DL2022-HW2-P3.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "# Deep Learning\n",
        "## HW2 - Problem 3 -sign language (80+16pts)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "\n",
        "\n",
        "1.   Full Name: Radmehr Karimian\n",
        "2.   Student Number: 98103556\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "id": "Pd_FdLm34EY4"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense, Activation, InputLayer, Dropout"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Fe36KZ3J4s8P",
        "outputId": "7b0a4a71-64aa-4c00-a090-6f0e0d439182"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Mounted at /content/gdrive\n"
          ]
        }
      ],
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/gdrive')\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0jmuBs1R414M",
        "outputId": "b51c62b9-9c9b-459c-97b5-108bca8dfd6c"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "/content/gdrive/MyDrive/DeepLearning\n"
          ]
        }
      ],
      "source": [
        "cd /content/gdrive/MyDrive/DeepLearning"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 335
        },
        "id": "Pif5mU_U5jZp",
        "outputId": "d636679f-e457-421e-ee2d-70cbb15fc08f"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Shape of training dataframe:(27455, 785)\n",
            "Shape of test dataframe:(7172, 785)\n"
          ]
        },
        {
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-b6a47d6f-cd27-4206-93e5-32b13f80ae79\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>label</th>\n",
              "      <th>pixel1</th>\n",
              "      <th>pixel2</th>\n",
              "      <th>pixel3</th>\n",
              "      <th>pixel4</th>\n",
              "      <th>pixel5</th>\n",
              "      <th>pixel6</th>\n",
              "      <th>pixel7</th>\n",
              "      <th>pixel8</th>\n",
              "      <th>pixel9</th>\n",
              "      <th>...</th>\n",
              "      <th>pixel775</th>\n",
              "      <th>pixel776</th>\n",
              "      <th>pixel777</th>\n",
              "      <th>pixel778</th>\n",
              "      <th>pixel779</th>\n",
              "      <th>pixel780</th>\n",
              "      <th>pixel781</th>\n",
              "      <th>pixel782</th>\n",
              "      <th>pixel783</th>\n",
              "      <th>pixel784</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>3</td>\n",
              "      <td>107</td>\n",
              "      <td>118</td>\n",
              "      <td>127</td>\n",
              "      <td>134</td>\n",
              "      <td>139</td>\n",
              "      <td>143</td>\n",
              "      <td>146</td>\n",
              "      <td>150</td>\n",
              "      <td>153</td>\n",
              "      <td>...</td>\n",
              "      <td>207</td>\n",
              "      <td>207</td>\n",
              "      <td>207</td>\n",
              "      <td>207</td>\n",
              "      <td>206</td>\n",
              "      <td>206</td>\n",
              "      <td>206</td>\n",
              "      <td>204</td>\n",
              "      <td>203</td>\n",
              "      <td>202</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>6</td>\n",
              "      <td>155</td>\n",
              "      <td>157</td>\n",
              "      <td>156</td>\n",
              "      <td>156</td>\n",
              "      <td>156</td>\n",
              "      <td>157</td>\n",
              "      <td>156</td>\n",
              "      <td>158</td>\n",
              "      <td>158</td>\n",
              "      <td>...</td>\n",
              "      <td>69</td>\n",
              "      <td>149</td>\n",
              "      <td>128</td>\n",
              "      <td>87</td>\n",
              "      <td>94</td>\n",
              "      <td>163</td>\n",
              "      <td>175</td>\n",
              "      <td>103</td>\n",
              "      <td>135</td>\n",
              "      <td>149</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>187</td>\n",
              "      <td>188</td>\n",
              "      <td>188</td>\n",
              "      <td>187</td>\n",
              "      <td>187</td>\n",
              "      <td>186</td>\n",
              "      <td>187</td>\n",
              "      <td>188</td>\n",
              "      <td>187</td>\n",
              "      <td>...</td>\n",
              "      <td>202</td>\n",
              "      <td>201</td>\n",
              "      <td>200</td>\n",
              "      <td>199</td>\n",
              "      <td>198</td>\n",
              "      <td>199</td>\n",
              "      <td>198</td>\n",
              "      <td>195</td>\n",
              "      <td>194</td>\n",
              "      <td>195</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>2</td>\n",
              "      <td>211</td>\n",
              "      <td>211</td>\n",
              "      <td>212</td>\n",
              "      <td>212</td>\n",
              "      <td>211</td>\n",
              "      <td>210</td>\n",
              "      <td>211</td>\n",
              "      <td>210</td>\n",
              "      <td>210</td>\n",
              "      <td>...</td>\n",
              "      <td>235</td>\n",
              "      <td>234</td>\n",
              "      <td>233</td>\n",
              "      <td>231</td>\n",
              "      <td>230</td>\n",
              "      <td>226</td>\n",
              "      <td>225</td>\n",
              "      <td>222</td>\n",
              "      <td>229</td>\n",
              "      <td>163</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>13</td>\n",
              "      <td>164</td>\n",
              "      <td>167</td>\n",
              "      <td>170</td>\n",
              "      <td>172</td>\n",
              "      <td>176</td>\n",
              "      <td>179</td>\n",
              "      <td>180</td>\n",
              "      <td>184</td>\n",
              "      <td>185</td>\n",
              "      <td>...</td>\n",
              "      <td>92</td>\n",
              "      <td>105</td>\n",
              "      <td>105</td>\n",
              "      <td>108</td>\n",
              "      <td>133</td>\n",
