37_entity_embedding.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "T4"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "source": [
        "!mkdir ~/.kaggle\n",
        "!mv kaggle.json ~/.kaggle"
      ],
      "metadata": {
        "id": "9FSoKRLSvwIt"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!kaggle competitions download -c bike-sharing-demand"
      ],
      "metadata": {
        "id": "dV0maTCfvvNE"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!unzip bike-sharing-demand.zip"
      ],
      "metadata": {
        "id": "lM7HX1-8wG9y"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.preprocessing import LabelEncoder, StandardScaler\n",
        "from tensorflow.keras.layers import Input, Embedding, Flatten, concatenate, Dense, Dropout\n",
        "from tensorflow.keras.models import Model"
      ],
      "metadata": {
        "id": "dw4LLdiNwP4m"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Load the Bike Sharing Demand dataset\n",
        "df = pd.read_csv(\"train.csv\")"
      ],
      "metadata": {
        "id": "KWWvFjT_wWqL"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Preprocess the dataset (you may need to do more data cleaning and feature engineering)\n",
        "df[\"datetime\"] = pd.to_datetime(df[\"datetime\"])\n",
        "df[\"day_of_week\"] = df[\"datetime\"].dt.dayofweek\n",
        "features = [\"day_of_week\", \"temp\", \"humidity\", \"windspeed\", \"count\"]\n",
        "df = df[features]"
      ],
      "metadata": {
        "id": "vqgFQ5ov62oY"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Split the dataset into training and testing sets\n",
        "X = df.drop(\"count\", axis=1)\n",
        "y = df[\"count\"]"
      ],
      "metadata": {
        "id": "WZWbcqtP7GVA"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)"
      ],
      "metadata": {
        "id": "jVTgq3_n7ILi"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Scale numerical features\n",
        "scaler_temp = StandardScaler()\n",
        "scaler_humidity = StandardScaler()\n",
        "scaler_windspeed = StandardScaler()\n",
        "\n",
        "X_train[[\"temp\"]] = scaler_temp.fit_transform(X_train[[\"temp\"]])\n",
        "X_train[[\"humidity\"]] = scaler_humidity.fit_transform(X_train[[\"humidity\"]])\n",
        "X_train[[\"windspeed\"]] = scaler_windspeed.fit_transform(X_train[[\"windspeed\"]])\n",
        "\n",
        "X_test[[\"temp\"]] = scaler_temp.transform(X_test[[\"temp\"]])\n",
        "X_test[[\"humidity\"]] = scaler_humidity.transform(X_test[[\"humidity\"]])\n",
        "X_test[[\"windspeed\"]] = scaler_windspeed.transform(X_test[[\"windspeed\"]])"
      ],
      "metadata": {
        "id": "_BiKGGyA7IgG"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Scale target\n",
        "m = y_train.max()\n",
        "y_train = y_train /m\n",
        "y_test = y_test /m"
      ],
      "metadata": {
        "id": "grUQtYsO7cZg"
      },
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Define input layers for categorical variables\n",
        "input_day_of_week = Input(shape=(1,))\n",
        "\n",
        "# Define embedding layers for categorical variables\n",
        "embedding_dim = 3  # You can adjust this dimension\n",
        "embedding_day_of_week = Embedding(input_dim=7, output_dim=embedding_dim)(input_day_of_week)  # 7 days in a week\n",
        "\n",
        "# Flatten the embedding layer\n",
        "flatten_day_of_week = Flatten()(embedding_day_of_week)\n",
        "\n",
        "# Define input layer for numerical variables\n",
        "input_numerical = Input(shape=(X_train.shape[1] - 1,))  # Minus 1 for the categorical input\n",
        "\n",
        "# Concatenate all input layers\n",
        "concatenated = concatenate([flatten_day_of_week, input_numerical])\n",
        "\n",
        "# Build the model\n",
        "x = Dense(64, activation=\"relu\")(concatenated)\n",
        "x = Dropout(0.5)(x)\n",
        "output = Dense(1, activation=\"linear\")(x)  # Linear activation for regression task\n",
        "\n",
        "model = Model(inputs=[input_day_of_week, input_numerical], outputs=output)\n"
      ],
      "metadata": {
        "id": "FqmIzXOH7Yg1"
      },
      "execution_count": 50,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "tf.keras.utils.plot_model(model, show_shapes=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 754
        },
        "id": "iGQB1M887yL7",
        "outputId": "03e41c72-7d9f-4a98-ac46-16ea11c87dc8"
      },
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 51
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from keras.callbacks import ReduceLROnPlateau\n",
        "\n",
        "# Define the callback\n",
        "reduce_lr = ReduceLROnPlateau(monitor='val_loss', factor=0.5, patience=30)\n",
        "\n",
        "# Compile the model\n",
        "model.compile(optimizer=\"RMSprop\", loss=\"mean_squared_error\", metrics=[\"mae\"])\n",
        "\n",
        "# Train the model\n",
        "h = model.fit(\n",
        "    [X_train[\"day_of_week\"], X_train[[\"temp\", \"humidity\", \"windspeed\"]]],\n",
        "    y_train,\n",
        "    callbacks= [reduce_lr],\n",
        "    epochs=300,\n",
        "    batch_size=512,\n",
        "    validation_split=0.1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cHoQrpKx74oI",
        "outputId": "2b99eac4-b5f7-4b3e-ff98-bd252caa2056"
      },
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/300\n",
            "16/16 [==============================] - 2s 21ms/step - loss: 0.1439 - mae: 0.2940 - val_loss: 0.0401 - val_mae: 0.1455 - lr: 0.0010\n",
            "Epoch 2/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.1019 - mae: 0.2479 - val_loss: 0.0319 - val_mae: 0.1331 - lr: 0.0010\n",
            "Epoch 3/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0799 - mae: 0.2188 - val_loss: 0.0297 - val_mae: 0.1295 - lr: 0.0010\n",
            "Epoch 4/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0658 - mae: 0.1965 - val_loss: 0.0289 - val_mae: 0.1277 - lr: 0.0010\n",
            "Epoch 5/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0558 - mae: 0.1806 - val_loss: 0.0278 - val_mae: 0.1262 - lr: 0.0010\n",
            "Epoch 6/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0465 - mae: 0.1654 - val_loss: 0.0273 - val_mae: 0.1249 - lr: 0.0010\n",
            "Epoch 7/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0409 - mae: 0.1540 - val_loss: 0.0270 - val_mae: 0.1245 - lr: 0.0010\n",
            "Epoch 8/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0359 - mae: 0.1444 - val_loss: 0.0268 - val_mae: 0.1240 - lr: 0.0010\n",
            "Epoch 9/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0333 - mae: 0.1373 - val_loss: 0.0269 - val_mae: 0.1268 - lr: 0.0010\n",
            "Epoch 10/300\n",
