training(h)_w_optimized_paramDense.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 52,
      "metadata": {
        "id": "XiSrIbnJK3Qs"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "import datetime\n",
        "import math\n",
        "\n",
        "import IPython\n",
        "import IPython.display\n",
        "import matplotlib as mpl\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "import tensorflow as tf\n",
        "from enum import Enum\n",
        "\n",
        "mpl.rcParams['figure.figsize'] = (25,8)\n",
        "mpl.rcParams['axes.grid'] = False"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "M0m9DlIoK3Qu"
      },
      "source": [
        "### Features:\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 53,
      "metadata": {
        "id": "roK9SbxqK3Qv"
      },
      "outputs": [],
      "source": [
        "DO_NOT_USE_EARLY_STOPPING = True\n",
        "DROP_INDEX_LIKE_COLUMNS = True\n",
        "LABEL_WIDTH_EQ_SHIFT = True\n",
        "WINDOW_WIDTH = 3\n",
        "LABEL_WIDTH = 1\n",
        "SHIFT = 1"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 54,
      "metadata": {
        "id": "Qrlu4mQoK3Qv"
      },
      "outputs": [],
      "source": [
        "USED_STORE_ID = 4051653300272\n",
        "FEATURE_SIZE = 87\n",
        "DATA_ORIGIN = \"./merged_cleaned_FE_imputed(v)_w.csv\"\n",
        "DROPPED_COLUMNS = [\n",
        "    \"WeekoftheYear_cos\",\n",
        "    \"WeekoftheYear_sin\",\n",
        "    \"DayoftheYear_cos\",\n",
        "    \"DayoftheYear_sin\",\n",
        "    \"Month_cos\",\n",
        "    \"Month_sin\",\n",
        "    \"Season_Autumn\",\n",
        "    \"Season_Spring\",\n",
        "    \"Season_Summer\",\n",
        "    \"Season_Winter\",\n",
        "]\n",
        "\n",
        "class PRODUCT_IDs(Enum):\n",
        "    BensLunchtime = \"4260705920294\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 55,
      "metadata": {
        "id": "f6Wm9En4K3Qw"
      },
      "outputs": [],
      "source": [
        "def load_data():\n",
        "    df = pd.read_csv(DATA_ORIGIN, parse_dates=[\"Date\"])\n",
        "    not_indexed = [\"ProductID\", \"Quantity\", \"Price_imputed\", \"Year\"]\n",
        "    df_pivot = df.pivot_table(\n",
        "        index=[col for col in list(df.columns) if col not in not_indexed],\n",
        "        columns=[\"ProductID\"],\n",
        "        values=\"Quantity\",\n",
        "    )\n",
        "    df_pivot[\"total_quantity_day\"] = df_pivot.sum(axis=1)\n",
        "\n",
        "    df_pivot = df_pivot.loc[df_pivot.index.get_level_values(\"StoreID\") == USED_STORE_ID]\n",
        "    df_pivot = df_pivot.droplevel(\"StoreID\")\n",
        "    df_pivot = df_pivot.reset_index()\n",
        "    df_pivot.columns = df_pivot.columns.astype(str)\n",
        "\n",
        "    return df_pivot\n",
        "\n",
        "\n",
        "def add_fourier_features(\n",
        "    data,\n",
        "    features_to_period={\n",
        "        \"Month\": 12.0,\n",
        "        \"DayoftheMonth\": 31.0,\n",
        "        \"WeekoftheMonth\": 4.0,\n",
        "        \"DayoftheWeek\": 7.0,\n",
        "        \"WeekoftheYear\": 52.0,\n",
        "        \"DayoftheYear\": 365.0,\n",
        "    },\n",
        "):\n",
        "    df = data.copy()\n",
        "\n",
        "    for column, period in features_to_period.items():\n",
        "        df[f\"{column}_cos\"] = np.cos(2 * np.pi * df[\"Month\"] / period)\n",
        "        df[f\"{column}_sin\"] = np.sin(2 * np.pi * df[\"Month\"] / period)\n",
        "\n",
        "    return df.drop(\n",
        "        features_to_period.keys(),\n",
        "        axis=1,\n",
        "    )\n",
        "\n",
        "\n",
        "def get_dataset():\n",
        "    df_pivot = load_data()\n",
        "    date_time = pd.to_datetime(df_pivot.pop(\"Date\"), format=\"%Y-%m-%d\")\n",
        "    data = add_fourier_features(df_pivot)\n",
        "    return data, date_time\n",
        "\n",
        "\n",
        "def split_data(data, train_frac=0.7, valid_frac=0.2, test_frac=0.1):\n",
        "    if not math.isclose(train_frac + valid_frac + test_frac, 1):\n",
