1D_DenseNet-kfoldcrossvalidation-Copy1.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 128
    },
    "colab_type": "code",
    "id": "pf32MK1QQp6b",
    "outputId": "78d0297e-b627-4a59-e46e-69633420c5b5"
   },
   "outputs": [],
   "source": [
    "from google.colab import drive\n",
    "drive.mount('/googledrive', force_remount=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 846
    },
    "colab_type": "code",
    "id": "wXz2orRxQsB6",
    "outputId": "1a2c1451-1f71-4354-834c-79e381a01f4e"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from __future__ import print_function\n",
    "import keras\n",
    "from keras.layers import Dense, Conv2D, BatchNormalization, Activation\n",
    "from keras.layers import AveragePooling2D, Input, Flatten\n",
    "from keras.optimizers import Adam, SGD\n",
    "from keras.callbacks import ModelCheckpoint, LearningRateScheduler\n",
    "from keras.callbacks import ReduceLROnPlateau, CSVLogger \n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from keras.regularizers import l2\n",
    "from keras import backend as K\n",
    "from keras.models import Model\n",
    "from keras.datasets import cifar10\n",
    "import numpy as np\n",
    "import os\n",
    "\n",
    "\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# !pip install livelossplot \n",
    "from livelossplot import PlotLossesKeras\n",
    "\n",
    "\n",
    "from keras import backend\n",
    "\n",
    "from keras.layers import AveragePooling1D, Input, GlobalMaxPooling1D\n",
    "from keras import layers, models\n",
    "from keras import backend\n",
    "\n",
    "        \n",
    "import tensorflow as tf\n",
    "config = tf.ConfigProto()\n",
    "config.gpu_options.per_process_gpu_memory_fraction = 0.4\n",
    "session = tf.Session(config=config)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "AVpGoW9HF3C5"
   },
   "source": [
    "# Data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "8ByRdciSCZ55"
   },
   "source": [
    "## Load Data set "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 329
    },
    "colab_type": "code",
    "id": "EwSMiMi2QyBr",
    "outputId": "b272f4f7-f070-4727-978a-dc0addd1a768"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(24000, 500) (24000,)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x720 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# data_set_path = './data_set/my_data/class_9'\n",
    "data_set_path = './data_set/svm/'\n",
    "\n",
    "save_path = './save_file/data_6_500_hard_svm'\n",
    "\n",
    "x_data = np.load(data_set_path+\"/x_data_6_500_hard_svm_acc+gyro.npy\")\n",
    "y_data = np.load(data_set_path+\"/y_data_6_500_hard_svm_acc+gyro.npy\")\n",
    "print(x_data.shape, y_data.shape)\n",
    "\n",
    "data_classes = 6\n",
    "\n",
    "# # Normalization\n",
    "# for idx in range(len(x_data)):\n",
    "#     min = np.min(x_data[idx])\n",
    "#     max = np.max(x_data[idx])\n",
    "#     x_data[idx] = (x_data[idx] - min) / (max - min)\n",
    "\n",
    "# fig = plt.figure(figsize=(25,10))\n",
    "\n",
    "# for id in range(4):\n",
    "#     ax = fig.add_subplot(2,4,id+1)\n",
    "#     ax.plot(x_data[id])\n",
    "#     ax.grid(True)\n",
    "#     ax.set_title(f\"x_train{id}\")\n",
    "\n",
    "# # Standardize\n",
    "# for idx in range(len(x_data)):\n",
    "#     standard_deviation = np.std(x_data[idx])\n",
    "#     mean = np.mean(x_data[idx])\n",
    "#     x_data[idx] = (x_data[idx] - mean) / standard_deviation\n",
    "\n",
    "\n",
    "# fig = plt.figure(figsize=(25,10))\n",
    "\n",
    "# for id in range(4):\n",
    "#     ax = fig.add_subplot(2,4,id+1)\n",
    "#     ax.plot(x_data[id])\n",
    "#     ax.grid(True)\n",
    "#     ax.set_title(f\"x_train{id}\")\n",
    "\n",
    "fig = plt.figure(figsize=(25,10))\n",
    "\n",
    "for id in range(4):\n",
    "    ax = fig.add_subplot(2,4,id+1)\n",
    "    ax.plot(x_data[id])\n",
    "    ax.grid(True)\n",
    "    ax.set_title(f\"x_train{id}\")\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "KjCs0sUyFeeV"
   },
   "source": [
    "## Data Augmentation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "M2MvkkwRBMkc"
   },
   "outputs": [],
   "source": [
    "\n",
    "class TimeSeriesDataAugmentation:\n",
    "\n",
    "    # def __init__(self):\n",
    "        # self.sigma = 0.05\n",
    "\n",
    "    def DA_Jitter(self, X, sigma=0.05):\n",
    "        myNoise = np.random.normal(loc=0, scale=sigma, size=X.shape)\n",
    "        return X+myNoise\n",
    "\n",
    "    def DA_Scaling(self, X, sigma=0.1):\n",
    "        scalingFactor = np.random.normal(loc=1.0, scale=sigma, size=(1)) # shape=(1,3)\n",
    "        myNoise = np.matmul(np.ones((X.shape[0],1)), scalingFactor)\n",
    "        return X*myNoise\n",
    "\n",
    "\n",
    "    def GenerateRandomCurves(self, X, sigma=0.2, knot=4):\n",
    "        \n",