              "      <td>163</td>\n",
              "      <td>157</td>\n",
              "      <td>163</td>\n",
              "      <td>164</td>\n",
              "      <td>179</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5 rows × 785 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-b6a47d6f-cd27-4206-93e5-32b13f80ae79')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-b6a47d6f-cd27-4206-93e5-32b13f80ae79 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-b6a47d6f-cd27-4206-93e5-32b13f80ae79');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ],
            "text/plain": [
              "   label  pixel1  pixel2  pixel3  pixel4  pixel5  pixel6  pixel7  pixel8  \\\n",
              "0      3     107     118     127     134     139     143     146     150   \n",
              "1      6     155     157     156     156     156     157     156     158   \n",
              "2      2     187     188     188     187     187     186     187     188   \n",
              "3      2     211     211     212     212     211     210     211     210   \n",
              "4     13     164     167     170     172     176     179     180     184   \n",
              "\n",
              "   pixel9  ...  pixel775  pixel776  pixel777  pixel778  pixel779  pixel780  \\\n",
              "0     153  ...       207       207       207       207       206       206   \n",
              "1     158  ...        69       149       128        87        94       163   \n",
              "2     187  ...       202       201       200       199       198       199   \n",
              "3     210  ...       235       234       233       231       230       226   \n",
              "4     185  ...        92       105       105       108       133       163   \n",
              "\n",
              "   pixel781  pixel782  pixel783  pixel784  \n",
              "0       206       204       203       202  \n",
              "1       175       103       135       149  \n",
              "2       198       195       194       195  \n",
              "3       225       222       229       163  \n",
              "4       157       163       164       179  \n",
              "\n",
              "[5 rows x 785 columns]"
            ]
          },
          "execution_count": 6,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "df_train = pd.read_csv('Q3_train.csv')\n",
        "df_test = pd.read_csv('Q3_test.csv')\n",
        "print(f\"Shape of training dataframe:{df_train.shape}\")\n",
        "print(f\"Shape of test dataframe:{df_test.shape}\")\n",
        "df_train.head()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "id": "30p6AHAs5oB3"
      },
      "outputs": [],
      "source": [
        "df_tmp = df_train.copy()\n",
        "y_train = df_tmp.pop('label')\n",
        "X_train = df_tmp\n",
        "\n",
        "df_tmp = df_test.copy()\n",
        "y_test = df_tmp.pop('label')\n",
        "X_test = df_tmp\n",
        "\n",
        "scaler1 = StandardScaler()\n",
        "scaler1.fit(X_train)\n",
        "X_train = scaler1.transform(X_train)\n",
        "X_test = scaler1.transform(X_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 78
        },
        "id": "uRzYr28v5tGO",
        "outputId": "98b3dd89-6e51-46d6-eea8-867928f260e7"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 9 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "np.random.seed(50)\n",
        "pics_no = np.random.randint(0, X_train.shape[0], 9)\n",
        "\n",
        "rows = 1\n",
        "cols = 9\n",
        "axes=[]\n",
        "fig=plt.figure()\n",
        "\n",
        "for j, i in enumerate(pics_no):\n",
        "    img, label = X_train[i], y_train[i]\n",
        "    axes.append(fig.add_subplot(rows, cols, j+1) )\n",
        "    subplot_title=(f\"label = {label}\")\n",
        "    axes[-1].set_title(subplot_title) \n",
        "    plt.axis(False) \n",
        "    plt.imshow(np.array(img).reshape((28,28)), cmap='gray')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 38,
      "metadata": {
        "id": "0xI1c6DE50ms"
      },
      "outputs": [],
      "source": [
        "train_df = df_train.sample(frac=1, random_state=42) # Shuffiling the entire dataset"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 39,
      "metadata": {
        "id": "v2hg0Rm9j4uj"
      },
      "outputs": [],
      "source": [
        "X, y = train_df.drop('label', axis=1), train_df['label'] # Split the dataset into X, y"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 40,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hj5PjsxxTx8c",
        "outputId": "516a9f3a-b257-49d7-f493-0da4be00d763"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "((27455, 784), (27455,))"
            ]
          },
          "execution_count": 40,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "X.shape, y.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 41,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VMFTh-yjkyn1",
        "outputId": "9e240ff8-4ea9-4f02-90b5-bfdcf4b7420a"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "(array([dtype('int64')], dtype=object), dtype('int64'))"
            ]
          },
          "execution_count": 41,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "np.unique(X.dtypes), y.dtype"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 43,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uUF8l7nQk9kX",
        "outputId": "899a977f-3c3f-4235-d832-fa4513e25ae8"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "array([dtype('float64')], dtype=object)"
            ]
          },
          "execution_count": 43,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "X = X/255.0 # Normalizing the training data and converting the data type to float\n",
        "np.unique(X.dtypes)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 44,
      "metadata": {
        "id": "bFees0GrT0eu"
      },
      "outputs": [],
      "source": [
        "from sklearn.preprocessing import LabelBinarizer\n",
        "label_binarizer = LabelBinarizer() # Binarize labels in a one-vs-all fashion (return one-hot encoded vectors)\n",
        "y = label_binarizer.fit_transform(y)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 45,
      "metadata": {