        "        raise Error(\"Ivalid split sizes\")\n",
        "\n",
        "    n = len(data)\n",
        "    train_df = data[0 : int(n * train_frac)]\n",
        "    val_df = data[int(n * train_frac) : int(n * (train_frac + valid_frac))]\n",
        "    test_df = data[int(n * (train_frac + valid_frac)) :]\n",
        "\n",
        "    return train_df, val_df, test_df\n",
        "\n",
        "\n",
        "def normalize(data, norm_values_to_use=None):\n",
        "    norm_values_used = norm_values_to_use\n",
        "    if norm_values_to_use is None:\n",
        "        std = data.std()\n",
        "        std = std.where(lambda x: np.invert(np.isclose(x, 0)), other=0.3)\n",
        "        norm_values_used = {\n",
        "            \"std\": std,\n",
        "            \"mean\": data.mean(),\n",
        "        }\n",
        "\n",
        "    return (data - norm_values_used[\"mean\"]) / (\n",
        "        norm_values_used[\"std\"]\n",
        "    ), norm_values_used"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 56,
      "metadata": {
        "id": "l0_k3_xVK3Qw"
      },
      "outputs": [],
      "source": [
        "data, time = get_dataset()\n",
        "if DROP_INDEX_LIKE_COLUMNS:\n",
        "    data = data.drop(\n",
        "        labels=DROPPED_COLUMNS,\n",
        "        axis=\"columns\",\n",
        "    )\n",
        "train_df, val_df, test_df = split_data(data)\n",
        "\n",
        "train_df, normalization_params = normalize(train_df)\n",
        "val_df, _ = normalize(val_df, normalization_params)\n",
        "test_df, _ = normalize(test_df, normalization_params)\n",
        "assert FEATURE_SIZE == len(train_df.columns)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 57,
      "metadata": {
        "id": "smOq7gVTK3Qx"
      },
      "outputs": [],
      "source": [
        "# Code reference: https://github.com/anencore94/SlidingWindowGenerator/blob/master/slidingwindow_generator/slidingwindow_generator.py\n",
        "class WindowGenerator():\n",
        "    def __init__(self, input_width, label_width, shift,\n",
        "                train_df, val_df, test_df,\n",
        "                label_columns=None):\n",
        "        '''\n",
        "        This drawing indicates the meaning of the attributes:\n",
        "\n",
        "        ------------------------Time Dimension ----------------->\n",
        "                                               |\n",
        "        <----------total window size---------->|\n",
        "        <------input width------><---shift---->|\n",
        "                    <------label width-------->|\n",
        "                                               |\n",
        "\n",
        "        '''\n",
        "        # Store the raw data.\n",
        "        self.train_df = train_df\n",
        "        self.val_df = val_df\n",
        "        self.test_df = test_df\n",
        "\n",
        "        # Work out the label column indices.\n",
        "        self.label_columns = label_columns\n",
        "        if label_columns is not None:\n",
        "            self.label_columns_indices = {name: i for i, name in\n",
        "                                        enumerate(label_columns)}\n",
        "        self.column_indices = {name: i for i, name in\n",
        "                                enumerate(train_df.columns)}\n",
        "\n",
        "        # Work out the window parameters.\n",
        "        self.input_width = input_width\n",
        "        self.label_width = label_width\n",
        "        self.shift = shift\n",
        "\n",
        "        self.total_window_size = input_width + shift\n",
        "\n",
        "        self.input_slice = slice(0, input_width)\n",
        "        self.input_indices = np.arange(self.total_window_size)[self.input_slice]\n",
        "\n",
        "        label_start = self.total_window_size - self.label_width\n",
        "        self.labels_slice = slice(label_start, None)\n",
        "        self.label_indices = np.arange(self.total_window_size)[self.labels_slice]\n",
        "\n",
