Train.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 384.0
    },
    "colab_type": "code",
    "id": "7736EJluandP",
    "outputId": "bd1dd0cb-d36b-4e67-e647-0168f858072a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(100, 100, 3) 40 43\n",
      "(100, 100) 40 43\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import cv2\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import os\n",
    "\n",
    "data_path = 'data'\n",
    "\n",
    "data_set = np.load(os.path.join(data_path, 'array.npy'))\n",
    "\n",
    "scale = 1\n",
    "\n",
    "print(data_set[0][0].shape, data_set[0][1], data_set[0][2])\n",
    "\n",
    "for im in data_set:\n",
    "    im[0] = cv2.resize(im[0], None, None, scale, scale)\n",
    "    im[0] = cv2.cvtColor(im[0], cv2.COLOR_BGR2GRAY)\n",
    "    im[0] = cv2.equalizeHist(im[0])\n",
    "    im[1] = int(im[1] * scale)\n",
    "    im[2] = int(im[2] * scale)\n",
    "\n",
    "plt.imshow(data_set[0][0])\n",
    "print(data_set[0][0].shape, data_set[0][1], data_set[0][2])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51.0
    },
    "colab_type": "code",
    "id": "WCUv4Dzqap1Y",
    "outputId": "3dca1139-743e-49ed-fca7-f3f80135478c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2013 (2013, 100, 100, 1)\n",
      "2013 (2013, 4)\n"
     ]
    }
   ],
   "source": [
    "X_ = data_set[:, 0:1]\n",
    "Y_ = data_set[:, 1:5]\n",
    "\n",
    "Xx = []\n",
    "Y = []\n",
    "\n",
    "for x in X_:\n",
    "    Xx.append(x[0] * 1. / 255)\n",
    "\n",
    "Xx = np.array(Xx)\n",
    "Y = np.array(Y_)\n",
    "\n",
    "Y *= 1. / (100 * scale)\n",
    "\n",
    "X = Xx[..., np.newaxis]\n",
    "\n",
    "print(len(X), X.shape)\n",
    "print(len(Y), Y.shape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34.0
    },
    "colab_type": "code",
    "id": "MWxfN78Ncwl0",
    "outputId": "781317ba-33d5-4a42-f2f0-f81d6cbf965f"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Bartosz\\Anaconda3\\lib\\site-packages\\h5py\\__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  from ._conv import register_converters as _register_converters\n",
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Conv2D, MaxPooling2D, Flatten, Activation, Dense, Dropout\n",
    "np.random.seed(58)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 680.0
    },
    "colab_type": "code",
    "id": "tqAQ2f9hc1OM",
    "outputId": "66748209-ce5d-49d5-a705-9f1c65e7f327"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_1 (Conv2D)            (None, 98, 98, 128)       1280      \n",
      "_________________________________________________________________\n",
      "activation_1 (Activation)    (None, 98, 98, 128)       0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 49, 49, 128)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_2 (Conv2D)            (None, 49, 49, 256)       295168    \n",
      "_________________________________________________________________\n",
      "activation_2 (Activation)    (None, 49, 49, 256)       0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2 (None, 24, 24, 256)       0         \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 24, 24, 256)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_3 (Conv2D)            (None, 24, 24, 256)       590080    \n",
      "_________________________________________________________________\n",
      "activation_3 (Activation)    (None, 24, 24, 256)       0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_3 (MaxPooling2 (None, 12, 12, 256)       0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 36864)             0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 512)               18874880  \n",
      "_________________________________________________________________\n",
      "activation_4 (Activation)    (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 4)                 2052      \n",
      "_________________________________________________________________\n",
      "activation_5 (Activation)    (None, 4)                 0         \n",
      "=================================================================\n",
      "Total params: 19,763,460\n",
      "Trainable params: 19,763,460\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "\n",
    "model.add(Conv2D(128, (3, 3), input_shape=(int(100 * scale), int(100 * scale), 1)))\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "\n",
    "model.add(Conv2D(256, (3, 3), padding='same'))\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "model.add(Dropout(0.25))\n",
    "\n",
    "model.add(Conv2D(256, (3, 3), padding='same'))\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "\n",
    "model.add(Flatten())\n",
    "model.add(Dense(512))\n",
    "model.add(Activation('relu'))\n",
    "model.add(Dropout(0.5))\n",