        "id": "HtmHgAvMT4GL"
      },
      "outputs": [],
      "source": [
        "X = tf.reshape(X, [-1, 28, 28, 1])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 34,
      "metadata": {
        "id": "bY2l46fVT--G"
      },
      "outputs": [],
      "source": [
        "X_train, X_valid = X[:25000], X[25000:]\n",
        "y_train, y_valid = y[:25000], y[25000:]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "metadata": {
        "id": "zOziWX2bUCU4"
      },
      "outputs": [],
      "source": [
        "# Defining the CNN\n",
        "\n",
        "model = keras.models.Sequential()\n",
        "model.add(keras.layers.Conv2D(32, (5, 5), padding='same', activation='relu', input_shape=(28, 28, 1)))\n",
        "model.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
        "model.add(keras.layers.Conv2D(64, (5, 5), padding='same', activation='relu'))\n",
        "model.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
        "model.add(keras.layers.Conv2D(128, (5, 5), padding='same', activation='relu'))\n",
        "model.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
        "model.add(keras.layers.Flatten())\n",
        "model.add(keras.layers.Dense(128, activation='relu'))\n",
        "model.add(keras.layers.Dense(24, activation='softmax'))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "id": "ro0UMFeeUG76"
      },
      "outputs": [],
      "source": [
        "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4kKAnzjgUKAs",
        "outputId": "7c83eac4-3028-4e71-b9c2-ac523b58c645"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " conv2d (Conv2D)             (None, 28, 28, 32)        832       \n",
            "                                                                 \n",
            " max_pooling2d (MaxPooling2D  (None, 14, 14, 32)       0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " conv2d_1 (Conv2D)           (None, 14, 14, 64)        51264     \n",
            "                                                                 \n",
            " max_pooling2d_1 (MaxPooling  (None, 7, 7, 64)         0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_2 (Conv2D)           (None, 7, 7, 128)         204928    \n",
            "                                                                 \n",
            " max_pooling2d_2 (MaxPooling  (None, 3, 3, 128)        0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " flatten (Flatten)           (None, 1152)              0         \n",
            "                                                                 \n",
            " dense (Dense)               (None, 128)               147584    \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 24)                3096      \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 407,704\n",
            "Trainable params: 407,704\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "model.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 19,
      "metadata": {
        "id": "LwH9tzraUN9K"
      },
      "outputs": [],
      "source": [
        "save_best_cb = keras.callbacks.ModelCheckpoint('models/initial-end-to-end', save_best_only=True) # Saves the best model so far\n",
        "early_stopping_cb = keras.callbacks.EarlyStopping(patience=5) # Interupts training when there is no progress"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 22,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gG7v6Wq1USPS",
        "outputId": "77ddf4a2-6f10-4176-d55b-7e19fcd7c7a6"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/20\n",
            "781/782 [============================>.] - ETA: 0s - loss: 2.6621e-08 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r782/782 [==============================] - 103s 131ms/step - loss: 2.6617e-08 - accuracy: 1.0000 - val_loss: 5.8465e-07 - val_accuracy: 1.0000\n",
            "Epoch 2/20\n",
            "781/782 [============================>.] - ETA: 0s - loss: 1.8064e-08 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r782/782 [==============================] - 111s 142ms/step - loss: 1.8058e-08 - accuracy: 1.0000 - val_loss: 3.8912e-07 - val_accuracy: 1.0000\n",
            "Epoch 3/20\n",
            "781/782 [============================>.] - ETA: 0s - loss: 1.2054e-08 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r782/782 [==============================] - 101s 129ms/step - loss: 1.2054e-08 - accuracy: 1.0000 - val_loss: 2.9839e-07 - val_accuracy: 1.0000\n",
            "Epoch 4/20\n",
            "781/782 [============================>.] - ETA: 0s - loss: 8.5906e-09 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r782/782 [==============================] - 99s 126ms/step - loss: 8.5878e-09 - accuracy: 1.0000 - val_loss: 1.7788e-07 - val_accuracy: 1.0000\n",
            "Epoch 5/20\n",
            "781/782 [============================>.] - ETA: 0s - loss: 5.9910e-09 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r782/782 [==============================] - 101s 129ms/step - loss: 5.9891e-09 - accuracy: 1.0000 - val_loss: 1.1143e-07 - val_accuracy: 1.0000\n",
            "Epoch 6/20\n",
            "781/782 [============================>.] - ETA: 0s - loss: 4.4885e-09 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r782/782 [==============================] - 101s 130ms/step - loss: 4.4870e-09 - accuracy: 1.0000 - val_loss: 5.9625e-08 - val_accuracy: 1.0000\n",
            "Epoch 7/20\n",
            "781/782 [============================>.] - ETA: 0s - loss: 3.2292e-09 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r782/782 [==============================] - 102s 130ms/step - loss: 3.2282e-09 - accuracy: 1.0000 - val_loss: 3.5786e-08 - val_accuracy: 1.0000\n",
            "Epoch 8/20\n",
            "781/782 [============================>.] - ETA: 0s - loss: 2.6091e-09 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r782/782 [==============================] - 111s 141ms/step - loss: 2.6083e-09 - accuracy: 1.0000 - val_loss: 2.3939e-08 - val_accuracy: 1.0000\n",
            "Epoch 9/20\n",
            "782/782 [==============================] - 102s 131ms/step - loss: 2.1029e-09 - accuracy: 1.0000 - val_loss: 3.0299e-08 - val_accuracy: 1.0000\n",
            "Epoch 10/20\n",