        "    def __repr__(self):\n",
        "        return '\\n'.join([\n",
        "            f'Total window size: {self.total_window_size}',\n",
        "            f'Input indices: {self.input_indices}',\n",
        "            f'Label indices: {self.label_indices}',\n",
        "            f'Label column name(s): {self.label_columns}'])\n",
        "\n",
        "    def split_window(self, features):\n",
        "        '''\n",
        "        splits the incoming data: (batch_size, total_window_size, all_features)\n",
        "        into:\n",
        "        inputs: (batch_size, input_width, all_features)\n",
        "        labels: (batch_size, label_width, #label_columns)\n",
        "\n",
        "        while the timesteps are according to the parameter meaning (see description of class)\n",
        "        '''\n",
        "        inputs = features[:, self.input_slice, :]\n",
        "        labels = features[:, self.labels_slice, :]\n",
        "        if self.label_columns is not None:\n",
        "            labels = tf.stack(\n",
        "                [labels[:, :, self.column_indices[name]] for name in self.label_columns],\n",
        "                axis=-1)\n",
        "\n",
        "        # Slicing doesn't preserve static shape information, so set the shapes\n",
        "        # manually. This way the `tf.data.Datasets` are easier to inspect.\n",
        "        inputs.set_shape([None, self.input_width, None])\n",
        "        labels.set_shape([None, self.label_width, None])\n",
        "\n",
        "        return inputs, labels\n",
        "\n",
        "    def plot(self, model=None, plot_col=PRODUCT_IDs.BensLunchtime.value, max_subplots=3, used_normalization=None):\n",
        "        def denormalize(data):\n",
        "            if used_normalization is None:\n",
        "                return data\n",
        "            return data * used_normalization['std'][plot_col] + used_normalization['mean'][plot_col]\n",
        "\n",
        "        inputs, labels = self.example\n",
        "        plt.figure(figsize=(12, 8))\n",
        "        plot_col_index = self.column_indices[plot_col]\n",
        "        max_n = min(max_subplots, len(inputs))\n",
        "        for n in range(max_n):\n",
        "            plt.subplot(max_n, 1, n+1)\n",
        "            plt.ylabel(f'{plot_col} [normed]')\n",
        "            plt.plot(self.input_indices, denormalize(inputs[n, :, plot_col_index]),\n",
        "                    label='Inputs', marker='.', zorder=-10)\n",
        "\n",
        "            if self.label_columns:\n",
        "                label_col_index = self.label_columns_indices.get(plot_col, None)\n",
        "            else:\n",
        "                label_col_index = plot_col_index\n",
        "\n",
        "            if label_col_index is None:\n",
        "                continue\n",
        "\n",
        "            plt.scatter(self.label_indices, denormalize(labels[n, :, label_col_index]),\n",
        "                        edgecolors='k', label='Labels', c='#2ca02c', s=64)\n",
        "\n",
        "            if model is not None:\n",
        "                predictions = model(inputs)\n",
        "                plt.scatter(self.label_indices, denormalize(predictions[n, :, label_col_index]),\n",
        "                        marker='X', edgecolors='k', label='Predictions',\n",
        "                        c='#ff7f0e', s=64)\n",
        "\n",
        "            if n == 0:\n",
        "                plt.legend()\n",
        "\n",
        "            plt.xlabel('Time [day]')\n",
        "\n",
        "    def make_dataset(self, data):\n",
        "        data = np.array(data, dtype=np.float32)\n",
        "        ds = tf.keras.utils.timeseries_dataset_from_array(\n",
        "            data=data,\n",
        "            targets=None,\n",
        "            sequence_length=self.total_window_size,\n",
        "            sequence_stride=1,\n",
        "            shuffle=True,\n",
        "            batch_size=32,)\n",
        "\n",
        "        ds = ds.map(self.split_window)\n",
        "\n",