    "\n",
    "model.add(Dense(4))\n",
    "model.add(Activation('linear'))\n",
    "\n",
    "model.summary()\n",
    "model.compile(loss='mean_absolute_error', optimizer='adam', metrics=['mean_absolute_error'])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 6851.0
    },
    "colab_type": "code",
    "id": "0ayhpum5dBul",
    "outputId": "4e691598-1e2d-4785-9165-0af1aeb50065"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 1610 samples, validate on 403 samples\n",
      "Epoch 1/200\n",
      "1610/1610 [==============================] - 8s 5ms/step - loss: 0.1922 - mean_absolute_error: 0.1922 - val_loss: 0.0918 - val_mean_absolute_error: 0.0918\n",
      "Epoch 2/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0916 - mean_absolute_error: 0.0916 - val_loss: 0.0756 - val_mean_absolute_error: 0.0756\n",
      "Epoch 3/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0802 - mean_absolute_error: 0.0802 - val_loss: 0.1018 - val_mean_absolute_error: 0.1018\n",
      "Epoch 4/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0742 - mean_absolute_error: 0.0742 - val_loss: 0.0767 - val_mean_absolute_error: 0.0767\n",
      "Epoch 5/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0711 - mean_absolute_error: 0.0711 - val_loss: 0.0829 - val_mean_absolute_error: 0.0829\n",
      "Epoch 6/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0664 - mean_absolute_error: 0.0664 - val_loss: 0.0758 - val_mean_absolute_error: 0.0758\n",
      "Epoch 7/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0636 - mean_absolute_error: 0.0636 - val_loss: 0.0811 - val_mean_absolute_error: 0.0811\n",
      "Epoch 8/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0608 - mean_absolute_error: 0.0608 - val_loss: 0.0688 - val_mean_absolute_error: 0.0688\n",
      "Epoch 9/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0587 - mean_absolute_error: 0.0587 - val_loss: 0.0578 - val_mean_absolute_error: 0.0578\n",
      "Epoch 10/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0585 - mean_absolute_error: 0.0585 - val_loss: 0.0526 - val_mean_absolute_error: 0.0526\n",
      "Epoch 11/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0598 - mean_absolute_error: 0.0598 - val_loss: 0.0702 - val_mean_absolute_error: 0.0702\n",
      "Epoch 12/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0556 - mean_absolute_error: 0.0556 - val_loss: 0.0434 - val_mean_absolute_error: 0.0434\n",
      "Epoch 13/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0544 - mean_absolute_error: 0.0544 - val_loss: 0.0604 - val_mean_absolute_error: 0.0604\n",
      "Epoch 14/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0576 - mean_absolute_error: 0.0576 - val_loss: 0.0548 - val_mean_absolute_error: 0.0548\n",
      "Epoch 15/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0562 - mean_absolute_error: 0.0562 - val_loss: 0.0432 - val_mean_absolute_error: 0.0432\n",
      "Epoch 16/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0527 - mean_absolute_error: 0.0527 - val_loss: 0.0451 - val_mean_absolute_error: 0.0451\n",
      "Epoch 17/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0530 - mean_absolute_error: 0.0530 - val_loss: 0.0378 - val_mean_absolute_error: 0.0378\n",
      "Epoch 18/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0505 - mean_absolute_error: 0.0505 - val_loss: 0.0468 - val_mean_absolute_error: 0.0468\n",
      "Epoch 19/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0498 - mean_absolute_error: 0.0498 - val_loss: 0.0436 - val_mean_absolute_error: 0.0436\n",
      "Epoch 20/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0495 - mean_absolute_error: 0.0495 - val_loss: 0.0437 - val_mean_absolute_error: 0.0437\n",
      "Epoch 21/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0482 - mean_absolute_error: 0.0482 - val_loss: 0.0478 - val_mean_absolute_error: 0.0478\n",
      "Epoch 22/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0488 - mean_absolute_error: 0.0488 - val_loss: 0.0397 - val_mean_absolute_error: 0.0397\n",
      "Epoch 23/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0470 - mean_absolute_error: 0.0470 - val_loss: 0.0399 - val_mean_absolute_error: 0.0399\n",
      "Epoch 24/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0455 - mean_absolute_error: 0.0455 - val_loss: 0.0390 - val_mean_absolute_error: 0.0390\n",
      "Epoch 25/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0451 - mean_absolute_error: 0.0451 - val_loss: 0.0393 - val_mean_absolute_error: 0.0393\n",
      "Epoch 26/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0453 - mean_absolute_error: 0.0453 - val_loss: 0.0383 - val_mean_absolute_error: 0.0383\n",
      "Epoch 27/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0438 - mean_absolute_error: 0.0438 - val_loss: 0.0420 - val_mean_absolute_error: 0.0420\n",