            "781/782 [============================>.] - ETA: 0s - loss: 1.7124e-09 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r782/782 [==============================] - 99s 127ms/step - loss: 1.7118e-09 - accuracy: 1.0000 - val_loss: 1.6121e-08 - val_accuracy: 1.0000\n",
            "Epoch 11/20\n",
            "781/782 [============================>.] - ETA: 0s - loss: 1.4548e-09 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r782/782 [==============================] - 101s 129ms/step - loss: 1.4544e-09 - accuracy: 1.0000 - val_loss: 1.1654e-08 - val_accuracy: 1.0000\n",
            "Epoch 12/20\n",
            "781/782 [============================>.] - ETA: 0s - loss: 1.3165e-09 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r782/782 [==============================] - 101s 129ms/step - loss: 1.3161e-09 - accuracy: 1.0000 - val_loss: 6.9437e-09 - val_accuracy: 1.0000\n",
            "Epoch 13/20\n",
            "782/782 [==============================] - 98s 125ms/step - loss: 0.1032 - accuracy: 0.9730 - val_loss: 0.0376 - val_accuracy: 0.9894\n",
            "Epoch 14/20\n",
            "782/782 [==============================] - 99s 127ms/step - loss: 0.0085 - accuracy: 0.9976 - val_loss: 1.2898e-04 - val_accuracy: 1.0000\n",
            "Epoch 15/20\n",
            "782/782 [==============================] - 99s 127ms/step - loss: 6.4249e-05 - accuracy: 1.0000 - val_loss: 6.7462e-05 - val_accuracy: 1.0000\n",
            "Epoch 16/20\n",
            "782/782 [==============================] - 98s 125ms/step - loss: 3.2023e-05 - accuracy: 1.0000 - val_loss: 4.1584e-05 - val_accuracy: 1.0000\n",
            "Epoch 17/20\n",
            "782/782 [==============================] - 99s 127ms/step - loss: 1.8764e-05 - accuracy: 1.0000 - val_loss: 2.6685e-05 - val_accuracy: 1.0000\n"
          ]
        }
      ],
      "source": [
        "history = model.fit(X_train, y_train, epochs=20, validation_data=(X_valid, y_valid), callbacks=[save_best_cb, early_stopping_cb])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 24,
      "metadata": {
        "id": "y0EDsKrLUXQh"
      },
      "outputs": [],
      "source": [
        "import pickle\n",
        "with open('models/intial-end-to-end-history', 'wb') as history_file:\n",
        "     pickle.dump(history.history, history_file)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 25,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 338
        },
        "id": "EdLDacp_c-fc",
        "outputId": "b84c93c0-ff8f-4a86-81dc-2f46dce08b66"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f7a26f97040>"
            ]
          },
          "execution_count": 25,
          "metadata": {},
          "output_type": "execute_result"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "h = np.load('models/intial-end-to-end-history', allow_pickle=True)\n",
        "best_model = keras.models.load_model('models/initial-end-to-end') \n",
        "fig, ax = plt.subplots(figsize=(10, 5))\n",
        "n_epochs = len(h['loss'])\n",
        "ax.plot(range(1, n_epochs+1), h['loss'], color='b', label='train_loss')\n",
        "ax.plot(range(1, n_epochs+1), h['val_loss'], color='c', label='val_loss')\n",
        "ax.plot(range(1, n_epochs+1), h['accuracy'], color='b', label='train_accuracy', linestyle='--')\n",
        "ax.plot(range(1, n_epochs+1), h['val_accuracy'], color='c', label='val_accuracy', linestyle='--')\n",
        "ax.set_xticks(range(1, n_epochs+1))\n",
        "ax.legend()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 91,
      "metadata": {
        "id": "7eS4qGrnjxGm"
      },
      "outputs": [],
      "source": [
        "X_test, y_test = df_test.drop('label', axis=1), df_test['label']"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 92,
      "metadata": {
        "id": "YxtyxO1CkSyI"
      },
      "outputs": [],
      "source": [
        "X_test = tf.reshape(X_test, [-1, 28, 28, 1])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 93,
      "metadata": {
        "id": "OI-fUSckkbpw"
      },
      "outputs": [],
      "source": [
        "y_test = label_binarizer.transform(y_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 30,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Xbmb48kDkdqZ",
        "outputId": "6ed0f3a1-3e23-4299-cb6c-577bd02a12d2"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "225/225 [==============================] - 8s 34ms/step - loss: 0.2588 - accuracy: 0.9506\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "[0.25876376032829285, 0.950641393661499]"
            ]
          },
          "execution_count": 30,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "best_model.evaluate(X_test, y_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 81,
      "metadata": {
        "id": "hcoxHrWSknhN"
      },
      "outputs": [],
      "source": [
        "\n",
        "model2 = keras.models.Sequential()\n",
        "model2.add(keras.layers.Conv2D(32, (5, 5), padding='same', activation='relu', input_shape=(28, 28, 1)))\n",
        "model2.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
        "model2.add(keras.layers.Conv2D(64, (5, 5), padding='same', activation='relu'))\n",
        "model2.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
        "model2.add(keras.layers.Conv2D(128, (5, 5), padding='same', activation='relu'))\n",
        "model2.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
        "model2.add(keras.layers.Flatten())\n",
        "model2.add(keras.layers.Dense(128, activation='relu'))\n",
        "model2.add(keras.layers.Dense(24, activation='softmax'))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 82,
      "metadata": {
        "id": "rRdkWxfOlzTu"
      },
      "outputs": [],
      "source": [
        " model2.compile(loss='categorical_crossentropy', optimizer='sgd', metrics=['accuracy'])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 83,
      "metadata": {
        "id": "xlcWzttQmA4S"
      },
      "outputs": [],
      "source": [
        "save_best_cb = keras.callbacks.ModelCheckpoint('models/initial-end-to-end2', save_best_only=True) # Saves the best model so far\n",