        "        return ds\n",
        "\n",
        "    @property\n",
        "    def train(self):\n",
        "        return self.make_dataset(self.train_df)\n",
        "\n",
        "    @property\n",
        "    def val(self):\n",
        "        return self.make_dataset(self.val_df)\n",
        "\n",
        "    @property\n",
        "    def test(self):\n",
        "        return self.make_dataset(self.test_df)\n",
        "\n",
        "    @property\n",
        "    def example(self):\n",
        "        \"\"\"Get and cache an example batch of `inputs, labels` for plotting.\"\"\"\n",
        "        result = getattr(self, '_example', None)\n",
        "        if result is None:\n",
        "            # No example batch was found, so get one from the `.train` dataset\n",
        "            result = next(iter(self.train))\n",
        "            # And cache it for next time\n",
        "            self._example = result\n",
        "        return result"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "o7LhpajDK3Qy"
      },
      "source": [
        "-----\n",
        "# General Settings"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 58,
      "metadata": {
        "id": "ToZAaq2BK3Qy"
      },
      "outputs": [],
      "source": [
        "## For tracking Model Performances:\n",
        "val_performance = {}\n",
        "performance = {}\n",
        "train_performance = {}\n",
        "test_performance = {}\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 59,
      "metadata": {
        "id": "TXe8fZw5K3Qy"
      },
      "outputs": [],
      "source": [
        "def get_loss():\n",
        "    ## !!! change function save_performance when changing here\n",
        "    return tf.keras.losses.MeanSquaredError()\n",
        "\n",
        "\n",
        "def get_metrics():\n",
        "    ## !!! change function save_performance when changing here\n",
        "    return [tf.keras.metrics.MeanAbsoluteError()]\n",
        "\n",
        "\n",
        "def save_performance(where):\n",
        "    df = pd.DataFrame(columns=[\"Instances_Train\", \"Loss_Train\", \"MAE_Train\", \"Latency_Train\", \"Trainable_Parameters\", \"Non_Trainable_Parameters\",\n",
        "                               \"Instances_Val\", \"Loss_Val\", \"MAE_Val\", \"Latency_Val\",\n",
        "                               \"Instances_Test\", \"Loss_Test\", \"MAE_Test\", \"Latency_Test\"])\n",
        "\n",
        "    for model_name in train_performance.keys():\n",
        "        df.loc[model_name] = train_performance[model_name] + val_performance[model_name] + test_performance[model_name]\n",
        "\n",
        "    df.to_csv(where)\n",
        "\n",
        "def get_model_size(model):\n",
        "    import os\n",
        "    import tempfile\n",
        "\n",
        "    _, keras_file = tempfile.mkstemp('.h5')\n",
        "    tf.keras.models.save_model(model, keras_file, include_optimizer=False)\n",
        "    print(f\"Model size: {os.path.getsize(keras_file) / 1024 / 1024} MB\")\n",
        "\n",
        "def register_performance(name, model, window):\n",
        "    import time\n",
        "    n_train = window.train_df.shape[0]\n",
        "    n_val = window.val_df.shape[0]\n",
        "    n_test = window.test_df.shape[0]\n",
        "\n",
        "    start_time = time.time()\n",
        "    train_performance[name] = model.evaluate(window.train, verbose=0)\n",
        "    train_time = time.time() - start_time\n",
        "\n",
        "    start_time = time.time()\n",
        "    val_performance[name] = model.evaluate(window.val, verbose=0)\n",
        "    val_time = time.time() - start_time\n",
        "\n",
        "    start_time = time.time()\n",
        "    test_performance[name] = model.evaluate(window.test, verbose=0)\n",
        "    test_time = time.time() - start_time\n",
        "\n",
        "    performance = {\n",
        "        \"n_train\": n_train,\n",
        "        \"n_val\": n_val,\n",
        "        \"n_test\": n_test,\n",
        "        \"train_loss\": train_performance[name][0],\n",
        "        \"train_mae\": train_performance[name][1],\n",