      "Epoch 28/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0444 - mean_absolute_error: 0.0444 - val_loss: 0.0360 - val_mean_absolute_error: 0.0360\n",
      "Epoch 29/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0437 - mean_absolute_error: 0.0437 - val_loss: 0.0371 - val_mean_absolute_error: 0.0371\n",
      "Epoch 30/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0441 - mean_absolute_error: 0.0441 - val_loss: 0.0390 - val_mean_absolute_error: 0.0390\n",
      "Epoch 31/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0423 - mean_absolute_error: 0.0423 - val_loss: 0.0376 - val_mean_absolute_error: 0.0376\n",
      "Epoch 32/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0423 - mean_absolute_error: 0.0423 - val_loss: 0.0401 - val_mean_absolute_error: 0.0401\n",
      "Epoch 33/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0406 - mean_absolute_error: 0.0406 - val_loss: 0.0390 - val_mean_absolute_error: 0.0390\n",
      "Epoch 34/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0396 - mean_absolute_error: 0.0396 - val_loss: 0.0376 - val_mean_absolute_error: 0.0376\n",
      "Epoch 35/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0397 - mean_absolute_error: 0.0397 - val_loss: 0.0387 - val_mean_absolute_error: 0.0387\n",
      "Epoch 36/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0393 - mean_absolute_error: 0.0393 - val_loss: 0.0422 - val_mean_absolute_error: 0.0422\n",
      "Epoch 37/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0385 - mean_absolute_error: 0.0385 - val_loss: 0.0384 - val_mean_absolute_error: 0.0384\n",
      "Epoch 38/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0372 - mean_absolute_error: 0.0372 - val_loss: 0.0394 - val_mean_absolute_error: 0.0394\n",
      "Epoch 39/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0375 - mean_absolute_error: 0.0375 - val_loss: 0.0399 - val_mean_absolute_error: 0.0399\n",
      "Epoch 40/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0365 - mean_absolute_error: 0.0365 - val_loss: 0.0374 - val_mean_absolute_error: 0.0374\n",
      "Epoch 41/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0373 - mean_absolute_error: 0.0373 - val_loss: 0.0392 - val_mean_absolute_error: 0.0392\n",
      "Epoch 42/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0358 - mean_absolute_error: 0.0358 - val_loss: 0.0366 - val_mean_absolute_error: 0.0366\n",
      "Epoch 43/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0352 - mean_absolute_error: 0.0352 - val_loss: 0.0379 - val_mean_absolute_error: 0.0379\n",
      "Epoch 44/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0345 - mean_absolute_error: 0.0345 - val_loss: 0.0396 - val_mean_absolute_error: 0.0396\n",
      "Epoch 45/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0349 - mean_absolute_error: 0.0349 - val_loss: 0.0347 - val_mean_absolute_error: 0.0347\n",
      "Epoch 46/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0336 - mean_absolute_error: 0.0336 - val_loss: 0.0369 - val_mean_absolute_error: 0.0369\n",
      "Epoch 47/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0332 - mean_absolute_error: 0.0332 - val_loss: 0.0382 - val_mean_absolute_error: 0.0382\n",
      "Epoch 48/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0331 - mean_absolute_error: 0.0331 - val_loss: 0.0363 - val_mean_absolute_error: 0.0363\n",
      "Epoch 49/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0325 - mean_absolute_error: 0.0325 - val_loss: 0.0366 - val_mean_absolute_error: 0.0366\n",
      "Epoch 50/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0324 - mean_absolute_error: 0.0324 - val_loss: 0.0365 - val_mean_absolute_error: 0.0365\n",
      "Epoch 51/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0316 - mean_absolute_error: 0.0316 - val_loss: 0.0353 - val_mean_absolute_error: 0.0353\n",
      "Epoch 52/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0314 - mean_absolute_error: 0.0314 - val_loss: 0.0371 - val_mean_absolute_error: 0.0371\n",
      "Epoch 53/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0304 - mean_absolute_error: 0.0304 - val_loss: 0.0366 - val_mean_absolute_error: 0.0366\n",
      "Epoch 54/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0303 - mean_absolute_error: 0.0303 - val_loss: 0.0370 - val_mean_absolute_error: 0.0370\n",
      "Epoch 55/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0302 - mean_absolute_error: 0.0302 - val_loss: 0.0351 - val_mean_absolute_error: 0.0351\n",
      "Epoch 56/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0296 - mean_absolute_error: 0.0296 - val_loss: 0.0364 - val_mean_absolute_error: 0.0364\n",
      "Epoch 57/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0294 - mean_absolute_error: 0.0294 - val_loss: 0.0351 - val_mean_absolute_error: 0.0351\n",