            "16/16 [==============================] - 0s 14ms/step - loss: 0.0312 - mae: 0.1327 - val_loss: 0.0266 - val_mae: 0.1224 - lr: 0.0010\n",
            "Epoch 11/300\n",
            "16/16 [==============================] - 0s 13ms/step - loss: 0.0297 - mae: 0.1288 - val_loss: 0.0265 - val_mae: 0.1231 - lr: 0.0010\n",
            "Epoch 12/300\n",
            "16/16 [==============================] - 0s 14ms/step - loss: 0.0282 - mae: 0.1260 - val_loss: 0.0268 - val_mae: 0.1259 - lr: 0.0010\n",
            "Epoch 13/300\n",
            "16/16 [==============================] - 0s 14ms/step - loss: 0.0280 - mae: 0.1261 - val_loss: 0.0266 - val_mae: 0.1215 - lr: 0.0010\n",
            "Epoch 14/300\n",
            "16/16 [==============================] - 0s 12ms/step - loss: 0.0277 - mae: 0.1242 - val_loss: 0.0264 - val_mae: 0.1241 - lr: 0.0010\n",
            "Epoch 15/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0275 - mae: 0.1246 - val_loss: 0.0266 - val_mae: 0.1253 - lr: 0.0010\n",
            "Epoch 16/300\n",
            "16/16 [==============================] - 0s 10ms/step - loss: 0.0271 - mae: 0.1232 - val_loss: 0.0264 - val_mae: 0.1242 - lr: 0.0010\n",
            "Epoch 17/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0270 - mae: 0.1228 - val_loss: 0.0264 - val_mae: 0.1239 - lr: 0.0010\n",
            "Epoch 18/300\n",
            "16/16 [==============================] - 0s 10ms/step - loss: 0.0267 - mae: 0.1222 - val_loss: 0.0265 - val_mae: 0.1237 - lr: 0.0010\n",
            "Epoch 19/300\n",
            "16/16 [==============================] - 0s 11ms/step - loss: 0.0266 - mae: 0.1225 - val_loss: 0.0264 - val_mae: 0.1238 - lr: 0.0010\n",
            "Epoch 20/300\n",
            "16/16 [==============================] - 0s 11ms/step - loss: 0.0265 - mae: 0.1223 - val_loss: 0.0264 - val_mae: 0.1234 - lr: 0.0010\n",
            "Epoch 21/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0266 - mae: 0.1223 - val_loss: 0.0264 - val_mae: 0.1234 - lr: 0.0010\n",
            "Epoch 22/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0264 - mae: 0.1218 - val_loss: 0.0263 - val_mae: 0.1227 - lr: 0.0010\n",
            "Epoch 23/300\n",
            "16/16 [==============================] - 0s 13ms/step - loss: 0.0264 - mae: 0.1220 - val_loss: 0.0263 - val_mae: 0.1237 - lr: 0.0010\n",
            "Epoch 24/300\n",
            "16/16 [==============================] - 0s 10ms/step - loss: 0.0263 - mae: 0.1216 - val_loss: 0.0265 - val_mae: 0.1259 - lr: 0.0010\n",
            "Epoch 25/300\n",
            "16/16 [==============================] - 0s 10ms/step - loss: 0.0263 - mae: 0.1218 - val_loss: 0.0263 - val_mae: 0.1237 - lr: 0.0010\n",
            "Epoch 26/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0263 - mae: 0.1218 - val_loss: 0.0263 - val_mae: 0.1243 - lr: 0.0010\n",
            "Epoch 27/300\n",
            "16/16 [==============================] - 0s 10ms/step - loss: 0.0265 - mae: 0.1223 - val_loss: 0.0263 - val_mae: 0.1230 - lr: 0.0010\n",
            "Epoch 28/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0262 - mae: 0.1214 - val_loss: 0.0263 - val_mae: 0.1225 - lr: 0.0010\n",
            "Epoch 29/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0264 - mae: 0.1216 - val_loss: 0.0263 - val_mae: 0.1232 - lr: 0.0010\n",
            "Epoch 30/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0261 - mae: 0.1208 - val_loss: 0.0263 - val_mae: 0.1240 - lr: 0.0010\n",
            "Epoch 31/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0263 - mae: 0.1218 - val_loss: 0.0262 - val_mae: 0.1226 - lr: 0.0010\n",
            "Epoch 32/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0261 - mae: 0.1208 - val_loss: 0.0262 - val_mae: 0.1229 - lr: 0.0010\n",
            "Epoch 33/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0262 - mae: 0.1213 - val_loss: 0.0261 - val_mae: 0.1234 - lr: 0.0010\n",
            "Epoch 34/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0262 - mae: 0.1213 - val_loss: 0.0262 - val_mae: 0.1237 - lr: 0.0010\n",
            "Epoch 35/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0261 - mae: 0.1212 - val_loss: 0.0261 - val_mae: 0.1234 - lr: 0.0010\n",
            "Epoch 36/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0261 - mae: 0.1210 - val_loss: 0.0261 - val_mae: 0.1239 - lr: 0.0010\n",
            "Epoch 37/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0261 - mae: 0.1213 - val_loss: 0.0262 - val_mae: 0.1236 - lr: 0.0010\n",
            "Epoch 38/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0260 - mae: 0.1209 - val_loss: 0.0261 - val_mae: 0.1232 - lr: 0.0010\n",
            "Epoch 39/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0261 - mae: 0.1213 - val_loss: 0.0261 - val_mae: 0.1230 - lr: 0.0010\n",
            "Epoch 40/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0259 - mae: 0.1208 - val_loss: 0.0262 - val_mae: 0.1242 - lr: 0.0010\n",
            "Epoch 41/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0258 - mae: 0.1205 - val_loss: 0.0263 - val_mae: 0.1253 - lr: 0.0010\n",
            "Epoch 42/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0261 - mae: 0.1209 - val_loss: 0.0261 - val_mae: 0.1229 - lr: 0.0010\n",
            "Epoch 43/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0260 - mae: 0.1209 - val_loss: 0.0261 - val_mae: 0.1232 - lr: 0.0010\n",
            "Epoch 44/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0259 - mae: 0.1204 - val_loss: 0.0263 - val_mae: 0.1241 - lr: 0.0010\n",
            "Epoch 45/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0259 - mae: 0.1207 - val_loss: 0.0262 - val_mae: 0.1241 - lr: 0.0010\n",
            "Epoch 46/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0259 - mae: 0.1203 - val_loss: 0.0261 - val_mae: 0.1233 - lr: 0.0010\n",
            "Epoch 47/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0260 - mae: 0.1208 - val_loss: 0.0260 - val_mae: 0.1229 - lr: 0.0010\n",
            "Epoch 48/300\n",
            "16/16 [==============================] - 0s 10ms/step - loss: 0.0259 - mae: 0.1207 - val_loss: 0.0260 - val_mae: 0.1226 - lr: 0.0010\n",
            "Epoch 49/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0259 - mae: 0.1201 - val_loss: 0.0260 - val_mae: 0.1226 - lr: 0.0010\n",
            "Epoch 50/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0257 - mae: 0.1199 - val_loss: 0.0260 - val_mae: 0.1224 - lr: 0.0010\n",
            "Epoch 51/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0258 - mae: 0.1200 - val_loss: 0.0259 - val_mae: 0.1227 - lr: 0.0010\n",
            "Epoch 52/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0259 - mae: 0.1204 - val_loss: 0.0260 - val_mae: 0.1239 - lr: 0.0010\n",