        "        \"val_loss\": val_performance[name][0],\n",
        "        \"val_mae\": val_performance[name][1],\n",
        "        \"test_loss\": test_performance[name][0],\n",
        "        \"test_mae\": test_performance[name][1],\n",
        "        \"train_latency\": train_time / n_train,\n",
        "        \"val_latency\": val_time / n_val,\n",
        "        \"test_latency\": test_time / n_test,\n",
        "        \"network_size_MB\": get_model_size(model),\n",
        "        \"n_trainable_params\": np.sum([tf.size(w_matrix).numpy() for w_matrix in model.trainable_variables]),\n",
        "        \"n_non_trainable_params\": np.sum([tf.size(w_matrix).numpy() for w_matrix in model.non_trainable_variables])\n",
        "    }\n",
        "    return performance"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jjIcxyP2K3Qz"
      },
      "source": [
        "---"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9YeXfTt1K3Qz"
      },
      "source": [
        "## Single Step Models"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 60,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "URzer7bHK3Qz",
        "outputId": "262126c2-e14c-4ac4-f914-53ec86cc1bc9"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "Total window size: 4\n",
              "Input indices: [0 1 2]\n",
              "Label indices: [3]\n",
              "Label column name(s): ['4260705920294']"
            ]
          },
          "execution_count": 60,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "single_step_window = WindowGenerator(\n",
        "    input_width=WINDOW_WIDTH,\n",
        "    label_width=LABEL_WIDTH,\n",
        "    shift=SHIFT,\n",
        "    label_columns=[PRODUCT_IDs.BensLunchtime.value],\n",
        "    train_df=train_df,\n",
        "    val_df=val_df,\n",
        "    test_df=test_df,\n",
        ")\n",
        "single_step_window"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 61,
      "metadata": {
        "id": "e0NT_SvhK3Q1"
      },
      "outputs": [],
      "source": [
        "MAX_EPOCHS = 50\n",
        "\n",
        "\n",
        "def compile_and_fit(model, window, patience=2):\n",
        "    if not DO_NOT_USE_EARLY_STOPPING:\n",
        "        early_stopping = tf.keras.callbacks.EarlyStopping(\n",
        "            monitor=\"val_loss\", patience=patience, mode=\"min\"\n",
        "        )\n",
        "\n",
        "    model.compile(\n",
        "        loss=get_loss(), optimizer=tf.keras.optimizers.Adam(), metrics=get_metrics()\n",
        "    )\n",
        "\n",
        "    callbacks = [early_stopping if not DO_NOT_USE_EARLY_STOPPING else None]\n",
        "\n",
        "    history = model.fit(\n",
        "        window.train,\n",
        "        epochs=MAX_EPOCHS,\n",
        "        validation_data=window.val,\n",
        "        callbacks=[func for func in callbacks if func is not None],\n",
        "    )\n",
        "    return history"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vlJzGIe9K3Q2"
      },
      "source": [
        "## Dense Model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 62,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "R4fXYtFyK3Q2",
        "outputId": "4612ec89-dbea-4287-946b-1a6f48be76bd"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Model: \"sequential_7\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " flatten_7 (Flatten)         (None, 261)               0         \n",
            "                                                                 \n",
            " dense_19 (Dense)            (None, 64)                16768     \n",
            "                                                                 \n",
            " dense_20 (Dense)            (None, 64)                4160      \n",
            "                                                                 \n",
            " dense_21 (Dense)            (None, 1)                 65        \n",
            "                                                                 \n",
            " reshape_7 (Reshape)         (None, 1, 1)              0         \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 20,993\n",