      "Epoch 58/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0296 - mean_absolute_error: 0.0296 - val_loss: 0.0368 - val_mean_absolute_error: 0.0368\n",
      "Epoch 59/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0287 - mean_absolute_error: 0.0287 - val_loss: 0.0359 - val_mean_absolute_error: 0.0359\n",
      "Epoch 60/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0289 - mean_absolute_error: 0.0289 - val_loss: 0.0372 - val_mean_absolute_error: 0.0372\n",
      "Epoch 61/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0286 - mean_absolute_error: 0.0286 - val_loss: 0.0368 - val_mean_absolute_error: 0.0368\n",
      "Epoch 62/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0283 - mean_absolute_error: 0.0283 - val_loss: 0.0368 - val_mean_absolute_error: 0.0368\n",
      "Epoch 63/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0279 - mean_absolute_error: 0.0279 - val_loss: 0.0371 - val_mean_absolute_error: 0.0371\n",
      "Epoch 64/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0276 - mean_absolute_error: 0.0276 - val_loss: 0.0367 - val_mean_absolute_error: 0.0367\n",
      "Epoch 65/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0275 - mean_absolute_error: 0.0275 - val_loss: 0.0380 - val_mean_absolute_error: 0.0380\n",
      "Epoch 66/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0270 - mean_absolute_error: 0.0270 - val_loss: 0.0354 - val_mean_absolute_error: 0.0354\n",
      "Epoch 67/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0273 - mean_absolute_error: 0.0273 - val_loss: 0.0353 - val_mean_absolute_error: 0.0353\n",
      "Epoch 68/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0269 - mean_absolute_error: 0.0269 - val_loss: 0.0353 - val_mean_absolute_error: 0.0353\n",
      "Epoch 69/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0264 - mean_absolute_error: 0.0264 - val_loss: 0.0355 - val_mean_absolute_error: 0.0355\n",
      "Epoch 70/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0263 - mean_absolute_error: 0.0263 - val_loss: 0.0355 - val_mean_absolute_error: 0.0355\n",
      "Epoch 71/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0265 - mean_absolute_error: 0.0265 - val_loss: 0.0362 - val_mean_absolute_error: 0.0362\n",
      "Epoch 72/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0266 - mean_absolute_error: 0.0266 - val_loss: 0.0355 - val_mean_absolute_error: 0.0355\n",
      "Epoch 73/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0263 - mean_absolute_error: 0.0263 - val_loss: 0.0361 - val_mean_absolute_error: 0.0361\n",
      "Epoch 74/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0261 - mean_absolute_error: 0.0261 - val_loss: 0.0377 - val_mean_absolute_error: 0.0377\n",
      "Epoch 75/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0261 - mean_absolute_error: 0.0261 - val_loss: 0.0391 - val_mean_absolute_error: 0.0391\n",
      "Epoch 76/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0254 - mean_absolute_error: 0.0254 - val_loss: 0.0363 - val_mean_absolute_error: 0.0363\n",
      "Epoch 77/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0258 - mean_absolute_error: 0.0258 - val_loss: 0.0352 - val_mean_absolute_error: 0.0352\n",
      "Epoch 78/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0253 - mean_absolute_error: 0.0253 - val_loss: 0.0359 - val_mean_absolute_error: 0.0359\n",
      "Epoch 79/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0250 - mean_absolute_error: 0.0250 - val_loss: 0.0358 - val_mean_absolute_error: 0.0358\n",
      "Epoch 80/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0250 - mean_absolute_error: 0.0250 - val_loss: 0.0369 - val_mean_absolute_error: 0.0369\n",
      "Epoch 81/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0247 - mean_absolute_error: 0.0247 - val_loss: 0.0390 - val_mean_absolute_error: 0.0390\n",
      "Epoch 82/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0252 - mean_absolute_error: 0.0252 - val_loss: 0.0362 - val_mean_absolute_error: 0.0362\n",
      "Epoch 83/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0248 - mean_absolute_error: 0.0248 - val_loss: 0.0351 - val_mean_absolute_error: 0.0351\n",
      "Epoch 84/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0247 - mean_absolute_error: 0.0247 - val_loss: 0.0375 - val_mean_absolute_error: 0.0375\n",
      "Epoch 85/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0244 - mean_absolute_error: 0.0244 - val_loss: 0.0350 - val_mean_absolute_error: 0.0350\n",
      "Epoch 86/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0244 - mean_absolute_error: 0.0244 - val_loss: 0.0348 - val_mean_absolute_error: 0.0348\n",