            "Epoch 53/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0259 - mae: 0.1208 - val_loss: 0.0259 - val_mae: 0.1218 - lr: 0.0010\n",
            "Epoch 54/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0260 - mae: 0.1207 - val_loss: 0.0259 - val_mae: 0.1235 - lr: 0.0010\n",
            "Epoch 55/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0257 - mae: 0.1199 - val_loss: 0.0258 - val_mae: 0.1225 - lr: 0.0010\n",
            "Epoch 56/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0257 - mae: 0.1203 - val_loss: 0.0259 - val_mae: 0.1220 - lr: 0.0010\n",
            "Epoch 57/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0256 - mae: 0.1199 - val_loss: 0.0258 - val_mae: 0.1220 - lr: 0.0010\n",
            "Epoch 58/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0259 - mae: 0.1203 - val_loss: 0.0259 - val_mae: 0.1220 - lr: 0.0010\n",
            "Epoch 59/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0258 - mae: 0.1205 - val_loss: 0.0258 - val_mae: 0.1234 - lr: 0.0010\n",
            "Epoch 60/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0256 - mae: 0.1198 - val_loss: 0.0262 - val_mae: 0.1258 - lr: 0.0010\n",
            "Epoch 61/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0257 - mae: 0.1203 - val_loss: 0.0260 - val_mae: 0.1233 - lr: 0.0010\n",
            "Epoch 62/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0257 - mae: 0.1200 - val_loss: 0.0260 - val_mae: 0.1243 - lr: 0.0010\n",
            "Epoch 63/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0258 - mae: 0.1203 - val_loss: 0.0258 - val_mae: 0.1232 - lr: 0.0010\n",
            "Epoch 64/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0257 - mae: 0.1197 - val_loss: 0.0259 - val_mae: 0.1248 - lr: 0.0010\n",
            "Epoch 65/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0257 - mae: 0.1201 - val_loss: 0.0260 - val_mae: 0.1248 - lr: 0.0010\n",
            "Epoch 66/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0257 - mae: 0.1204 - val_loss: 0.0257 - val_mae: 0.1234 - lr: 0.0010\n",
            "Epoch 67/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0255 - mae: 0.1196 - val_loss: 0.0259 - val_mae: 0.1244 - lr: 0.0010\n",
            "Epoch 68/300\n",
            "16/16 [==============================] - 0s 10ms/step - loss: 0.0257 - mae: 0.1204 - val_loss: 0.0259 - val_mae: 0.1233 - lr: 0.0010\n",
            "Epoch 69/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0256 - mae: 0.1200 - val_loss: 0.0258 - val_mae: 0.1222 - lr: 0.0010\n",
            "Epoch 70/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0256 - mae: 0.1193 - val_loss: 0.0257 - val_mae: 0.1230 - lr: 0.0010\n",
            "Epoch 71/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0257 - mae: 0.1202 - val_loss: 0.0257 - val_mae: 0.1228 - lr: 0.0010\n",
            "Epoch 72/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0256 - mae: 0.1198 - val_loss: 0.0256 - val_mae: 0.1226 - lr: 0.0010\n",
            "Epoch 73/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0257 - mae: 0.1200 - val_loss: 0.0257 - val_mae: 0.1218 - lr: 0.0010\n",
            "Epoch 74/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0257 - mae: 0.1199 - val_loss: 0.0257 - val_mae: 0.1212 - lr: 0.0010\n",
            "Epoch 75/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0255 - mae: 0.1197 - val_loss: 0.0261 - val_mae: 0.1248 - lr: 0.0010\n",
            "Epoch 76/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0257 - mae: 0.1198 - val_loss: 0.0259 - val_mae: 0.1242 - lr: 0.0010\n",
            "Epoch 77/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1197 - val_loss: 0.0256 - val_mae: 0.1226 - lr: 0.0010\n",
            "Epoch 78/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0256 - mae: 0.1201 - val_loss: 0.0261 - val_mae: 0.1253 - lr: 0.0010\n",
            "Epoch 79/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0257 - mae: 0.1203 - val_loss: 0.0259 - val_mae: 0.1238 - lr: 0.0010\n",
            "Epoch 80/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0259 - mae: 0.1207 - val_loss: 0.0257 - val_mae: 0.1222 - lr: 0.0010\n",
            "Epoch 81/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0257 - mae: 0.1199 - val_loss: 0.0259 - val_mae: 0.1243 - lr: 0.0010\n",
            "Epoch 82/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0257 - mae: 0.1203 - val_loss: 0.0257 - val_mae: 0.1221 - lr: 0.0010\n",
            "Epoch 83/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0256 - mae: 0.1193 - val_loss: 0.0260 - val_mae: 0.1244 - lr: 0.0010\n",
            "Epoch 84/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0256 - mae: 0.1199 - val_loss: 0.0257 - val_mae: 0.1227 - lr: 0.0010\n",
            "Epoch 85/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0257 - mae: 0.1197 - val_loss: 0.0258 - val_mae: 0.1234 - lr: 0.0010\n",
            "Epoch 86/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0258 - mae: 0.1203 - val_loss: 0.0258 - val_mae: 0.1235 - lr: 0.0010\n",
            "Epoch 87/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0258 - mae: 0.1205 - val_loss: 0.0257 - val_mae: 0.1222 - lr: 0.0010\n",
            "Epoch 88/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0256 - mae: 0.1197 - val_loss: 0.0257 - val_mae: 0.1234 - lr: 0.0010\n",
            "Epoch 89/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0254 - mae: 0.1196 - val_loss: 0.0257 - val_mae: 0.1239 - lr: 0.0010\n",
            "Epoch 90/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0256 - mae: 0.1200 - val_loss: 0.0257 - val_mae: 0.1226 - lr: 0.0010\n",
            "Epoch 91/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0256 - mae: 0.1203 - val_loss: 0.0256 - val_mae: 0.1213 - lr: 0.0010\n",
            "Epoch 92/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1196 - val_loss: 0.0257 - val_mae: 0.1234 - lr: 0.0010\n",
            "Epoch 93/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0256 - mae: 0.1200 - val_loss: 0.0258 - val_mae: 0.1249 - lr: 0.0010\n",
            "Epoch 94/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0256 - mae: 0.1203 - val_loss: 0.0257 - val_mae: 0.1237 - lr: 0.0010\n",
            "Epoch 95/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0256 - mae: 0.1201 - val_loss: 0.0256 - val_mae: 0.1217 - lr: 0.0010\n",
            "Epoch 96/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1194 - val_loss: 0.0256 - val_mae: 0.1223 - lr: 0.0010\n",
            "Epoch 97/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0255 - mae: 0.1195 - val_loss: 0.0258 - val_mae: 0.1239 - lr: 0.0010\n",
            "Epoch 98/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0256 - mae: 0.1196 - val_loss: 0.0255 - val_mae: 0.1221 - lr: 0.0010\n",
            "Epoch 99/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1194 - val_loss: 0.0257 - val_mae: 0.1234 - lr: 0.0010\n",