        "early_stopping_cb = keras.callbacks.EarlyStopping(patience=5) # Interupts training when there is no progress"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 85,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mpNG00NPmJXU",
        "outputId": "1966eef1-f220-42cd-9f66-b0c703f5ba84"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/20\n",
            "610/610 [==============================] - ETA: 0s - loss: 9.5661e-04 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 112s 184ms/step - loss: 9.5661e-04 - accuracy: 1.0000 - val_loss: 0.0011 - val_accuracy: 1.0000\n",
            "Epoch 2/20\n",
            "610/610 [==============================] - ETA: 0s - loss: 8.7911e-04 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 104s 170ms/step - loss: 8.7911e-04 - accuracy: 1.0000 - val_loss: 9.9812e-04 - val_accuracy: 1.0000\n",
            "Epoch 3/20\n",
            "610/610 [==============================] - ETA: 0s - loss: 8.2263e-04 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 98s 161ms/step - loss: 8.2263e-04 - accuracy: 1.0000 - val_loss: 9.5983e-04 - val_accuracy: 1.0000\n",
            "Epoch 4/20\n",
            "610/610 [==============================] - ETA: 0s - loss: 7.6273e-04 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 97s 160ms/step - loss: 7.6273e-04 - accuracy: 1.0000 - val_loss: 8.7744e-04 - val_accuracy: 1.0000\n",
            "Epoch 5/20\n",
            "610/610 [==============================] - ETA: 0s - loss: 7.2445e-04 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 96s 157ms/step - loss: 7.2445e-04 - accuracy: 1.0000 - val_loss: 8.4031e-04 - val_accuracy: 1.0000\n",
            "Epoch 6/20\n",
            "610/610 [==============================] - ETA: 0s - loss: 6.7658e-04 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 99s 162ms/step - loss: 6.7658e-04 - accuracy: 1.0000 - val_loss: 7.8365e-04 - val_accuracy: 1.0000\n",
            "Epoch 7/20\n",
            "610/610 [==============================] - ETA: 0s - loss: 6.3897e-04 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 97s 159ms/step - loss: 6.3897e-04 - accuracy: 1.0000 - val_loss: 7.3486e-04 - val_accuracy: 1.0000\n",
            "Epoch 8/20\n",
            "610/610 [==============================] - ETA: 0s - loss: 6.0330e-04 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 106s 174ms/step - loss: 6.0330e-04 - accuracy: 1.0000 - val_loss: 6.9299e-04 - val_accuracy: 1.0000\n",
            "Epoch 9/20\n",
            "610/610 [==============================] - ETA: 0s - loss: 5.7168e-04 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 97s 160ms/step - loss: 5.7168e-04 - accuracy: 1.0000 - val_loss: 6.7150e-04 - val_accuracy: 1.0000\n",
            "Epoch 10/20\n",
            "610/610 [==============================] - ETA: 0s - loss: 5.4105e-04 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 99s 163ms/step - loss: 5.4105e-04 - accuracy: 1.0000 - val_loss: 6.6047e-04 - val_accuracy: 1.0000\n",
            "Epoch 11/20\n",
            "610/610 [==============================] - ETA: 0s - loss: 5.1651e-04 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 97s 160ms/step - loss: 5.1651e-04 - accuracy: 1.0000 - val_loss: 5.9836e-04 - val_accuracy: 1.0000\n",
            "Epoch 12/20\n",
            "610/610 [==============================] - ETA: 0s - loss: 4.9575e-04 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 99s 162ms/step - loss: 4.9575e-04 - accuracy: 1.0000 - val_loss: 5.7800e-04 - val_accuracy: 1.0000\n",
            "Epoch 13/20\n",
            "610/610 [==============================] - ETA: 0s - loss: 4.7785e-04 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 98s 161ms/step - loss: 4.7785e-04 - accuracy: 1.0000 - val_loss: 5.6214e-04 - val_accuracy: 1.0000\n",
            "Epoch 14/20\n",
            "610/610 [==============================] - ETA: 0s - loss: 4.5361e-04 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 97s 158ms/step - loss: 4.5361e-04 - accuracy: 1.0000 - val_loss: 5.3563e-04 - val_accuracy: 1.0000\n",
            "Epoch 15/20\n",
            "610/610 [==============================] - 98s 160ms/step - loss: 4.3372e-04 - accuracy: 1.0000 - val_loss: 5.5264e-04 - val_accuracy: 1.0000\n",
            "Epoch 16/20\n",
            "610/610 [==============================] - ETA: 0s - loss: 4.1685e-04 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 97s 159ms/step - loss: 4.1685e-04 - accuracy: 1.0000 - val_loss: 5.1422e-04 - val_accuracy: 1.0000\n",
            "Epoch 17/20\n",
            "610/610 [==============================] - ETA: 0s - loss: 4.0248e-04 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 98s 161ms/step - loss: 4.0248e-04 - accuracy: 1.0000 - val_loss: 4.7765e-04 - val_accuracy: 1.0000\n",
            "Epoch 18/20\n",
            "610/610 [==============================] - ETA: 0s - loss: 3.8579e-04 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 97s 159ms/step - loss: 3.8579e-04 - accuracy: 1.0000 - val_loss: 4.6860e-04 - val_accuracy: 1.0000\n",
            "Epoch 19/20\n",
            "610/610 [==============================] - ETA: 0s - loss: 3.7281e-04 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 98s 160ms/step - loss: 3.7281e-04 - accuracy: 1.0000 - val_loss: 4.4951e-04 - val_accuracy: 1.0000\n",
            "Epoch 20/20\n",
            "610/610 [==============================] - ETA: 0s - loss: 3.5841e-04 - accuracy: 1.0000"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r610/610 [==============================] - 97s 159ms/step - loss: 3.5841e-04 - accuracy: 1.0000 - val_loss: 4.3319e-04 - val_accuracy: 1.0000\n"
          ]
        }
      ],
      "source": [
        "history2 = model2.fit(X_train, y_train, epochs=20, validation_data=(X_valid, y_valid), callbacks=[save_best_cb, early_stopping_cb])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 86,
      "metadata": {
        "id": "VyGuomalmToo"
      },
      "outputs": [],
      "source": [
        "import pickle\n",
        "with open('models/intial-end-to-end-history2', 'wb') as history_file:\n",
        "     pickle.dump(history2.history, history_file)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 94,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 338
        },
        "id": "tQQWlgNnt_Y0",