      "Epoch 87/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0248 - mean_absolute_error: 0.0248 - val_loss: 0.0353 - val_mean_absolute_error: 0.0353\n",
      "Epoch 88/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0242 - mean_absolute_error: 0.0242 - val_loss: 0.0364 - val_mean_absolute_error: 0.0364\n",
      "Epoch 89/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0249 - mean_absolute_error: 0.0249 - val_loss: 0.0386 - val_mean_absolute_error: 0.0386\n",
      "Epoch 90/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0242 - mean_absolute_error: 0.0242 - val_loss: 0.0356 - val_mean_absolute_error: 0.0356\n",
      "Epoch 91/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0244 - mean_absolute_error: 0.0244 - val_loss: 0.0353 - val_mean_absolute_error: 0.0353\n",
      "Epoch 92/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0240 - mean_absolute_error: 0.0240 - val_loss: 0.0376 - val_mean_absolute_error: 0.0376\n",
      "Epoch 93/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0242 - mean_absolute_error: 0.0242 - val_loss: 0.0344 - val_mean_absolute_error: 0.0344\n",
      "Epoch 94/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0242 - mean_absolute_error: 0.0242 - val_loss: 0.0386 - val_mean_absolute_error: 0.0386\n",
      "Epoch 95/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0240 - mean_absolute_error: 0.0240 - val_loss: 0.0366 - val_mean_absolute_error: 0.0366\n",
      "Epoch 96/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0235 - mean_absolute_error: 0.0235 - val_loss: 0.0362 - val_mean_absolute_error: 0.0362\n",
      "Epoch 97/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0240 - mean_absolute_error: 0.0240 - val_loss: 0.0376 - val_mean_absolute_error: 0.0376\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 98/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0234 - mean_absolute_error: 0.0234 - val_loss: 0.0370 - val_mean_absolute_error: 0.0370\n",
      "Epoch 99/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0239 - mean_absolute_error: 0.0239 - val_loss: 0.0352 - val_mean_absolute_error: 0.0352\n",
      "Epoch 100/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0232 - mean_absolute_error: 0.0232 - val_loss: 0.0381 - val_mean_absolute_error: 0.0381\n",
      "Epoch 101/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0239 - mean_absolute_error: 0.0239 - val_loss: 0.0347 - val_mean_absolute_error: 0.0347\n",
      "Epoch 102/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0236 - mean_absolute_error: 0.0236 - val_loss: 0.0378 - val_mean_absolute_error: 0.0378\n",
      "Epoch 103/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0231 - mean_absolute_error: 0.0231 - val_loss: 0.0363 - val_mean_absolute_error: 0.0363\n",
      "Epoch 104/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0231 - mean_absolute_error: 0.0231 - val_loss: 0.0367 - val_mean_absolute_error: 0.0367\n",
      "Epoch 105/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0235 - mean_absolute_error: 0.0235 - val_loss: 0.0346 - val_mean_absolute_error: 0.0346\n",
      "Epoch 106/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0229 - mean_absolute_error: 0.0229 - val_loss: 0.0351 - val_mean_absolute_error: 0.0351\n",
      "Epoch 107/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0232 - mean_absolute_error: 0.0232 - val_loss: 0.0360 - val_mean_absolute_error: 0.0360\n",
      "Epoch 108/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0231 - mean_absolute_error: 0.0231 - val_loss: 0.0361 - val_mean_absolute_error: 0.0361\n",
      "Epoch 109/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0230 - mean_absolute_error: 0.0230 - val_loss: 0.0350 - val_mean_absolute_error: 0.0350\n",
      "Epoch 110/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0224 - mean_absolute_error: 0.0224 - val_loss: 0.0350 - val_mean_absolute_error: 0.0350\n",
      "Epoch 111/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0231 - mean_absolute_error: 0.0231 - val_loss: 0.0342 - val_mean_absolute_error: 0.0342\n",
      "Epoch 112/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0228 - mean_absolute_error: 0.0228 - val_loss: 0.0383 - val_mean_absolute_error: 0.0383\n",
      "Epoch 113/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0226 - mean_absolute_error: 0.0226 - val_loss: 0.0356 - val_mean_absolute_error: 0.0356\n",
      "Epoch 114/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0226 - mean_absolute_error: 0.0226 - val_loss: 0.0360 - val_mean_absolute_error: 0.0360\n",
      "Epoch 115/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0224 - mean_absolute_error: 0.0224 - val_loss: 0.0374 - val_mean_absolute_error: 0.0374\n",
      "Epoch 116/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0233 - mean_absolute_error: 0.0233 - val_loss: 0.0364 - val_mean_absolute_error: 0.0364\n",