            "Epoch 100/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0257 - mae: 0.1200 - val_loss: 0.0258 - val_mae: 0.1239 - lr: 0.0010\n",
            "Epoch 101/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1199 - val_loss: 0.0258 - val_mae: 0.1234 - lr: 0.0010\n",
            "Epoch 102/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0255 - mae: 0.1195 - val_loss: 0.0256 - val_mae: 0.1220 - lr: 0.0010\n",
            "Epoch 103/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0256 - mae: 0.1197 - val_loss: 0.0255 - val_mae: 0.1214 - lr: 5.0000e-04\n",
            "Epoch 104/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0254 - mae: 0.1189 - val_loss: 0.0256 - val_mae: 0.1225 - lr: 5.0000e-04\n",
            "Epoch 105/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1194 - val_loss: 0.0256 - val_mae: 0.1229 - lr: 5.0000e-04\n",
            "Epoch 106/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1197 - val_loss: 0.0255 - val_mae: 0.1217 - lr: 5.0000e-04\n",
            "Epoch 107/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1216 - lr: 5.0000e-04\n",
            "Epoch 108/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0254 - mae: 0.1194 - val_loss: 0.0256 - val_mae: 0.1229 - lr: 5.0000e-04\n",
            "Epoch 109/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0255 - mae: 0.1198 - val_loss: 0.0256 - val_mae: 0.1227 - lr: 5.0000e-04\n",
            "Epoch 110/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1196 - val_loss: 0.0255 - val_mae: 0.1220 - lr: 5.0000e-04\n",
            "Epoch 111/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1196 - val_loss: 0.0256 - val_mae: 0.1217 - lr: 5.0000e-04\n",
            "Epoch 112/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0254 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1218 - lr: 5.0000e-04\n",
            "Epoch 113/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0256 - mae: 0.1195 - val_loss: 0.0256 - val_mae: 0.1221 - lr: 5.0000e-04\n",
            "Epoch 114/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0255 - mae: 0.1195 - val_loss: 0.0257 - val_mae: 0.1235 - lr: 5.0000e-04\n",
            "Epoch 115/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1195 - val_loss: 0.0256 - val_mae: 0.1233 - lr: 5.0000e-04\n",
            "Epoch 116/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1196 - val_loss: 0.0256 - val_mae: 0.1224 - lr: 5.0000e-04\n",
            "Epoch 117/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1195 - val_loss: 0.0256 - val_mae: 0.1219 - lr: 5.0000e-04\n",
            "Epoch 118/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0257 - mae: 0.1198 - val_loss: 0.0256 - val_mae: 0.1233 - lr: 5.0000e-04\n",
            "Epoch 119/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0255 - mae: 0.1197 - val_loss: 0.0255 - val_mae: 0.1222 - lr: 5.0000e-04\n",
            "Epoch 120/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0256 - mae: 0.1194 - val_loss: 0.0257 - val_mae: 0.1238 - lr: 5.0000e-04\n",
            "Epoch 121/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1196 - val_loss: 0.0256 - val_mae: 0.1223 - lr: 5.0000e-04\n",
            "Epoch 122/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0256 - mae: 0.1196 - val_loss: 0.0256 - val_mae: 0.1225 - lr: 5.0000e-04\n",
            "Epoch 123/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1192 - val_loss: 0.0257 - val_mae: 0.1231 - lr: 5.0000e-04\n",
            "Epoch 124/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1194 - val_loss: 0.0257 - val_mae: 0.1227 - lr: 5.0000e-04\n",
            "Epoch 125/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1196 - val_loss: 0.0256 - val_mae: 0.1222 - lr: 5.0000e-04\n",
            "Epoch 126/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0256 - mae: 0.1196 - val_loss: 0.0255 - val_mae: 0.1222 - lr: 5.0000e-04\n",
            "Epoch 127/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0253 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1215 - lr: 5.0000e-04\n",
            "Epoch 128/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0254 - mae: 0.1191 - val_loss: 0.0255 - val_mae: 0.1220 - lr: 5.0000e-04\n",
            "Epoch 129/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0254 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1226 - lr: 5.0000e-04\n",
            "Epoch 130/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0255 - mae: 0.1196 - val_loss: 0.0256 - val_mae: 0.1227 - lr: 5.0000e-04\n",
            "Epoch 131/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0256 - mae: 0.1195 - val_loss: 0.0256 - val_mae: 0.1218 - lr: 5.0000e-04\n",
            "Epoch 132/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0256 - mae: 0.1193 - val_loss: 0.0256 - val_mae: 0.1233 - lr: 5.0000e-04\n",
            "Epoch 133/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0255 - mae: 0.1200 - val_loss: 0.0255 - val_mae: 0.1216 - lr: 5.0000e-04\n",
            "Epoch 134/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0256 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 2.5000e-04\n",
            "Epoch 135/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0256 - mae: 0.1199 - val_loss: 0.0256 - val_mae: 0.1224 - lr: 2.5000e-04\n",
            "Epoch 136/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1190 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 2.5000e-04\n",
            "Epoch 137/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0254 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1226 - lr: 2.5000e-04\n",
            "Epoch 138/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0253 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1222 - lr: 2.5000e-04\n",
            "Epoch 139/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1193 - val_loss: 0.0256 - val_mae: 0.1229 - lr: 2.5000e-04\n",
            "Epoch 140/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0255 - mae: 0.1197 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 2.5000e-04\n",
            "Epoch 141/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0254 - mae: 0.1191 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 2.5000e-04\n",
            "Epoch 142/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1192 - val_loss: 0.0256 - val_mae: 0.1228 - lr: 2.5000e-04\n",
            "Epoch 143/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1189 - val_loss: 0.0256 - val_mae: 0.1232 - lr: 2.5000e-04\n",
            "Epoch 144/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1226 - lr: 2.5000e-04\n",
            "Epoch 145/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1187 - val_loss: 0.0256 - val_mae: 0.1230 - lr: 2.5000e-04\n",