        "outputId": "766571d0-267f-440b-cc1d-9027e7963b61"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f98a2c0bbe0>"
            ]
          },
          "execution_count": 94,
          "metadata": {},
          "output_type": "execute_result"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "h = np.load('models/intial-end-to-end-history2', allow_pickle=True)\n",
        "best_model = keras.models.load_model('models/initial-end-to-end2') \n",
        "fig, ax = plt.subplots(figsize=(10, 5))\n",
        "n_epochs = len(h['loss'])\n",
        "ax.plot(range(1, n_epochs+1), h['loss'], color='b', label='train_loss')\n",
        "ax.plot(range(1, n_epochs+1), h['val_loss'], color='c', label='val_loss')\n",
        "ax.plot(range(1, n_epochs+1), h['accuracy'], color='b', label='train_accuracy', linestyle='--')\n",
        "ax.plot(range(1, n_epochs+1), h['val_accuracy'], color='c', label='val_accuracy', linestyle='--')\n",
        "ax.set_xticks(range(1, n_epochs+1))\n",
        "ax.legend()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 95,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vQjeMxbxuGQY",
        "outputId": "71cf37f9-94fe-4ec8-ee2e-4f0bad148bca"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "225/225 [==============================] - 9s 37ms/step - loss: 98.9735 - accuracy: 0.8985\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "[98.97352600097656, 0.8984941244125366]"
            ]
          },
          "execution_count": 95,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "best_model.evaluate(X_test, y_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 46,
      "metadata": {
        "id": "gZDmg_88uO5u"
      },
      "outputs": [],
      "source": [
        "\n",
        "X_train, X_valid = X[:19500], X[19500:]\n",
        "y_train, y_valid = y[:19500], y[19500:]\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kQrPGzcNvN13",
        "outputId": "7cce1365-c808-4710-c9e4-cd7c66ae6d71"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 1.9226 - accuracy: 0.3937"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 69s 113ms/step - loss: 1.9218 - accuracy: 0.3939 - val_loss: 0.3556 - val_accuracy: 0.9065\n",
            "Epoch 2/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.4113 - accuracy: 0.8573"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 69s 114ms/step - loss: 0.4111 - accuracy: 0.8574 - val_loss: 0.0534 - val_accuracy: 0.9913\n",
            "Epoch 3/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.1762 - accuracy: 0.9420"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 70s 116ms/step - loss: 0.1761 - accuracy: 0.9421 - val_loss: 0.0134 - val_accuracy: 0.9977\n",
            "Epoch 4/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.1092 - accuracy: 0.9638"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 68s 112ms/step - loss: 0.1091 - accuracy: 0.9638 - val_loss: 0.0078 - val_accuracy: 0.9986\n",
            "Epoch 5/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.0815 - accuracy: 0.9739"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 68s 111ms/step - loss: 0.0815 - accuracy: 0.9739 - val_loss: 0.0015 - val_accuracy: 1.0000\n",
            "Epoch 6/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.0711 - accuracy: 0.9778"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 68s 112ms/step - loss: 0.0710 - accuracy: 0.9778 - val_loss: 0.0010 - val_accuracy: 1.0000\n",
            "Epoch 7/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.0535 - accuracy: 0.9834"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 69s 114ms/step - loss: 0.0535 - accuracy: 0.9834 - val_loss: 9.4717e-04 - val_accuracy: 0.9999\n",
            "Epoch 8/10\n",
            "610/610 [==============================] - 67s 110ms/step - loss: 0.0528 - accuracy: 0.9833 - val_loss: 0.0012 - val_accuracy: 0.9999\n",
            "Epoch 9/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.0465 - accuracy: 0.9856"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 69s 113ms/step - loss: 0.0465 - accuracy: 0.9856 - val_loss: 3.5268e-04 - val_accuracy: 1.0000\n",
            "Epoch 10/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.0483 - accuracy: 0.9849"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 70s 115ms/step - loss: 0.0485 - accuracy: 0.9848 - val_loss: 1.6194e-04 - val_accuracy: 1.0000\n",
            "Epoch 1/10\n",
            "610/610 [==============================] - ETA: 0s - loss: 2.1944 - accuracy: 0.3166"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 71s 115ms/step - loss: 2.1944 - accuracy: 0.3166 - val_loss: 0.5187 - val_accuracy: 0.8641\n",
            "Epoch 2/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.6061 - accuracy: 0.7904"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 72s 118ms/step - loss: 0.6060 - accuracy: 0.7905 - val_loss: 0.1081 - val_accuracy: 0.9779\n",
            "Epoch 3/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.3143 - accuracy: 0.8912"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 69s 114ms/step - loss: 0.3142 - accuracy: 0.8912 - val_loss: 0.0348 - val_accuracy: 0.9953\n",
            "Epoch 4/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.2115 - accuracy: 0.9290"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 69s 114ms/step - loss: 0.2114 - accuracy: 0.9290 - val_loss: 0.0126 - val_accuracy: 0.9995\n",
            "Epoch 5/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.1570 - accuracy: 0.9480"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 71s 117ms/step - loss: 0.1570 - accuracy: 0.9481 - val_loss: 0.0056 - val_accuracy: 0.9992\n",
            "Epoch 6/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.1285 - accuracy: 0.9578"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 69s 113ms/step - loss: 0.1285 - accuracy: 0.9577 - val_loss: 0.0039 - val_accuracy: 1.0000\n",
            "Epoch 7/10\n",
            "610/610 [==============================] - 67s 110ms/step - loss: 0.1054 - accuracy: 0.9644 - val_loss: 0.0044 - val_accuracy: 0.9991\n",
            "Epoch 8/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.1019 - accuracy: 0.9671"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 71s 116ms/step - loss: 0.1018 - accuracy: 0.9671 - val_loss: 0.0016 - val_accuracy: 1.0000\n",