            "Trainable params: 20,993\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n",
            "Epoch 1/50\n",
            "6/6 [==============================] - 1s 50ms/step - loss: 3.5373 - mean_absolute_error: 1.4559 - val_loss: 4.6932 - val_mean_absolute_error: 1.6972\n",
            "Epoch 2/50\n",
            "6/6 [==============================] - 0s 20ms/step - loss: 1.2615 - mean_absolute_error: 0.8892 - val_loss: 1.6119 - val_mean_absolute_error: 1.0664\n",
            "Epoch 3/50\n",
            "6/6 [==============================] - 0s 14ms/step - loss: 1.0866 - mean_absolute_error: 0.8735 - val_loss: 2.1535 - val_mean_absolute_error: 1.1938\n",
            "Epoch 4/50\n",
            "6/6 [==============================] - 0s 22ms/step - loss: 0.8530 - mean_absolute_error: 0.7658 - val_loss: 1.5081 - val_mean_absolute_error: 1.0098\n",
            "Epoch 5/50\n",
            "6/6 [==============================] - 0s 25ms/step - loss: 0.5438 - mean_absolute_error: 0.5959 - val_loss: 1.3004 - val_mean_absolute_error: 0.9497\n",
            "Epoch 6/50\n",
            "6/6 [==============================] - 0s 20ms/step - loss: 0.4274 - mean_absolute_error: 0.5206 - val_loss: 1.4477 - val_mean_absolute_error: 0.9864\n",
            "Epoch 7/50\n",
            "6/6 [==============================] - 0s 24ms/step - loss: 0.3547 - mean_absolute_error: 0.4677 - val_loss: 1.4601 - val_mean_absolute_error: 0.9828\n",
            "Epoch 8/50\n",
            "6/6 [==============================] - 0s 24ms/step - loss: 0.2726 - mean_absolute_error: 0.4070 - val_loss: 1.3688 - val_mean_absolute_error: 0.9552\n",
            "Epoch 9/50\n",
            "6/6 [==============================] - 0s 24ms/step - loss: 0.2240 - mean_absolute_error: 0.3722 - val_loss: 1.2880 - val_mean_absolute_error: 0.9267\n",
            "Epoch 10/50\n",
            "6/6 [==============================] - 0s 24ms/step - loss: 0.1746 - mean_absolute_error: 0.3242 - val_loss: 1.2887 - val_mean_absolute_error: 0.9266\n",
            "Epoch 11/50\n",
            "6/6 [==============================] - 0s 19ms/step - loss: 0.1411 - mean_absolute_error: 0.2816 - val_loss: 1.3273 - val_mean_absolute_error: 0.9344\n",
            "Epoch 12/50\n",
            "6/6 [==============================] - 0s 22ms/step - loss: 0.1153 - mean_absolute_error: 0.2517 - val_loss: 1.3621 - val_mean_absolute_error: 0.9406\n",
            "Epoch 13/50\n",
            "6/6 [==============================] - 0s 21ms/step - loss: 0.0902 - mean_absolute_error: 0.2205 - val_loss: 1.3931 - val_mean_absolute_error: 0.9456\n",
            "Epoch 14/50\n",
            "6/6 [==============================] - 0s 21ms/step - loss: 0.0735 - mean_absolute_error: 0.1968 - val_loss: 1.4316 - val_mean_absolute_error: 0.9526\n",
            "Epoch 15/50\n",
            "6/6 [==============================] - 0s 14ms/step - loss: 0.0599 - mean_absolute_error: 0.1754 - val_loss: 1.4188 - val_mean_absolute_error: 0.9519\n",
            "Epoch 16/50\n",
            "6/6 [==============================] - 0s 20ms/step - loss: 0.0477 - mean_absolute_error: 0.1551 - val_loss: 1.4264 - val_mean_absolute_error: 0.9552\n",
            "Epoch 17/50\n",
            "6/6 [==============================] - 0s 14ms/step - loss: 0.0386 - mean_absolute_error: 0.1362 - val_loss: 1.4482 - val_mean_absolute_error: 0.9591\n",
            "Epoch 18/50\n",
            "6/6 [==============================] - 0s 14ms/step - loss: 0.0303 - mean_absolute_error: 0.1190 - val_loss: 1.4613 - val_mean_absolute_error: 0.9616\n",
            "Epoch 19/50\n",
            "6/6 [==============================] - 0s 21ms/step - loss: 0.0246 - mean_absolute_error: 0.1058 - val_loss: 1.4672 - val_mean_absolute_error: 0.9633\n",
            "Epoch 20/50\n",
            "6/6 [==============================] - 0s 15ms/step - loss: 0.0195 - mean_absolute_error: 0.0914 - val_loss: 1.4579 - val_mean_absolute_error: 0.9637\n",
            "Epoch 21/50\n",
            "6/6 [==============================] - 0s 15ms/step - loss: 0.0158 - mean_absolute_error: 0.0796 - val_loss: 1.4584 - val_mean_absolute_error: 0.9651\n",