            "Epoch 146/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1195 - val_loss: 0.0256 - val_mae: 0.1229 - lr: 2.5000e-04\n",
            "Epoch 147/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0255 - mae: 0.1199 - val_loss: 0.0256 - val_mae: 0.1226 - lr: 2.5000e-04\n",
            "Epoch 148/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0255 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1226 - lr: 2.5000e-04\n",
            "Epoch 149/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1190 - val_loss: 0.0256 - val_mae: 0.1229 - lr: 2.5000e-04\n",
            "Epoch 150/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1194 - val_loss: 0.0255 - val_mae: 0.1222 - lr: 2.5000e-04\n",
            "Epoch 151/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1227 - lr: 2.5000e-04\n",
            "Epoch 152/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1194 - val_loss: 0.0255 - val_mae: 0.1226 - lr: 2.5000e-04\n",
            "Epoch 153/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0256 - mae: 0.1201 - val_loss: 0.0254 - val_mae: 0.1218 - lr: 2.5000e-04\n",
            "Epoch 154/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1190 - val_loss: 0.0255 - val_mae: 0.1228 - lr: 2.5000e-04\n",
            "Epoch 155/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0254 - mae: 0.1197 - val_loss: 0.0255 - val_mae: 0.1227 - lr: 2.5000e-04\n",
            "Epoch 156/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0256 - mae: 0.1198 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 2.5000e-04\n",
            "Epoch 157/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0255 - mae: 0.1195 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 2.5000e-04\n",
            "Epoch 158/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1220 - lr: 2.5000e-04\n",
            "Epoch 159/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0255 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1228 - lr: 2.5000e-04\n",
            "Epoch 160/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1195 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 2.5000e-04\n",
            "Epoch 161/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0252 - mae: 0.1188 - val_loss: 0.0255 - val_mae: 0.1220 - lr: 2.5000e-04\n",
            "Epoch 162/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1194 - val_loss: 0.0255 - val_mae: 0.1223 - lr: 2.5000e-04\n",
            "Epoch 163/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0256 - mae: 0.1195 - val_loss: 0.0256 - val_mae: 0.1229 - lr: 2.5000e-04\n",
            "Epoch 164/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1193 - val_loss: 0.0256 - val_mae: 0.1228 - lr: 1.2500e-04\n",
            "Epoch 165/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1195 - val_loss: 0.0256 - val_mae: 0.1226 - lr: 1.2500e-04\n",
            "Epoch 166/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0255 - mae: 0.1197 - val_loss: 0.0256 - val_mae: 0.1226 - lr: 1.2500e-04\n",
            "Epoch 167/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1226 - lr: 1.2500e-04\n",
            "Epoch 168/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0253 - mae: 0.1188 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.2500e-04\n",
            "Epoch 169/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0253 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1222 - lr: 1.2500e-04\n",
            "Epoch 170/300\n",
            "16/16 [==============================] - 0s 10ms/step - loss: 0.0254 - mae: 0.1190 - val_loss: 0.0255 - val_mae: 0.1222 - lr: 1.2500e-04\n",
            "Epoch 171/300\n",
            "16/16 [==============================] - 0s 10ms/step - loss: 0.0254 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1222 - lr: 1.2500e-04\n",
            "Epoch 172/300\n",
            "16/16 [==============================] - 0s 10ms/step - loss: 0.0256 - mae: 0.1195 - val_loss: 0.0256 - val_mae: 0.1227 - lr: 1.2500e-04\n",
            "Epoch 173/300\n",
            "16/16 [==============================] - 0s 10ms/step - loss: 0.0255 - mae: 0.1198 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 1.2500e-04\n",
            "Epoch 174/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0256 - mae: 0.1191 - val_loss: 0.0255 - val_mae: 0.1226 - lr: 1.2500e-04\n",
            "Epoch 175/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0255 - mae: 0.1196 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.2500e-04\n",
            "Epoch 176/300\n",
            "16/16 [==============================] - 0s 10ms/step - loss: 0.0254 - mae: 0.1193 - val_loss: 0.0256 - val_mae: 0.1228 - lr: 1.2500e-04\n",
            "Epoch 177/300\n",
            "16/16 [==============================] - 0s 10ms/step - loss: 0.0253 - mae: 0.1191 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 1.2500e-04\n",
            "Epoch 178/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0255 - mae: 0.1190 - val_loss: 0.0255 - val_mae: 0.1223 - lr: 1.2500e-04\n",
            "Epoch 179/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0253 - mae: 0.1187 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 1.2500e-04\n",
            "Epoch 180/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0256 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1223 - lr: 1.2500e-04\n",
            "Epoch 181/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0255 - mae: 0.1195 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 1.2500e-04\n",
            "Epoch 182/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0254 - mae: 0.1190 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 1.2500e-04\n",
            "Epoch 183/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0254 - mae: 0.1190 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.2500e-04\n",
            "Epoch 184/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0255 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1226 - lr: 1.2500e-04\n",
            "Epoch 185/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0254 - mae: 0.1191 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.2500e-04\n",
            "Epoch 186/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0253 - mae: 0.1190 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.2500e-04\n",
            "Epoch 187/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0252 - mae: 0.1188 - val_loss: 0.0255 - val_mae: 0.1223 - lr: 1.2500e-04\n",
            "Epoch 188/300\n",
            "16/16 [==============================] - 0s 10ms/step - loss: 0.0256 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1222 - lr: 1.2500e-04\n",
            "Epoch 189/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0254 - mae: 0.1190 - val_loss: 0.0255 - val_mae: 0.1223 - lr: 1.2500e-04\n",