            "Epoch 9/10\n",
            "610/610 [==============================] - 68s 111ms/step - loss: 0.0909 - accuracy: 0.9707 - val_loss: 0.0017 - val_accuracy: 1.0000\n",
            "Epoch 10/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.0831 - accuracy: 0.9723"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 69s 113ms/step - loss: 0.0831 - accuracy: 0.9723 - val_loss: 5.3527e-04 - val_accuracy: 1.0000\n",
            "Epoch 1/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 2.6870 - accuracy: 0.1780"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 70s 113ms/step - loss: 2.6865 - accuracy: 0.1782 - val_loss: 1.1665 - val_accuracy: 0.6603\n",
            "Epoch 2/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 1.1522 - accuracy: 0.6077"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 69s 112ms/step - loss: 1.1519 - accuracy: 0.6077 - val_loss: 0.4339 - val_accuracy: 0.8993\n",
            "Epoch 3/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.7097 - accuracy: 0.7527"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 70s 114ms/step - loss: 0.7094 - accuracy: 0.7528 - val_loss: 0.2123 - val_accuracy: 0.9625\n",
            "Epoch 4/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.5291 - accuracy: 0.8173"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 69s 114ms/step - loss: 0.5292 - accuracy: 0.8173 - val_loss: 0.1158 - val_accuracy: 0.9840\n",
            "Epoch 5/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.4105 - accuracy: 0.8573"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 70s 114ms/step - loss: 0.4105 - accuracy: 0.8572 - val_loss: 0.0612 - val_accuracy: 0.9908\n",
            "Epoch 6/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.3361 - accuracy: 0.8827"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 68s 111ms/step - loss: 0.3360 - accuracy: 0.8828 - val_loss: 0.0503 - val_accuracy: 0.9955\n",
            "Epoch 7/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.2916 - accuracy: 0.9001"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 69s 113ms/step - loss: 0.2915 - accuracy: 0.9002 - val_loss: 0.0252 - val_accuracy: 0.9984\n",
            "Epoch 8/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.2515 - accuracy: 0.9152"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "610/610 [==============================] - 70s 115ms/step - loss: 0.2519 - accuracy: 0.9151 - val_loss: 0.0132 - val_accuracy: 0.9989\n",
            "Epoch 9/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.2349 - accuracy: 0.9196"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING:absl:Found untraced functions such as _jit_compiled_convolution_op, _jit_compiled_convolution_op, _jit_compiled_convolution_op while saving (showing 3 of 3). These functions will not be directly callable after loading.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r610/610 [==============================] - 69s 113ms/step - loss: 0.2350 - accuracy: 0.9196 - val_loss: 0.0109 - val_accuracy: 0.9985\n",
            "Epoch 10/10\n",
            "609/610 [============================>.] - ETA: 0s - loss: 0.2154 - accuracy: 0.9271"
          ]
        }
      ],
      "source": [
        "# Models\n",
        "# 'models/experiment-dropout-0'\n",
        "# 'models/experiment-dropout-1'\n",
        "# 'models/experiment-dropout-2'\n",
        "\n",
        "# History objects\n",
        "# 'models/experiment-dropout-0-history'\n",
        "# 'models/experiment-dropout-1-history'\n",
        "# 'models/experiment-dropout-2-history'\n",
        "import pickle\n",
        "early_stopping_cb = keras.callbacks.EarlyStopping(patience=5)\n",
        "\n",
        "dropout_rates = [0.3, 0.4, 0.5]\n",
        "\n",
        "for index, i in enumerate(dropout_rates):\n",
        "    model = keras.models.Sequential()\n",
        "    model.add(keras.layers.Conv2D(24, (5, 5), padding='same', activation='relu'))\n",
        "    model.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
        "    model.add(keras.layers.Dropout(i))\n",
        "    model.add(keras.layers.Conv2D(48, (5, 5), padding='same', activation='relu'))\n",
        "    model.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
        "    model.add(keras.layers.Dropout(i))\n",
        "    model.add(keras.layers.Conv2D(96, (5, 5), padding='same', activation='relu'))\n",
        "    model.add(keras.layers.MaxPooling2D(pool_size=(2, 2)))\n",
        "    model.add(keras.layers.Dropout(i))\n",
        "    model.add(keras.layers.Flatten())\n",
        "    model.add(keras.layers.Dense(128, activation='relu'))\n",
        "    model.add(keras.layers.Dropout(i))\n",
        "    model.add(keras.layers.Dense(24, activation='softmax'))\n",
        "    model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
        "    save_best_cb = keras.callbacks.ModelCheckpoint(f'models/experiment-dropout-{index}', save_best_only=True)\n",
        "    history = model.fit(X_train, y_train, epochs=10, validation_data=(X_valid, y_valid), callbacks=[save_best_cb, early_stopping_cb])\n",
        "    with open(f'models/experiment-dropout-{index}-history', 'wb') as history_file:\n",
        "        pickle.dump(history.history, history_file)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 56,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GMja5btEz6m9",
        "outputId": "9ac396c3-3422-48cb-84b1-4b80d5250c43"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "249/249 [==============================] - 7s 26ms/step - loss: 1.6194e-04 - accuracy: 1.0000\n",
            "249/249 [==============================] - 13s 50ms/step - loss: 5.3527e-04 - accuracy: 1.0000\n",
            "249/249 [==============================] - 13s 51ms/step - loss: 0.0054 - accuracy: 1.0000\n"
          ]
        }
      ],
      "source": [
        "for index in range(3):\n",
        "    model = keras.models.load_model(f'models/experiment-dropout-{index}')\n",
        "    model.evaluate(X_valid, y_valid)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 57,
      "metadata": {
        "id": "ziqwAzFs338N"
      },
      "outputs": [],
      "source": [
        "h_0 = np.load('models/experiment-dropout-0-history', allow_pickle=True)\n",
