ディープラーニング_mnist_Good.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "collapsed_sections": [
        "hnNpHgBNp73W",
        "MeuJ09R82rdq"
      ],
      "authorship_tag": "ABX9TyOKNgfjzj1q22qUI/Saer9J",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Ken-huea/Ken-huea/blob/main/%E3%83%87%E3%82%A3%E3%83%BC%E3%83%97%E3%83%A9%E3%83%BC%E3%83%8B%E3%83%B3%E3%82%B0_mnist_Good.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### mnistにつて。"
      ],
      "metadata": {
        "id": "OMxVAIGqf14_"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![2022-08-23_15h58_47.png](data:image/png;base64,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)"
      ],
      "metadata": {
        "id": "D7edff4Cfy8f"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "今回の作成物の目標\n",
        "- 私たちが0～9の数字が書かれた画像を提出する。　モデル(AI)が読み込み判断し、予測を返す。\n",
        "\n",
        "このような仕組みを作成しようと思います。"
      ],
      "metadata": {
        "id": "AQCScoDo54f5"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 実装　\n",
        "⇒一般的なディープラーニング。こちらを参照致しました。\n",
        "\n",
        "[概要はこちら](https://www.youtube.com/watch?v=L3qf-jMj_fI&list=PL1P-NJggk9lbilEP-IHapQRCdYlvX43yC&index=105&t=1124s)\n",
        "　\n",
        "[こちらも取り込んだ](https://www.youtube.com/watch?v=ZAWvHvOBGDU&list=PL1P-NJggk9lbilEP-IHapQRCdYlvX43yC&index=105)"
      ],
      "metadata": {
        "id": "a6A3vAN-l3Nf"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 準備\n",
        "モジュールインポート、データの取得"
      ],
      "metadata": {
        "id": "4qnmrSEbmaIe"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# -*- coding: utf-8 -*-\n",
        "import numpy as np\n",
        "from tensorflow.keras.models import Sequential, model_from_json\n",
        "from tensorflow.keras.layers import Dense, Dropout\n",
        "from tensorflow.keras.optimizers import RMSprop\n",
        "from tensorflow.keras.datasets import mnist\n",
        "from tensorflow.keras.utils import to_categorical\n",
        "import os\n",
        "import pickle"
      ],
      "metadata": {
        "id": "oks92jeHmXl0"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "入力画像のパラメータ 画像サイズの入力。　画像サイズを統一します。\n",
        "\n",
        "今回作成するモデルは28*28の画像に対応するようにします。　つまり、与えられた画像をこのスケールに変換する。」という事です。\n"
      ],
      "metadata": {
        "id": "cJ4xD-al5c4c"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "img_width = 28  # 入力画像の幅\n",
        "img_height = 28 # 入力画像の高さ\n",
        "\n",
        "# 入力データ数（今回は28*28=784個）\n",
        "num_input = int(img_width * img_height)\n",
        "\n",
        "# 分類クラス数(今回は0～9の手書き文字なので10)10パターン\n",
        "num_classes = 10       "
      ],
      "metadata": {
        "id": "6X7lVhsfmrBP"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### データの取得と整形"
      ],
      "metadata": {
        "id": "DkOFoNNQoAKz"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "①データ取得と分割"
      ],
      "metadata": {
        "id": "9fqW_UV3o5F8"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "データはmnistから取得します。取得時にデータ分割も同時にしておきましょう。　この分割方法はこのようなものと考えて頂ければ。\n",
        "\n",
        "そして分割されたデータのshape確認はクセで行う事をお勧めします。shapeが合ってないエラーは多発しますので。"
      ],
      "metadata": {
        "id": "HQk8UFMA6YNQ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# mnistデータセット（訓練用データと検証用データ）をネットから取得。そして訓練用とテスト用に分割。\n",
        "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
        "\n",
        "x_train.shape, y_train.shape,x_test.shape, y_test.shape\n",
        "print(\"訓練データの画像データ\",x_train.shape) # 28*28pxの60000枚。\n",
        "print(\"訓練データの正解ラベル\",y_train.shape) # 60000枚で１～９\n",
        "print(\"テストデータの画像データ\",x_test.shape) # 28*28pxの10000枚。\n",
        "print(\"テストデータの正解ラベル\",y_test.shape) # 10000枚で１～９"
      ],
      "metadata": {
        "id": "t6NndlHln9dk",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "167c604c-e305-4ed0-c0f8-f7139a910872"
      },
      "execution_count": null,
      "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",
            "訓練データの画像データ (60000, 28, 28)\n",
            "訓練データの正解ラベル (60000,)\n",
            "テストデータの画像データ (10000, 28, 28)\n",