      "Epoch 117/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0229 - mean_absolute_error: 0.0229 - val_loss: 0.0363 - val_mean_absolute_error: 0.0363\n",
      "Epoch 118/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0226 - mean_absolute_error: 0.0226 - val_loss: 0.0367 - val_mean_absolute_error: 0.0367\n",
      "Epoch 119/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0226 - mean_absolute_error: 0.0226 - val_loss: 0.0377 - val_mean_absolute_error: 0.0377\n",
      "Epoch 120/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0222 - mean_absolute_error: 0.0222 - val_loss: 0.0364 - val_mean_absolute_error: 0.0364\n",
      "Epoch 121/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0223 - mean_absolute_error: 0.0223 - val_loss: 0.0354 - val_mean_absolute_error: 0.0354\n",
      "Epoch 122/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0223 - mean_absolute_error: 0.0223 - val_loss: 0.0375 - val_mean_absolute_error: 0.0375\n",
      "Epoch 123/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0222 - mean_absolute_error: 0.0222 - val_loss: 0.0361 - val_mean_absolute_error: 0.0361\n",
      "Epoch 124/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0224 - mean_absolute_error: 0.0224 - val_loss: 0.0356 - val_mean_absolute_error: 0.0356\n",
      "Epoch 125/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0219 - mean_absolute_error: 0.0219 - val_loss: 0.0368 - val_mean_absolute_error: 0.0368\n",
      "Epoch 126/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0222 - mean_absolute_error: 0.0222 - val_loss: 0.0364 - val_mean_absolute_error: 0.0364\n",
      "Epoch 127/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0221 - mean_absolute_error: 0.0221 - val_loss: 0.0356 - val_mean_absolute_error: 0.0356\n",
      "Epoch 128/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0224 - mean_absolute_error: 0.0224 - val_loss: 0.0388 - val_mean_absolute_error: 0.0388\n",
      "Epoch 129/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0221 - mean_absolute_error: 0.0221 - val_loss: 0.0369 - val_mean_absolute_error: 0.0369\n",
      "Epoch 130/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0220 - mean_absolute_error: 0.0220 - val_loss: 0.0360 - val_mean_absolute_error: 0.0360\n",
      "Epoch 131/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0216 - mean_absolute_error: 0.0216 - val_loss: 0.0357 - val_mean_absolute_error: 0.0357\n",
      "Epoch 132/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0219 - mean_absolute_error: 0.0219 - val_loss: 0.0354 - val_mean_absolute_error: 0.0354\n",
      "Epoch 133/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0220 - mean_absolute_error: 0.0220 - val_loss: 0.0363 - val_mean_absolute_error: 0.0363\n",
      "Epoch 134/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0217 - mean_absolute_error: 0.0217 - val_loss: 0.0369 - val_mean_absolute_error: 0.0369\n",
      "Epoch 135/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0218 - mean_absolute_error: 0.0218 - val_loss: 0.0362 - val_mean_absolute_error: 0.0362\n",
      "Epoch 136/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0217 - mean_absolute_error: 0.0217 - val_loss: 0.0353 - val_mean_absolute_error: 0.0353\n",
      "Epoch 137/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0212 - mean_absolute_error: 0.0212 - val_loss: 0.0349 - val_mean_absolute_error: 0.0349\n",
      "Epoch 138/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0220 - mean_absolute_error: 0.0220 - val_loss: 0.0366 - val_mean_absolute_error: 0.0366\n",
      "Epoch 139/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0216 - mean_absolute_error: 0.0216 - val_loss: 0.0362 - val_mean_absolute_error: 0.0362\n",
      "Epoch 140/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0217 - mean_absolute_error: 0.0217 - val_loss: 0.0372 - val_mean_absolute_error: 0.0372\n",
      "Epoch 141/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0216 - mean_absolute_error: 0.0216 - val_loss: 0.0358 - val_mean_absolute_error: 0.0358\n",
      "Epoch 142/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0218 - mean_absolute_error: 0.0218 - val_loss: 0.0366 - val_mean_absolute_error: 0.036618 - mean_absolute_error: 0.02\n",
      "Epoch 143/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0215 - mean_absolute_error: 0.0215 - val_loss: 0.0358 - val_mean_absolute_error: 0.0358\n",
      "Epoch 144/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0213 - mean_absolute_error: 0.0213 - val_loss: 0.0362 - val_mean_absolute_error: 0.0362\n",
      "Epoch 145/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0212 - mean_absolute_error: 0.0212 - val_loss: 0.0356 - val_mean_absolute_error: 0.0356\n",