            "Epoch 190/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0255 - mae: 0.1190 - val_loss: 0.0255 - val_mae: 0.1223 - lr: 1.2500e-04\n",
            "Epoch 191/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0255 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1226 - lr: 1.2500e-04\n",
            "Epoch 192/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0256 - mae: 0.1203 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.2500e-04\n",
            "Epoch 193/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0251 - mae: 0.1188 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.2500e-04\n",
            "Epoch 194/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0256 - mae: 0.1195 - val_loss: 0.0255 - val_mae: 0.1226 - lr: 6.2500e-05\n",
            "Epoch 195/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 6.2500e-05\n",
            "Epoch 196/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0252 - mae: 0.1188 - val_loss: 0.0255 - val_mae: 0.1226 - lr: 6.2500e-05\n",
            "Epoch 197/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1197 - val_loss: 0.0255 - val_mae: 0.1226 - lr: 6.2500e-05\n",
            "Epoch 198/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1191 - val_loss: 0.0255 - val_mae: 0.1226 - lr: 6.2500e-05\n",
            "Epoch 199/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1188 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 6.2500e-05\n",
            "Epoch 200/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0254 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 6.2500e-05\n",
            "Epoch 201/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0253 - mae: 0.1188 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 6.2500e-05\n",
            "Epoch 202/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1194 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 6.2500e-05\n",
            "Epoch 203/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1190 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 6.2500e-05\n",
            "Epoch 204/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0256 - mae: 0.1197 - val_loss: 0.0255 - val_mae: 0.1223 - lr: 6.2500e-05\n",
            "Epoch 205/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0254 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 6.2500e-05\n",
            "Epoch 206/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0255 - mae: 0.1195 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 6.2500e-05\n",
            "Epoch 207/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0254 - mae: 0.1194 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 6.2500e-05\n",
            "Epoch 208/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1191 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 6.2500e-05\n",
            "Epoch 209/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1187 - val_loss: 0.0255 - val_mae: 0.1223 - lr: 6.2500e-05\n",
            "Epoch 210/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1188 - val_loss: 0.0255 - val_mae: 0.1223 - lr: 6.2500e-05\n",
            "Epoch 211/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0256 - mae: 0.1195 - val_loss: 0.0255 - val_mae: 0.1223 - lr: 6.2500e-05\n",
            "Epoch 212/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0254 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 6.2500e-05\n",
            "Epoch 213/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1191 - val_loss: 0.0255 - val_mae: 0.1226 - lr: 6.2500e-05\n",
            "Epoch 214/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0254 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 6.2500e-05\n",
            "Epoch 215/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0254 - mae: 0.1194 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 6.2500e-05\n",
            "Epoch 216/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0254 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 6.2500e-05\n",
            "Epoch 217/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0252 - mae: 0.1188 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 6.2500e-05\n",
            "Epoch 218/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1195 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 6.2500e-05\n",
            "Epoch 219/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 6.2500e-05\n",
            "Epoch 220/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1189 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 6.2500e-05\n",
            "Epoch 221/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0254 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 6.2500e-05\n",
            "Epoch 222/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0255 - mae: 0.1195 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 6.2500e-05\n",
            "Epoch 223/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0256 - mae: 0.1196 - val_loss: 0.0255 - val_mae: 0.1223 - lr: 6.2500e-05\n",
            "Epoch 224/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1190 - val_loss: 0.0255 - val_mae: 0.1223 - lr: 3.1250e-05\n",
            "Epoch 225/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1188 - val_loss: 0.0255 - val_mae: 0.1223 - lr: 3.1250e-05\n",
            "Epoch 226/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1191 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 3.1250e-05\n",
            "Epoch 227/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0252 - mae: 0.1187 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 3.1250e-05\n",
            "Epoch 228/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0254 - mae: 0.1195 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 3.1250e-05\n",
            "Epoch 229/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0253 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 3.1250e-05\n",
            "Epoch 230/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1188 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 3.1250e-05\n",
            "Epoch 231/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 3.1250e-05\n",
            "Epoch 232/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0252 - mae: 0.1191 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 3.1250e-05\n",
            "Epoch 233/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0255 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 3.1250e-05\n",
            "Epoch 234/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 3.1250e-05\n",
            "Epoch 235/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0254 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 3.1250e-05\n",
            "Epoch 236/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0254 - mae: 0.1194 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 3.1250e-05\n",