            "テストデータの正解ラベル (10000,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "- ②　取得したデータを、2次元配列から1次元配列へ変換（今回は28*28=784個の要素数)：\n",
        "\n",
        "28＊28のピクセルをreshapeに784の数字の羅列に変換しました。こっちの方が処理しやすいので。\n",
        "\n",
        "- ③　そのデータ型をfloat32(浮動小数点型)に変換\n",
        "\n",
        "これらの処理はディープラーニングに組み込む上での必要な処理です。"
      ],
      "metadata": {
        "id": "9TTBZ8eWoQBQ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 2次元配列から1次元配列へ変換\n",
        "x_train = x_train.reshape(60000, num_input)\n",
        "x_test = x_test.reshape(10000, num_input)\n",
        "\n",
        "\n",
        "# データ型をfloat32(浮動小数点型)に変換\n",
        "x_train = x_train.astype('float32')\n",
        "x_test = x_test.astype('float32')"
      ],
      "metadata": {
        "id": "lOVi6mOXoS4Y"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# 処理後の結果が気になる方は実行して確認してみてください。\n",
        "print(x_train[0].shape)\n",
        "x_train[0]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mDhhVYz78lWm",
        "outputId": "6d3c8f77-0982-4128-abb1-0f35abb7ec07"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(784,)"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "④　正規化(0-255から0.0-1.0に変換）\n",
        "\n",
        "これも同様でディープラーニングに必要な処理。　今回は全てのデータに255を割っていますね。(正規化の方法は色々あります。)"
      ],
      "metadata": {
        "id": "fZaqcP93oqvg"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "x_train /= 255\n",
        "x_test /= 255"
      ],
      "metadata": {
        "id": "fdL-yNaDovtH"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# 処理後の結果が気になる方は実行して確認してみてください。 今回は全ての数字を0～１の範囲で抑えています。\n",
        "x_train[0]"
      ],
      "metadata": {
        "id": "NlThyc1W87Ke"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "⑤ one hot encording\n",
        "\n",
        "ディープラーニングにおいての鉄則ですね。答えのラベルをカテゴリカル変数、one hot encording致します。\n",
        "\n",
        "説明では分かりにくいですが、見ればわかるはずです。\n",
        "\n",
        "そしてその処理はkerasのモジュールを用いれば簡単。<font color=\"red\">**第一引数にデータ、第二引数に分類する数。**(今回は10。)</font>"
      ],
      "metadata": {
        "id": "vwJiVBlWpEeM"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "y_train = to_categorical(y_train)\n",
        "y_test = to_categorical(y_test)"
      ],
      "metadata": {
        "id": "pO0hQIwJpRwx"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# このように0から１～９までのカラムが生成されて、該当の正解ラベルに１、他は0になります。３ならば[0., 0., 3., ..., 0., 0., 0.] となります。\n",
        "y_train"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vObAIa7L_oY2",
        "outputId": "34068c23-6182-4e4b-84df-1052568c087e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0., 0., 0., ..., 0., 0., 0.],\n",
              "       [1., 0., 0., ..., 0., 0., 0.],\n",
              "       [0., 0., 0., ..., 0., 0., 0.],\n",
              "       ...,\n",
              "       [0., 0., 0., ..., 0., 0., 0.],\n",
              "       [0., 0., 0., ..., 0., 0., 0.],\n",
              "       [0., 0., 0., ..., 0., 1., 0.]], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# データセットの個数を表示 ⇒只の確認。\n",
        "print(x_train.shape[0], 'train samples')  # 60000枚の画像データ(1次元配列済み)\n",
        "print(x_test.shape[0], 'test samples')    # 60000枚の画像データ(1次元配列済み)\n",
        "print(x_test[0].shape)                    # 1次元配列になっていますね。"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UDSKHX8OpYO-",
        "outputId": "a6bfcac9-f0f8-4659-c135-3bdd58f6ecb6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "60000 train samples\n",
            "10000 test samples\n",
            "(784,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### モデルの構築\n",
        "今回は<font color=\"red\">ディープラーニング)</font>のモデルを設定"
      ],
      "metadata": {
        "id": "hnNpHgBNp73W"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "ニューロンを作成して、場合によってはドロップアウトも追加します。\n",
        "\n",
        "ドロップアウト：dropout_rate=0.2なら20%のユニットを無効化し、<font color=\"red\">過学習防止。 \n",
        "</font>"
      ],
      "metadata": {
        "id": "sCoEalIDrn1P"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# ハイパーパラメータの調整 ⇒適宜調整。\n",
        "\n",
        "num_middle_unit = 512  # 中間層のユニット数\n",
        "dropout_rate = 0.2     # 過学習防止用：入力の20%を0にする（破棄）"