            "Epoch 22/50\n",
            "6/6 [==============================] - 0s 13ms/step - loss: 0.0127 - mean_absolute_error: 0.0695 - val_loss: 1.4679 - val_mean_absolute_error: 0.9676\n",
            "Epoch 23/50\n",
            "6/6 [==============================] - 0s 14ms/step - loss: 0.0102 - mean_absolute_error: 0.0603 - val_loss: 1.4849 - val_mean_absolute_error: 0.9707\n",
            "Epoch 24/50\n",
            "6/6 [==============================] - 0s 21ms/step - loss: 0.0084 - mean_absolute_error: 0.0529 - val_loss: 1.4916 - val_mean_absolute_error: 0.9722\n",
            "Epoch 25/50\n",
            "6/6 [==============================] - 0s 15ms/step - loss: 0.0069 - mean_absolute_error: 0.0454 - val_loss: 1.4805 - val_mean_absolute_error: 0.9706\n",
            "Epoch 26/50\n",
            "6/6 [==============================] - 0s 21ms/step - loss: 0.0055 - mean_absolute_error: 0.0399 - val_loss: 1.4766 - val_mean_absolute_error: 0.9699\n",
            "Epoch 27/50\n",
            "6/6 [==============================] - 0s 14ms/step - loss: 0.0047 - mean_absolute_error: 0.0348 - val_loss: 1.4875 - val_mean_absolute_error: 0.9723\n",
            "Epoch 28/50\n",
            "6/6 [==============================] - 0s 21ms/step - loss: 0.0038 - mean_absolute_error: 0.0305 - val_loss: 1.4895 - val_mean_absolute_error: 0.9728\n",
            "Epoch 29/50\n",
            "6/6 [==============================] - 0s 14ms/step - loss: 0.0032 - mean_absolute_error: 0.0273 - val_loss: 1.4838 - val_mean_absolute_error: 0.9717\n",
            "Epoch 30/50\n",
            "6/6 [==============================] - 0s 21ms/step - loss: 0.0027 - mean_absolute_error: 0.0237 - val_loss: 1.4858 - val_mean_absolute_error: 0.9718\n",
            "Epoch 31/50\n",
            "6/6 [==============================] - 0s 21ms/step - loss: 0.0023 - mean_absolute_error: 0.0217 - val_loss: 1.4920 - val_mean_absolute_error: 0.9725\n",
            "Epoch 32/50\n",
            "6/6 [==============================] - 0s 21ms/step - loss: 0.0019 - mean_absolute_error: 0.0188 - val_loss: 1.4893 - val_mean_absolute_error: 0.9718\n",
            "Epoch 33/50\n",
            "6/6 [==============================] - 0s 20ms/step - loss: 0.0016 - mean_absolute_error: 0.0168 - val_loss: 1.4857 - val_mean_absolute_error: 0.9709\n",
            "Epoch 34/50\n",
            "6/6 [==============================] - 0s 15ms/step - loss: 0.0014 - mean_absolute_error: 0.0145 - val_loss: 1.4868 - val_mean_absolute_error: 0.9708\n",
            "Epoch 35/50\n",
            "6/6 [==============================] - 0s 15ms/step - loss: 0.0011 - mean_absolute_error: 0.0125 - val_loss: 1.4884 - val_mean_absolute_error: 0.9709\n",
            "Epoch 36/50\n",
            "6/6 [==============================] - 0s 21ms/step - loss: 9.5069e-04 - mean_absolute_error: 0.0113 - val_loss: 1.4860 - val_mean_absolute_error: 0.9702\n",
            "Epoch 37/50\n",
            "6/6 [==============================] - 0s 20ms/step - loss: 8.2412e-04 - mean_absolute_error: 0.0102 - val_loss: 1.4862 - val_mean_absolute_error: 0.9699\n",
            "Epoch 38/50\n",
            "6/6 [==============================] - 0s 21ms/step - loss: 6.9463e-04 - mean_absolute_error: 0.0090 - val_loss: 1.4881 - val_mean_absolute_error: 0.9702\n",
            "Epoch 39/50\n",
            "6/6 [==============================] - 0s 14ms/step - loss: 5.9532e-04 - mean_absolute_error: 0.0080 - val_loss: 1.4887 - val_mean_absolute_error: 0.9701\n",
            "Epoch 40/50\n",
            "6/6 [==============================] - 0s 23ms/step - loss: 5.2673e-04 - mean_absolute_error: 0.0074 - val_loss: 1.4874 - val_mean_absolute_error: 0.9696\n",
            "Epoch 41/50\n",
            "6/6 [==============================] - 0s 13ms/step - loss: 4.3902e-04 - mean_absolute_error: 0.0069 - val_loss: 1.4883 - val_mean_absolute_error: 0.9695\n",
            "Epoch 42/50\n",
            "6/6 [==============================] - 0s 22ms/step - loss: 3.7487e-04 - mean_absolute_error: 0.0062 - val_loss: 1.4890 - val_mean_absolute_error: 0.9693\n",