            "Epoch 237/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0254 - mae: 0.1187 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 3.1250e-05\n",
            "Epoch 238/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 3.1250e-05\n",
            "Epoch 239/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1194 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 3.1250e-05\n",
            "Epoch 240/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 3.1250e-05\n",
            "Epoch 241/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0253 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 3.1250e-05\n",
            "Epoch 242/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0256 - mae: 0.1197 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 3.1250e-05\n",
            "Epoch 243/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0255 - mae: 0.1195 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 3.1250e-05\n",
            "Epoch 244/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1194 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 3.1250e-05\n",
            "Epoch 245/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1195 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 3.1250e-05\n",
            "Epoch 246/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0252 - mae: 0.1186 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 3.1250e-05\n",
            "Epoch 247/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0255 - mae: 0.1194 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 3.1250e-05\n",
            "Epoch 248/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0254 - mae: 0.1194 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 3.1250e-05\n",
            "Epoch 249/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1194 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 3.1250e-05\n",
            "Epoch 250/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 3.1250e-05\n",
            "Epoch 251/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0254 - mae: 0.1191 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 3.1250e-05\n",
            "Epoch 252/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 3.1250e-05\n",
            "Epoch 253/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0252 - mae: 0.1190 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 3.1250e-05\n",
            "Epoch 254/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1194 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 1.5625e-05\n",
            "Epoch 255/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 1.5625e-05\n",
            "Epoch 256/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 257/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0254 - mae: 0.1190 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 1.5625e-05\n",
            "Epoch 258/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0254 - mae: 0.1191 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 1.5625e-05\n",
            "Epoch 259/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0253 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 1.5625e-05\n",
            "Epoch 260/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0254 - mae: 0.1189 - val_loss: 0.0255 - val_mae: 0.1224 - lr: 1.5625e-05\n",
            "Epoch 261/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0254 - mae: 0.1190 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 262/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0256 - mae: 0.1195 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 263/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 264/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0254 - mae: 0.1191 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 265/300\n",
            "16/16 [==============================] - 0s 5ms/step - loss: 0.0254 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 266/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0252 - mae: 0.1187 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 267/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1190 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 268/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0254 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 269/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1196 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 270/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0253 - mae: 0.1186 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 271/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0252 - mae: 0.1187 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 272/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0254 - mae: 0.1194 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 273/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1195 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 274/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0256 - mae: 0.1197 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 275/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0254 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 276/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0255 - mae: 0.1198 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 277/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0255 - mae: 0.1196 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 278/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1189 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 279/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0254 - mae: 0.1191 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 280/300\n",
            "16/16 [==============================] - 0s 6ms/step - loss: 0.0253 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 281/300\n",
            "16/16 [==============================] - 0s 7ms/step - loss: 0.0254 - mae: 0.1190 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 282/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0253 - mae: 0.1190 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 283/300\n",
            "16/16 [==============================] - 0s 10ms/step - loss: 0.0253 - mae: 0.1189 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 1.5625e-05\n",
            "Epoch 284/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0255 - mae: 0.1196 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 7.8125e-06\n",
            "Epoch 285/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0253 - mae: 0.1189 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 7.8125e-06\n",
            "Epoch 286/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0254 - mae: 0.1194 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 7.8125e-06\n",