      ],
      "metadata": {
        "id": "EJ3yqPvjmfBu"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# モデルの構築\n",
        "model = Sequential()"
      ],
      "metadata": {
        "id": "sFB1pk0MqBr9"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "①②は1行で一緒みたい。\n",
        "\n",
        "①層目の入力層:28×28=784個のユニット数(個数は変数num_inputに格納)\n",
        "\n",
        "②層目の中間層:ユニット数512(変数num_middle_unitに格納)、\n",
        "その値は自分で適宜調整し精度向上を目指す。\n",
        "\n",
        "活性化関数は<font color=\"red\">relu関数</font>"
      ],
      "metadata": {
        "id": "DwknE4Obqjv1"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(num_input)\n",
        "print(num_middle_unit)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wwTeVfJzGaIF",
        "outputId": "2b625322-e6d6-49a6-b053-654afb997580"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "784\n",
            "512\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "変数を用いる事で変更/調整を容易にしていますね。単純な構造では返って分かりにくいかもだけど、このような構築方法もアリという事か。"
      ],
      "metadata": {
        "id": "I0VLpCPTGvrt"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#①②とドロップアウト\n",
        "model.add(Dense(activation='relu', input_dim=num_input, units=num_middle_unit))\n",
        "\n",
        "model.add(Dropout(dropout_rate))"
      ],
      "metadata": {
        "id": "z_q4qt51qiNl"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "③層目の中間層:ユニット数512、活性化関数はrelu関数"
      ],
      "metadata": {
        "id": "aJM3jmCtsV0-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#③ とドロップアウト\n",
        "model.add(Dense(units=num_middle_unit, activation='relu'))\n",
        "\n",
        "model.add(Dropout(dropout_rate))"
      ],
      "metadata": {
        "id": "g-yjWGobsgwG"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "4層目、<font color=\"red\">出力層</font>:10分類（0から9まで）なので、ユニット数10, 分類問題なので<font color=\"red\">活性化関数はsoftmax関数</font>⇒それが定石。"
      ],
      "metadata": {
        "id": "eyRbkNShswZX"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model.add(Dense(num_classes, activation='softmax')) # 活性化関数：softmax\n",
        "#状態を確認\n",
        "model.summary()\n",
        "# model summaryでモデル構築を確認。全部512でいいのかな？"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "k2GUmXakssok",
        "outputId": "e0b664a4-ac57-4215-a29d-39b034484b4f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " dense (Dense)               (None, 512)               401920    \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 512)               0         \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 512)               262656    \n",
            "                                                                 \n",
            " dropout_1 (Dropout)         (None, 512)               0         \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 10)                5130      \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 669,706\n",
            "Trainable params: 669,706\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "数値の根拠は色々あるみたいですが、今回は教材通りに行います。是非試行錯誤をしてみましょう。"
      ],
      "metadata": {
        "id": "1ZW3nmue9vwg"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 学習"
      ],
      "metadata": {
        "id": "b46qnq1OynAb"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "コンパイルはどう学習するかを指定しています。\n",
        "- 'categorical_crossentropy':分類問題ならこれを選びましょう。\n",
        "- RMSprop: これは精度により推敲してみてください。\n",
        "- accuracy: 正確度です。分類問題ではメジャーな評価指標の一つです。"
      ],
      "metadata": {
        "id": "4e61VuPOHMDF"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# コンパイル（多クラス分類問題）\n",
        "model.compile(loss='categorical_crossentropy', optimizer=RMSprop(), metrics=['accuracy'])"
      ],
      "metadata": {
        "id": "V0crI3PUysUY"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "batch_size = 128       # バッチサイズ\n",
        "epochs = 20            # エポック数(学習の繰り返し回数)\n",
        "dropout_rate = 0.2     # 過学習防止用：入力の20%を0にする（破棄）"
      ],
      "metadata": {
        "id": "gkWEvonuy6oE"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "<font color=\"red\">バッチサイズ</font>は機械学習分野の慣習として、「2のn乗」（32, 64, 128, 256, 512）が使われることが多い。\n",