      "Epoch 146/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0213 - mean_absolute_error: 0.0213 - val_loss: 0.0368 - val_mean_absolute_error: 0.0368\n",
      "Epoch 147/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0216 - mean_absolute_error: 0.0216 - val_loss: 0.0366 - val_mean_absolute_error: 0.0366\n",
      "Epoch 148/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0212 - mean_absolute_error: 0.0212 - val_loss: 0.0376 - val_mean_absolute_error: 0.0376\n",
      "Epoch 149/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0206 - mean_absolute_error: 0.0206 - val_loss: 0.0369 - val_mean_absolute_error: 0.0369\n",
      "Epoch 150/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0211 - mean_absolute_error: 0.0211 - val_loss: 0.0361 - val_mean_absolute_error: 0.0361\n",
      "Epoch 151/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0213 - mean_absolute_error: 0.0213 - val_loss: 0.0354 - val_mean_absolute_error: 0.0354\n",
      "Epoch 152/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0210 - mean_absolute_error: 0.0210 - val_loss: 0.0376 - val_mean_absolute_error: 0.0376\n",
      "Epoch 153/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0206 - mean_absolute_error: 0.0206 - val_loss: 0.0363 - val_mean_absolute_error: 0.0363\n",
      "Epoch 154/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0211 - mean_absolute_error: 0.0211 - val_loss: 0.0368 - val_mean_absolute_error: 0.0368\n",
      "Epoch 155/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0209 - mean_absolute_error: 0.0209 - val_loss: 0.0393 - val_mean_absolute_error: 0.0393\n",
      "Epoch 156/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0208 - mean_absolute_error: 0.0208 - val_loss: 0.0359 - val_mean_absolute_error: 0.0359\n",
      "Epoch 157/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0208 - mean_absolute_error: 0.0208 - val_loss: 0.0387 - val_mean_absolute_error: 0.0387\n",
      "Epoch 158/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0211 - mean_absolute_error: 0.0211 - val_loss: 0.0368 - val_mean_absolute_error: 0.0368\n",
      "Epoch 159/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0208 - mean_absolute_error: 0.0208 - val_loss: 0.0374 - val_mean_absolute_error: 0.0374\n",
      "Epoch 160/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0209 - mean_absolute_error: 0.0209 - val_loss: 0.0374 - val_mean_absolute_error: 0.0374\n",
      "Epoch 161/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0208 - mean_absolute_error: 0.0208 - val_loss: 0.0381 - val_mean_absolute_error: 0.0381\n",
      "Epoch 162/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0205 - mean_absolute_error: 0.0205 - val_loss: 0.0362 - val_mean_absolute_error: 0.0362\n",
      "Epoch 163/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0205 - mean_absolute_error: 0.0205 - val_loss: 0.0363 - val_mean_absolute_error: 0.0363\n",
      "Epoch 164/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0208 - mean_absolute_error: 0.0208 - val_loss: 0.0359 - val_mean_absolute_error: 0.0359\n",
      "Epoch 165/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0206 - mean_absolute_error: 0.0206 - val_loss: 0.0384 - val_mean_absolute_error: 0.0384\n",
      "Epoch 166/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0207 - mean_absolute_error: 0.0207 - val_loss: 0.0376 - val_mean_absolute_error: 0.0376\n",
      "Epoch 167/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0207 - mean_absolute_error: 0.0207 - val_loss: 0.0365 - val_mean_absolute_error: 0.0365\n",
      "Epoch 168/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0207 - mean_absolute_error: 0.0207 - val_loss: 0.0370 - val_mean_absolute_error: 0.0370\n",
      "Epoch 169/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0204 - mean_absolute_error: 0.0204 - val_loss: 0.0357 - val_mean_absolute_error: 0.0357\n",
      "Epoch 170/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0201 - mean_absolute_error: 0.0201 - val_loss: 0.0359 - val_mean_absolute_error: 0.0359\n",
      "Epoch 171/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0204 - mean_absolute_error: 0.0204 - val_loss: 0.0370 - val_mean_absolute_error: 0.0370\n",
      "Epoch 172/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0204 - mean_absolute_error: 0.0204 - val_loss: 0.0366 - val_mean_absolute_error: 0.0366\n",
      "Epoch 173/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0203 - mean_absolute_error: 0.0203 - val_loss: 0.0383 - val_mean_absolute_error: 0.0383\n",
      "Epoch 174/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0205 - mean_absolute_error: 0.0205 - val_loss: 0.0371 - val_mean_absolute_error: 0.0371\n",
      "Epoch 175/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0203 - mean_absolute_error: 0.0203 - val_loss: 0.0386 - val_mean_absolute_error: 0.0386\n",