            "Epoch 287/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0255 - mae: 0.1194 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 7.8125e-06\n",
            "Epoch 288/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0253 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 7.8125e-06\n",
            "Epoch 289/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0254 - mae: 0.1195 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 7.8125e-06\n",
            "Epoch 290/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0253 - mae: 0.1191 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 7.8125e-06\n",
            "Epoch 291/300\n",
            "16/16 [==============================] - 0s 10ms/step - loss: 0.0254 - mae: 0.1196 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 7.8125e-06\n",
            "Epoch 292/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0254 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 7.8125e-06\n",
            "Epoch 293/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0252 - mae: 0.1187 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 7.8125e-06\n",
            "Epoch 294/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0253 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 7.8125e-06\n",
            "Epoch 295/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0254 - mae: 0.1194 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 7.8125e-06\n",
            "Epoch 296/300\n",
            "16/16 [==============================] - 0s 8ms/step - loss: 0.0254 - mae: 0.1194 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 7.8125e-06\n",
            "Epoch 297/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0254 - mae: 0.1193 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 7.8125e-06\n",
            "Epoch 298/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0254 - mae: 0.1192 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 7.8125e-06\n",
            "Epoch 299/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0253 - mae: 0.1191 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 7.8125e-06\n",
            "Epoch 300/300\n",
            "16/16 [==============================] - 0s 9ms/step - loss: 0.0255 - mae: 0.1197 - val_loss: 0.0255 - val_mae: 0.1225 - lr: 7.8125e-06\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "plt.plot(h.history['mae'])\n",
        "plt.plot(h.history['val_mae'])\n",
        "plt.legend([\"train\", \"val\"])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 448
        },
        "id": "P0lJ7ofX8ALt",
        "outputId": "d87bc9e7-724a-41b6-b206-8103d83b0b46"
      },
      "execution_count": 53,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7abbfa4961a0>"
            ]
          },
          "metadata": {},
          "execution_count": 53
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate the model\n",
        "loss, mae = model.evaluate(\n",
        "    [X_test[\"day_of_week\"], X_test[[\"temp\", \"humidity\", \"windspeed\"]]],\n",
        "    y_test\n",
        ")\n",
        "\n",
        "print(f'Test Loss (MSE): {loss:.4f}')\n",
        "print(f'Test MAE: {mae:.4f}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qQD2PLMR8VwY",
        "outputId": "d52b0fb8-8c92-47af-8728-8b6a5265e0b6"
      },
      "execution_count": 59,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "69/69 [==============================] - 0s 2ms/step - loss: 0.0244 - mae: 0.1157\n",
            "Test Loss (MSE): 0.0244\n",
            "Test MAE: 0.1157\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "result = model.predict([X_test[\"day_of_week\"], X_test[[\"temp\", \"humidity\", \"windspeed\"]]])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0zqYKwV59TUp",
        "outputId": "c2ea14df-1103-4e3d-fc0c-63964ac18d19"
      },
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "69/69 [==============================] - 0s 2ms/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "[print(round(list(y_test)[i]*m) , round(result[i][0]*m)) for i in range(20)];"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dNMmb97Z9gvP",
        "outputId": "0d149bde-1a07-48d5-a85f-093e1f45fb2e"
      },
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "127 269\n",
            "13 66\n",
            "163 94\n",
            "233 270\n",
            "222 274\n",
            "166 153\n",
            "144 184\n",
            "376 370\n",
            "601 138\n",
            "53 108\n",
            "6 53\n",
            "14 176\n",
            "209 277\n",
            "62 101\n",
            "139 343\n",
            "9 121\n",
            "96 160\n",
            "98 237\n",
            "230 148\n",
            "12 216\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Lets Visualize!"
      ],
      "metadata": {
        "id": "thvGd2sJG-o4"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Extract embeddings from the trained model\n",
        "embedding_layer = model.layers[1]  # Assuming the embedding layer is the 2nd layer in your model\n",
        "embedding_weights = embedding_layer.get_weights()[0]\n",
        "embedding_weights.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TiVKgfZwCA6F",
        "outputId": "4f42df42-252f-4fe4-8c1c-47f1ba6e9714"
      },
      "execution_count": 62,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(7, 3)"
            ]
          },
          "metadata": {},
          "execution_count": 62
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.decomposition import PCA\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "\n",
        "# Apply PCA to reduce embeddings to 2D space\n",
        "pca = PCA(n_components=2)\n",
        "pca_embeddings = pca.fit_transform(embedding_weights)\n",
        "\n",
        "# Day labels\n",
        "day_labels = [\"Mon\", \"Tue\", \"Wed\", \"Thu\", \"Fri\", \"Sat\", \"Sun\"]\n",
        "\n",
        "# Create a scatter plot with text labels\n",
        "plt.figure(figsize=(8, 6))\n",
        "for i, (x, y) in enumerate(pca_embeddings):\n",
        "    plt.scatter(x, y, s=100)\n",
        "    plt.annotate(day_labels[i], (x, y), textcoords=\"offset points\", xytext=(0, 10), ha=\"center\")\n",
        "\n",
        "plt.title(\"PCA Visualization of Day of Week Embeddings\")\n",
        "plt.xlabel(\"PCA Dimension 1\")\n",
        "plt.ylabel(\"PCA Dimension 2\")\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 564
        },
        "id": "OvktaLYCEM2F",
        "outputId": "b4f06441-215c-46d7-ee11-daed368c2049"
      },
      "execution_count": 63,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 800x600 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "https://class.vision"
      ],
      "metadata": {
        "id": "A6YWW8_4HGiQ"
      }
    }
  ]
}