        "h_1 = np.load('models/experiment-dropout-1-history', allow_pickle=True)\n",
        "h_2 = np.load('models/experiment-dropout-2-history', allow_pickle=True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 60,
      "metadata": {
        "id": "uyK5cJHi4HOk"
      },
      "outputs": [],
      "source": [
        "def get_train_val_plots(h, yticks=None, figsize=(10, 5)):\n",
        "    # h: Any dictionary like the history.history\n",
        "    \n",
        "    fig, ax = plt.subplots(figsize=figsize)\n",
        "    n_epochs = len(h['loss'])\n",
        "    x_loss = np.arange(n_epochs+1)-0.5\n",
        "    \n",
        "    ax.plot(x_loss[x_loss >= 0], h['loss'], color='b', label='train_loss')\n",
        "    ax.plot(range(1, n_epochs+1), h['val_loss'], color='r', label='val_loss')\n",
        "    ax.plot(range(1, n_epochs+1), h['accuracy'], color='b', label='train_accuracy', linestyle='--')\n",
        "    ax.plot(range(1, n_epochs+1), h['val_accuracy'], color='r', label='val_accuracy', linestyle='--')\n",
        "    ax.set_xlim(0, n_epochs)\n",
        "    ax.set_xticks(range(1, n_epochs+1))\n",
        "    if yticks is not None:\n",
        "        ax.set_yticks(yticks)\n",
        "    ax.legend()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 61,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 321
        },
        "id": "_cxToIoj4hHe",
        "outputId": "884e1527-3c2b-43c0-8ea3-03195bc8f300"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "get_train_val_plots(h_0)\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 62,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 320
        },
        "id": "tjn5HD1z5QBV",
        "outputId": "934dc05f-c65f-46fd-c8fd-d4b9be073f08"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "get_train_val_plots(h_1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 63,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 320
        },
        "id": "9-lq3OsH5RcB",
        "outputId": "0bf60848-c036-4f6e-cb57-dda42d759993"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "get_train_val_plots(h_2)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 64,
      "metadata": {
        "id": "7eciGV3u5UPb"
      },
      "outputs": [],
      "source": [
        "best_model = keras.models.load_model('models/experiment-dropout-0/')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 73,
      "metadata": {
        "id": "-uklSRmj5gKO"
      },
      "outputs": [],
      "source": [
        "X_test, y_test = df_test.drop('label', axis=1), df_test['label']"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 74,
      "metadata": {
        "id": "2QYemZJv5lLh"
      },
      "outputs": [],
      "source": [
        "X_test /= 255.0"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 77,
      "metadata": {
        "id": "TKcPZf-n7Vg1"
      },
      "outputs": [],
      "source": [
        "def test_model(model, X_test, y_test, label_binarizer):\n",
        "    X_test_reshape = tf.reshape(X_test, [-1, 28, 28, 1])\n",
        "    y_test_labels = label_binarizer.transform(y_test)\n",
        "    results = model.evaluate(X_test_reshape, y_test_labels)\n",
        "    print(f'Loss: {results[0]:.3f} Accuracy: {results[1]:.3f}')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 78,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BO3fdu8f7YIR",
        "outputId": "a1860195-2354-4fb0-e531-d2e26897405d"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "225/225 [==============================] - 6s 26ms/step - loss: 0.0918 - accuracy: 0.9845\n",
            "Loss: 0.092 Accuracy: 0.985\n"
          ]
        }
      ],
      "source": [
        "test_model(best_model, X_test, y_test, label_binarizer)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 80,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "_TzVI2ko9aQf",
        "outputId": "f895481e-7972-4092-9424-a1f74c23036d"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "execution_count": 80,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "from keras.utils import plot_model\n",
        "plot_model(best_model, to_file='model.png', show_shapes=True, show_layer_names=True)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "## Conclusion: \n",
        "By addition a dropout layer after each hidden layer, we are regularizing the network. As a result, we expect to see a model that is less overfitted, meaning that the validation accuracy increases.<br>\n",
        "As shown in the figures above, with the addition of the validation accuracy and train accuracy stay very close (since no overfitting occurs), however, without dropout layers, the training accuracy could reach 99% while the validation accuracy remained around 98%."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "## Conclusion: \n",
        "\n",
        "Adam is well known to perform worse than SGD for image classification tasks.\n",
        "But not in this senario!\n",
        "SGD converges much slower than ADAM. "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {},
      "outputs": [],
      "source": [
        "data = []\n",
        "labels = []\n",
        "columns = []\n",
        "for i in range(1,785):\n",
        "  columns.append('pixel'+str(i))\n",
        "print(columns)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {},
      "outputs": [],
      "source": [
        "import cv2\n",
        "image = cv2.imread(filename,0)\n",
        "image = cv2.resize(image, (28, 28),interpolation = cv2.INTER_LINEAR)\n",
        "cv2_imshow(image)\n",
        "\n",
        "data.append(list(np.array(image,dtype=np.float64).reshape(784)))\n",
        "labels.append(IMAGE_CLASS)  ##image_lcass"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {},
      "outputs": [],
      "source": [
        "data = np.array(data)\n",
        "dataf = pd.DataFrame(data/255, columns=columns)\n",
        "data = Data(dataf.to_numpy(), np.array())\n",
        "dataloader = DataLoader(dataset=data, batch_size=batch_size, shuffle=False)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {},
      "outputs": [],
      "source": [
        "test_model(best_model, data, labels, label_binarizer)"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}