        "\n",
        "512はせっかちさんかな？って思っちゃうよ。\n",
        "\n",
        "<font color=\"red\">verbose</font> : これはログを出すかどうか。"
      ],
      "metadata": {
        "id": "gaEesuxyXUZZ"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "そして学習で一番大事なことは<font color=\"red\">過学習をしない事</font>。\n",
        "\n",
        "その為にトレインデータだけでなく、val_dataも加味して学習させることが大事。"
      ],
      "metadata": {
        "id": "zzrBuN4RIoV-"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "こちらはvalidationを取り入れた学習。\n",
        "\n",
        "学習データ:trainのうち、10％を検証データ:validationとして使用 ⇒ \n",
        "`validation_split=0.1`\n",
        "\n"
      ],
      "metadata": {
        "id": "7Bng_eLjW3UB"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 構築したモデルで学習\n",
        "history = model.fit(x_train, \n",
        "                    y_train, \n",
        "                    batch_size=batch_size, \n",
        "                    epochs=epochs, \n",
        "                    verbose=1, \n",
        "                    validation_split=0.1\n",
        "                    )"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "I1IcZwA_W1EK",
        "outputId": "1ab290b3-4668-4ba1-80ef-8398a29dc021"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "422/422 [==============================] - 9s 19ms/step - loss: 0.2621 - accuracy: 0.9199 - val_loss: 0.0944 - val_accuracy: 0.9733\n",
            "Epoch 2/20\n",
            "422/422 [==============================] - 8s 19ms/step - loss: 0.1058 - accuracy: 0.9672 - val_loss: 0.0745 - val_accuracy: 0.9763\n",
            "Epoch 3/20\n",
            "422/422 [==============================] - 8s 19ms/step - loss: 0.0786 - accuracy: 0.9759 - val_loss: 0.0803 - val_accuracy: 0.9785\n",
            "Epoch 4/20\n",
            "422/422 [==============================] - 8s 19ms/step - loss: 0.0601 - accuracy: 0.9820 - val_loss: 0.0719 - val_accuracy: 0.9803\n",
            "Epoch 5/20\n",
            "422/422 [==============================] - 13s 31ms/step - loss: 0.0511 - accuracy: 0.9845 - val_loss: 0.0749 - val_accuracy: 0.9815\n",
            "Epoch 6/20\n",
            "422/422 [==============================] - 12s 28ms/step - loss: 0.0443 - accuracy: 0.9867 - val_loss: 0.0682 - val_accuracy: 0.9842\n",
            "Epoch 7/20\n",
            "422/422 [==============================] - 11s 26ms/step - loss: 0.0383 - accuracy: 0.9890 - val_loss: 0.0739 - val_accuracy: 0.9843\n",
            "Epoch 8/20\n",
            "422/422 [==============================] - 10s 24ms/step - loss: 0.0339 - accuracy: 0.9893 - val_loss: 0.0818 - val_accuracy: 0.9842\n",
            "Epoch 9/20\n",
            "422/422 [==============================] - 8s 20ms/step - loss: 0.0294 - accuracy: 0.9910 - val_loss: 0.0958 - val_accuracy: 0.9832\n",
            "Epoch 10/20\n",
            "422/422 [==============================] - 10s 23ms/step - loss: 0.0284 - accuracy: 0.9912 - val_loss: 0.1019 - val_accuracy: 0.9818\n",
            "Epoch 11/20\n",
            "422/422 [==============================] - 8s 19ms/step - loss: 0.0276 - accuracy: 0.9923 - val_loss: 0.1015 - val_accuracy: 0.9838\n",
            "Epoch 12/20\n",
            "422/422 [==============================] - 8s 19ms/step - loss: 0.0249 - accuracy: 0.9930 - val_loss: 0.0913 - val_accuracy: 0.9857\n",
            "Epoch 13/20\n",
            "422/422 [==============================] - 8s 19ms/step - loss: 0.0221 - accuracy: 0.9933 - val_loss: 0.1143 - val_accuracy: 0.9837\n",
            "Epoch 14/20\n",
            "422/422 [==============================] - 8s 19ms/step - loss: 0.0234 - accuracy: 0.9934 - val_loss: 0.1150 - val_accuracy: 0.9822\n",
            "Epoch 15/20\n",
            "422/422 [==============================] - 8s 19ms/step - loss: 0.0186 - accuracy: 0.9940 - val_loss: 0.1005 - val_accuracy: 0.9852\n",
            "Epoch 16/20\n",
            "422/422 [==============================] - 8s 19ms/step - loss: 0.0210 - accuracy: 0.9940 - val_loss: 0.1154 - val_accuracy: 0.9845\n",