            "Epoch 43/50\n",
            "6/6 [==============================] - 0s 23ms/step - loss: 3.3027e-04 - mean_absolute_error: 0.0056 - val_loss: 1.4887 - val_mean_absolute_error: 0.9691\n",
            "Epoch 44/50\n",
            "6/6 [==============================] - 0s 15ms/step - loss: 2.6824e-04 - mean_absolute_error: 0.0051 - val_loss: 1.4870 - val_mean_absolute_error: 0.9686\n",
            "Epoch 45/50\n",
            "6/6 [==============================] - 0s 12ms/step - loss: 2.3084e-04 - mean_absolute_error: 0.0047 - val_loss: 1.4882 - val_mean_absolute_error: 0.9686\n",
            "Epoch 46/50\n",
            "6/6 [==============================] - 0s 15ms/step - loss: 1.9862e-04 - mean_absolute_error: 0.0043 - val_loss: 1.4897 - val_mean_absolute_error: 0.9688\n",
            "Epoch 47/50\n",
            "6/6 [==============================] - 0s 21ms/step - loss: 1.6085e-04 - mean_absolute_error: 0.0039 - val_loss: 1.4892 - val_mean_absolute_error: 0.9686\n",
            "Epoch 48/50\n",
            "6/6 [==============================] - 0s 21ms/step - loss: 1.4943e-04 - mean_absolute_error: 0.0036 - val_loss: 1.4879 - val_mean_absolute_error: 0.9682\n",
            "Epoch 49/50\n",
            "6/6 [==============================] - 0s 13ms/step - loss: 1.1962e-04 - mean_absolute_error: 0.0034 - val_loss: 1.4886 - val_mean_absolute_error: 0.9683\n",
            "Epoch 50/50\n",
            "6/6 [==============================] - 0s 15ms/step - loss: 9.7695e-05 - mean_absolute_error: 0.0029 - val_loss: 1.4896 - val_mean_absolute_error: 0.9683\n",
            "Model size: 0.0982513427734375 MB\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "{'n_train': 190,\n",
              " 'n_val': 54,\n",
              " 'n_test': 28,\n",
              " 'train_loss': 8.68667266331613e-05,\n",
              " 'train_mae': 0.002683955477550626,\n",
              " 'val_loss': 1.4895983934402466,\n",
              " 'val_mae': 0.9683351516723633,\n",
              " 'test_loss': 2.5901131629943848,\n",
              " 'test_mae': 1.3686670064926147,\n",
              " 'train_latency': 0.0007835890117444491,\n",
              " 'val_latency': 0.0029895835452609593,\n",
              " 'test_latency': 0.004357576370239258,\n",
              " 'network_size_MB': None,\n",
              " 'n_trainable_params': 20993,\n",
              " 'n_non_trainable_params': 0.0}"
            ]
          },
          "execution_count": 62,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "dense = tf.keras.Sequential([\n",
        "    tf.keras.layers.Flatten(input_shape=(WINDOW_WIDTH, FEATURE_SIZE)),\n",
        "    tf.keras.layers.Dense(units=64, activation='relu'),\n",
        "    tf.keras.layers.Dense(units=64, activation='relu'),\n",
        "    tf.keras.layers.Dense(units=LABEL_WIDTH),\n",
        "    tf.keras.layers.Reshape(target_shape=(LABEL_WIDTH, 1))\n",
        "])\n",
        "print(dense.summary())\n",
        "\n",
        "history = compile_and_fit(dense, single_step_window)\n",
        "\n",
        "register_performance('Dense', dense, single_step_window)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 63,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7Da1xem_O_BG",
        "outputId": "d1aa8aa0-14a7-48da-dc46-e7c939ea0290"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Model size: 0.0982513427734375 MB\n"
          ]
        }
      ],
      "source": [
        "performance_metrics = register_performance('Dense', dense, single_step_window)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 64,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 696
        },
        "id": "SthhI2XrK3Q2",
        "outputId": "1b8abfba-e933-4424-b370-ad57f5e05a7b"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 1200x800 with 3 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "single_step_window.plot(dense, used_normalization=normalization_params)"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Melersed",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.9.16"
    },
    "orig_nbformat": 4
  },
  "nbformat": 4,
  "nbformat_minor": 0
}