      "Epoch 176/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0198 - mean_absolute_error: 0.0198 - val_loss: 0.0381 - val_mean_absolute_error: 0.0381\n",
      "Epoch 177/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0202 - mean_absolute_error: 0.0202 - val_loss: 0.0358 - val_mean_absolute_error: 0.0358\n",
      "Epoch 178/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0199 - mean_absolute_error: 0.0199 - val_loss: 0.0361 - val_mean_absolute_error: 0.0361\n",
      "Epoch 179/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0203 - mean_absolute_error: 0.0203 - val_loss: 0.0362 - val_mean_absolute_error: 0.0362\n",
      "Epoch 180/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0203 - mean_absolute_error: 0.0203 - val_loss: 0.0370 - val_mean_absolute_error: 0.0370\n",
      "Epoch 181/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0203 - mean_absolute_error: 0.0203 - val_loss: 0.0362 - val_mean_absolute_error: 0.0362\n",
      "Epoch 182/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0201 - mean_absolute_error: 0.0201 - val_loss: 0.0371 - val_mean_absolute_error: 0.0371\n",
      "Epoch 183/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0197 - mean_absolute_error: 0.0197 - val_loss: 0.0361 - val_mean_absolute_error: 0.0361\n",
      "Epoch 184/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0199 - mean_absolute_error: 0.0199 - val_loss: 0.0381 - val_mean_absolute_error: 0.0381\n",
      "Epoch 185/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0195 - mean_absolute_error: 0.0195 - val_loss: 0.0377 - val_mean_absolute_error: 0.0377\n",
      "Epoch 186/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0197 - mean_absolute_error: 0.0197 - val_loss: 0.0380 - val_mean_absolute_error: 0.0380\n",
      "Epoch 187/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0197 - mean_absolute_error: 0.0197 - val_loss: 0.0370 - val_mean_absolute_error: 0.0370\n",
      "Epoch 188/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0200 - mean_absolute_error: 0.0200 - val_loss: 0.0360 - val_mean_absolute_error: 0.0360\n",
      "Epoch 189/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0196 - mean_absolute_error: 0.0196 - val_loss: 0.0377 - val_mean_absolute_error: 0.0377\n",
      "Epoch 190/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0198 - mean_absolute_error: 0.0198 - val_loss: 0.0371 - val_mean_absolute_error: 0.0371\n",
      "Epoch 191/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0197 - mean_absolute_error: 0.0197 - val_loss: 0.0365 - val_mean_absolute_error: 0.0365\n",
      "Epoch 192/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0198 - mean_absolute_error: 0.0198 - val_loss: 0.0370 - val_mean_absolute_error: 0.0370\n",
      "Epoch 193/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0197 - mean_absolute_error: 0.0197 - val_loss: 0.0383 - val_mean_absolute_error: 0.0383\n",
      "Epoch 194/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0200 - mean_absolute_error: 0.0200 - val_loss: 0.0376 - val_mean_absolute_error: 0.0376\n",
      "Epoch 195/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0200 - mean_absolute_error: 0.0200 - val_loss: 0.0381 - val_mean_absolute_error: 0.0381\n",
      "Epoch 196/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0196 - mean_absolute_error: 0.0196 - val_loss: 0.0367 - val_mean_absolute_error: 0.0367\n",
      "Epoch 197/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0195 - mean_absolute_error: 0.0195 - val_loss: 0.0372 - val_mean_absolute_error: 0.0372\n",
      "Epoch 198/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0203 - mean_absolute_error: 0.0203 - val_loss: 0.0361 - val_mean_absolute_error: 0.0361\n",
      "Epoch 199/200\n",
      "1610/1610 [==============================] - 5s 3ms/step - loss: 0.0196 - mean_absolute_error: 0.0196 - val_loss: 0.0373 - val_mean_absolute_error: 0.0373\n",
      "Epoch 200/200\n",
      "1610/1610 [==============================] - 4s 3ms/step - loss: 0.0198 - mean_absolute_error: 0.0198 - val_loss: 0.0383 - val_mean_absolute_error: 0.0383\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x27791d435f8>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(X, Y, epochs=200, batch_size=32, shuffle=True, validation_split=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51.0
    },
    "colab_type": "code",
    "id": "jlw4dsUCrLCy",
    "outputId": "63cc9c2c-15c9-453b-beed-117786c95877"
   },
   "outputs": [],
   "source": [
    "model.save(os.path.join(data_path, 'model.h5'))\n",
    "model.save_weights(os.path.join(data_path, 'weights.h5'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "Train.ipynb",
   "provenance": [],
   "version": "0.3.2"
  },
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}