            "Epoch 17/20\n",
            "422/422 [==============================] - 9s 22ms/step - loss: 0.0209 - accuracy: 0.9944 - val_loss: 0.1305 - val_accuracy: 0.9828\n",
            "Epoch 18/20\n",
            "422/422 [==============================] - 12s 29ms/step - loss: 0.0194 - accuracy: 0.9946 - val_loss: 0.1260 - val_accuracy: 0.9837\n",
            "Epoch 19/20\n",
            "422/422 [==============================] - 10s 25ms/step - loss: 0.0202 - accuracy: 0.9951 - val_loss: 0.1197 - val_accuracy: 0.9835\n",
            "Epoch 20/20\n",
            "422/422 [==============================] - 10s 23ms/step - loss: 0.0152 - accuracy: 0.9955 - val_loss: 0.1404 - val_accuracy: 0.9843\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "観るべきところはaccuracyとval_accuracy。どちらも高い数字ですね。\n",
        "今のところ、過学習はしていないだろうと考えられる。\n",
        "\n",
        "例えばaccuracyが凄く高く、val_accuracyが低いと、<font color=\"red\">過学習</font>していますよね。　訓練した所しかできてなくて未知のデータに弱いという事なのだから。"
      ],
      "metadata": {
        "id": "Kg7oIL6fJb8A"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "学習結果を可視化してみましょう。"
      ],
      "metadata": {
        "id": "2Z3JLDmsKxhH"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "plt.figure(figsize=(10,6))\n",
        "# 学習過程のプロット\n",
        "# plt.plot(history.epoch, history.history[\"loss\"], label=\"loss\")                 # lossは今回見なくて良いかなと。\n",
        "# plt.plot(history.epoch, history.history[\"val_loss\"], label=\"val_loss\")         # lossは今回見なくて良いかなと。\n",
        "\n",
        "plt.plot(history.epoch, history.history[\"accuracy\"], label=\"accuancy\")\n",
        "plt.plot(history.epoch, history.history[\"val_accuracy\"], label=\"val_accuracy\")\n",
        "\n",
        "plt.legend()"
      ],
      "metadata": {
        "id": "qPrnruqRy-qM",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 392
        },
        "outputId": "8f1937e3-5af5-4438-cd88-d37e8d210a95"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f31b6c2d450>"
            ]
          },
          "metadata": {},
          "execution_count": 22
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x432 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 精度評価"
      ],
      "metadata": {
        "id": "FmK8lqZjMkcv"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# テストデータを用いた最終的、パフォーマンス計測\n",
        "score = model.evaluate(x_test, \n",
        "                            y_test,\n",
        "                            verbose=0\n",
        "                            )\n",
        " \n",
        "print(\"testデータでのloss：\",score[0]) # lossの表示。\n",
        "print(\"testデータでのaccuracy：\",score[1]) # accuracyの表示。"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KvDvSzEoK3lu",
        "outputId": "8be2e644-fb2b-4ff7-db7b-0c5ab897857b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "testデータでのloss： 0.13635851442813873\n",
            "testデータでのaccuracy： 0.9814000129699707\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "今回はかなり良いモデルが出来たんじゃないでしょうか。"
      ],
      "metadata": {
        "id": "4Cf7BfoUKLet"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### これ上のコードと同じじゃない？"
      ],
      "metadata": {
        "id": "MeuJ09R82rdq"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 一つしか予測できてないじゃん。」\n",
        "pred = model.predict(x_test)\n",
        "pred"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "L7JBkKX3q4hW",
        "outputId": "c4f4f9e9-2174-49b2-9cec-4effe1eadd06"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(10000, 10)"
            ]
          },
          "metadata": {},
          "execution_count": 35
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "pred_list = []\n",
        "\n",
        "for i in pred:\n",
        "  pred_list.append(np.argmax(i))\n",
        "pred_list"
      ],
      "metadata": {
        "id": "SzFdHgBIri4O"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(y_test.shape)\n",
        "y_test[0]          # カテゴリカル変数を元に戻さないと。"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cQrTB6f807Zr",
