\")\n",
        "    loss_ax.set_xlabel(\"Epoch\")\n",
        "    loss_ax.set_ylabel(\"Loss\")\n",
        "    loss_ax.legend(loc = \"upper left\")\n",
        "    \n",
        "    acc_ax.plot(hist.history[\"accuracy\"], \"b\", label=\"train_acc\")\n",
        "    acc_ax.plot(hist.history[\"val_accuracy\"], \"g\", label=\"val_acc\")\n",
        "    acc_ax.set_ylabel(\"Accuracy\")\n",
        "    acc_ax.legend(loc = \"upper right\")\n",
        "    \n",
        "    plt.show()"
      ],
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qVs8gpOZhBIT",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "outputId": "c9b6b845-454a-4fae-a0a5-264cae04424e"
      },
      "source": [
        "acc_loss_plot(history)"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kFD0TsyM9t24"
      },
      "source": [
        ""
      ],
      "execution_count": 11,
      "outputs": []
    }
  ]
}