        "outputId": "912c8125-b488-4b48-ec7d-c41e77bc9ff5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(10000, 10)\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0., 0., 0., 0., 0., 0., 0., 1., 0., 0.], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 49
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.metrics import accuracy_score\n",
        "accuracy_score(pred_list, y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 347
        },
        "id": "mQfHqHVGy3ra",
        "outputId": "58ee7dae-72d6-4074-a098-841a03b7ed0c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "error",
          "ename": "ValueError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-44-88166b3507ae>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0msklearn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmetrics\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0maccuracy_score\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0maccuracy_score\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpred_list\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_test\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py\u001b[0m in \u001b[0;36maccuracy_score\u001b[0;34m(y_true, y_pred, normalize, sample_weight)\u001b[0m\n\u001b[1;32m    209\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    210\u001b[0m     \u001b[0;31m# Compute accuracy for each possible representation\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 211\u001b[0;31m     \u001b[0my_type\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_true\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_pred\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_check_targets\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_true\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_pred\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    212\u001b[0m     \u001b[0mcheck_consistent_length\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_true\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_pred\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msample_weight\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    213\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0my_type\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstartswith\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"multilabel\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py\u001b[0m in \u001b[0;36m_check_targets\u001b[0;34m(y_true, y_pred)\u001b[0m\n\u001b[1;32m     93\u001b[0m         raise ValueError(\n\u001b[1;32m     94\u001b[0m             \"Classification metrics can't handle a mix of {0} and {1} targets\".format(\n\u001b[0;32m---> 95\u001b[0;31m                 \u001b[0mtype_true\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtype_pred\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     96\u001b[0m             )\n\u001b[1;32m     97\u001b[0m         )\n",
            "\u001b[0;31mValueError\u001b[0m: Classification metrics can't handle a mix of multiclass and multilabel-indicator targets"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 最後に確認\n",
        "実際に未知のデータを与えて正しく分類分けできるか確かめてみましょう。"
      ],
      "metadata": {
        "id": "921epoJYdg5N"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "そして画像をモデルに読込むにはある程度の処理が必要ですよね。\n",
        "\n",
        "\n",
        "①shapeを合わせる。（大きさの調整、色等）\n",
        "\n",
        "②型(float)に調整\n",
        "\n",
        "③1次元配列に変換"
      ],
      "metadata": {
        "id": "SL4S3HsENnIm"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 画像読込\n",
        "\n",
        "import cv2\n",
        "from google.colab.patches import cv2_imshow\n",
        "img = cv2.imread(\"/content/img_9.jpg\",cv2.IMREAD_GRAYSCALE)\n",
        "#　この際に第二引数にて｢白黒｣で読み込む。という処理を致しております。\n",
        "\n",
        "# これで表示。\n",
        "cv2_imshow(img)\n",
        "\n",
        "#google colab以外\n",
        "# cv2.imshow(\"画像タイトル名前つけ\", 変数名)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 45
        },
        "id": "fNLYxO-Hv0uX",
        "outputId": "5d513537-77e2-4570-8628-eb54faba3cc0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<PIL.Image.Image image mode=L size=28x28 at 0x7F31A5C70ED0>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAABwAAAAcCAAAAABXZoBIAAABxklEQVR4nL2SPWhTURiGn3PPuef+p7eNMRUU6qKDgjhkEl0tGnAQnMTBURzEyUlwcxYU7eDYTXCRggguLv5MooK6BKGljdT+JKk3zc39HG4Sad1914fvPc/h+8AB1wE0hikfVAQBDgCeAgNRBAlgLKAYQZTjaNAAphKjcfCIS4YF5Rvlogkg0qDAB8DYHOVlqLOn5+yl+vvHi4XMbA+yyvaoN1E0HvVERIq+yD0FUUmMHkoH79zVcPBxJbcHG9n52h21Q7pZYqMc7OyDJmnIgfsiKxPXSTQQcvhDXz6B2QcdG4FtSl9WQ1SpOykodhWk1XXb2yhQZHsmKzC18OXzWq/YuVmf+I7jpseWRaT9K5PnF/g3iy9fvzpz9IrIQlJ+ZGIbqM71zAm73ikYdu3uHltItInQjWUZPAvGkybHHRgKrwMOw8s1hup3OeG5hFABAjf1ZqLbbzf7rWrij/oikjpYA0zfFZFu69D4LUOM4cjTeePMNZekWJd2SjjatqeIqX7N1969aW3083zpx62YBPtX9GQ7k8FPEdmS47P4MF2SgBC49r3Yku6Tb6s3wCe1oztBuXj64okXD+drpLEOwaD3L/T/5Q9VsY6HXmviIgAAAABJRU5ErkJggg==\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "<font color=\"red\">**ポイント**</font>\n",
        "\n",
        "今回モデルに学習させたデータは`(28, 28)`の画像ですよね。\n",
        "\n",
        "そして私が読み込んだのは`(28, 28, 3)`。　⇒shapeが異なっています。この<font color=\"red\">3</font>は何？って話。\n",
        "　これは色の**RGB**です。\n",
        "\n",
        "この要素があるからモデルに拒絶(エラー)されたのです。\n",
        "`ValueError: cannot reshape array of size 2352 into shape (1,784)`\n",
        "\n",
        "<font color=\"red\">2352</font> これは1次元配列した値`28*28*3=2352` 　そこから<font color=\"red\">俺が読み込んだ画像colorじゃね？</font>と気づくことが出来ました。"
      ],
      "metadata": {
        "id": "9_FeQ8wvU5jj"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "だから画像を読み込む時に<font color=\"red\">白黒で読み込んでください</font>という指示をしたわけですよ。⇒　これでちゃんとモデルに適合して予測を算出できます。という事。"
      ],
      "metadata": {
        "id": "7jOyvMsBW0Lp"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "img.shape \n",
        "# 28 *28* 3 = 2352"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_B0_oaf5TQ3U",
        "outputId": "b843b008-5442-45f3-effc-2e692d06f787"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(28, 28)"
            ]
          },
          "metadata": {},
          "execution_count": 32
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "今回は独自関数を作成しました。これに合わせて説明致します。"
      ],
      "metadata": {
        "id": "UMXXBi4qA6Dd"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def predicter(img):\n",
        "  img = cv2.resize(img, (28, 28))       # 訓練データと同じサイズに整形\n",
        "  img = img.astype('float32')           # float32に変換正規化\n",
        "  img = np.array(img)/255               # 1～255正規化\n",
        "  img = img.reshape(1, num_input)       # 一次元配列に変換\n",
        "  predict_y = model.predict(img)        # 分類機に入力データを与えて予測（出力：各クラスの予想確率）\n",
        "  predict_number = np.argmax(predict_y) # 最も確率の高い要素番号（=予想する数字）\n",
        "  return \"この画像の番号は:\", predict_number "
      ],
      "metadata": {
        "id": "M_yNuxoz_tHN"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "predicter(img)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "V6h-FaRNAuUd",
        "outputId": "e38d4b18-a47b-4cc9-cbcd-2927961d796a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "('この画像の番号は:', 3)"
            ]
          },
          "metadata": {},
          "execution_count": 41
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "こうすればかなりイメージが湧くのではないでしょうか？\n",
        "確率的に一番高いものを返している。という事です。"
      ],
      "metadata": {
        "id": "k12Ckf25CygL"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "img = cv2.resize(img, (28, 28))       \n",
        "img = img.astype('float32')         \n",
        "img = np.array(img)/255               \n",
        "img = img.reshape(1, num_input)       \n",
        "predict_y = model.predict(img)        "
      ],
      "metadata": {
        "id": "Zqmnz7jvMvGZ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "predict_y"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "z_A_MHQxD1L0",
        "outputId": "d6daf623-d477-4d29-db10-3ddff966e09b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.0000000e+00, 3.0450313e-28, 2.4421430e-22, 1.0000000e+00,\n",
              "        5.2796795e-35, 6.4711155e-21, 0.0000000e+00, 4.2266945e-25,\n",
              "        6.2572861e-25, 8.8468763e-24]], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 44
        }
      ]
    }
  ]
}