    "        xx = np.arange(0, X.shape[0], (X.shape[0]-1) / (knot + 1)) #(3,1) * (1, 6) = (3,6) // (6,3)\n",
    "\n",
    "        yy = np.random.normal(loc = 1.0, scale = sigma, size = (knot + 2)) # (6, 3)\n",
    "        x_range = np.arange(X.shape[0])\n",
    "\n",
    "        cs_x = CubicSpline(xx, yy)\n",
    "\n",
    "        return np.array(cs_x(x_range))\n",
    "\n",
    "\n",
    "    def DA_MagWarp(self, X, sigma=0.2):\n",
    "        return X * self.GenerateRandomCurves(X, sigma)\n",
    "\n",
    "\n",
    "\n",
    "    def DistortTimesteps(self, X, sigma=0.2, knot=4):\n",
    "        tt = self.GenerateRandomCurves(X, sigma, knot) # Regard these samples aroun 1 as time intervals\n",
    "        tt_cum = np.cumsum(tt)        # Add intervals to make a cumulative graph\n",
    "        t_scale = (X.shape[0]-1) / tt_cum[-1] \n",
    "\n",
    "        tt_cum = tt_cum * t_scale\n",
    "\n",
    "        return tt_cum\n",
    "\n",
    "\n",
    "\n",
    "    def DA_TimeWarp(self, X, sigma=0.2):\n",
    "        tt_new = self.DistortTimesteps(X, sigma)\n",
    "\n",
    "        X_new = np.zeros(X.shape) #(3600, 3)      0\n",
    "\n",
    "        x_range = np.arange(X.shape[0]) #(3600,)  0~3599\n",
    "\n",
    "        X_new = np.interp(x_range, tt_new, X)\n",
    "\n",
    "        return X_new\n",
    "\n",
    "\n",
    "\n",
    "    def DA_Permutation(self, X, nPerm=4, minSegLength=10):\n",
    "        data = X\n",
    "        X_new = np.zeros(data.shape)     #(100,)\n",
    "        idx = np.random.permutation(nPerm) \n",
    "\n",
    "        bWhile = True\n",
    "        while bWhile == True:\n",
    "            segs = np.zeros(nPerm+1, dtype=int)\n",
    "            \n",
    "            segs[1:-1] = np.sort(np.random.randint(minSegLength, data.shape[0]-minSegLength, nPerm-1))\n",
    "            segs[-1] = data.shape[0]\n",
    "\n",
    "            if np.min(segs[1:] - segs[0:-1]) > minSegLength:\n",
    "                bWhile = False\n",
    "\n",
    "        pp = 0\n",
    "        for ii in range(nPerm):\n",
    "            x_temp = data[segs[idx[ii]]:segs[idx[ii]+1]]\n",
    "            X_new[pp:pp+len(x_temp)] = x_temp\n",
    "            pp += len(x_temp)\n",
    "\n",
    "        return X_new\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "    def DA_Permutation(self, X, nPerm=4, minSegLength=10):\n",
    "        data = X\n",
    "        X_new = np.zeros(data.shape)     #(100,)\n",
    "        idx = np.random.permutation(nPerm) \n",
    "\n",
    "        bWhile = True\n",
    "        while bWhile == True:\n",
    "            segs = np.zeros(nPerm+1, dtype=int)\n",
    "            \n",
    "            segs[1:-1] = np.sort(np.random.randint(minSegLength, data.shape[0]-minSegLength, nPerm-1))\n",
    "            segs[-1] = data.shape[0]\n",
    "\n",
    "            if np.min(segs[1:] - segs[0:-1]) > minSegLength:\n",
    "                bWhile = False\n",
    "\n",
    "        pp = 0\n",
    "        for ii in range(nPerm):\n",
    "            x_temp = data[segs[idx[ii]]:segs[idx[ii]+1]]\n",
    "            X_new[pp:pp+len(x_temp)] = x_temp\n",
    "            pp += len(x_temp)\n",
    "\n",
    "        return X_new\n",
    "\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "aH3x6zo2FiyT"
   },
   "source": [
    "## Data split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 54
    },
    "colab_type": "code",
    "id": "MNbBVlyURAd8",
    "outputId": "ce63d075-9322-4f3b-8a24-2eeebc8e7be1"
   },
   "outputs": [],
   "source": [
    "\n",
    "#data_set_구분\n",
    "from sklearn.model_selection import train_test_split\n",
    "x_train, x_test, y_train, y_test = train_test_split(x_data, y_data, test_size=0.3, random_state=1, shuffle=False)\n",
    "\n",
    "print(x_train.shape, x_test.shape, y_train.shape, y_test.shape)\n",
    "\n",
    "#y_data onehot encoding\n",
    "y_train = keras.utils.to_categorical(y_train, data_classes)\n",
    "y_test = keras.utils.to_categorical(y_test, data_classes)\n",
    "\n",
    "print(x_train.shape, x_test.shape, y_train.shape, y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "_OAzDrwbFqnn"
   },
   "source": [
    "## Data packaging"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "id": "ZYWOEh9sR1ss",
    "outputId": "12f68d36-1b16-4a83-f70d-fa64f469218b"
   },
   "outputs": [],
   "source": [
    "isPerdata = True\n",
    "# isPerdata = False\n",
    "\n",
    "if not isPerdata:\n",
    "    x_train = x_train.reshape(-1, 100)\n",
    "    x_test  = x_test.reshape(-1, 100)\n",
    "    y_train = y_train.reshape(-1, 9)\n",
    "    y_test = y_test.reshape(-1, 9)\n",
    "else:\n",
    "    # x_train = x_train.reshape(-1, 1, 100)\n",
    "    # x_test  = x_test.reshape(-1, 1, 100)\n",
    "#     y_train = y_train.reshape(-1, 1, 9)\n",
    "#     y_test = y_test.reshape(-1, 1, 9)\n",
    "    x_train = x_train.reshape(-1, 100,1)\n",
    "    x_test  = x_test.reshape(-1, 100,1)\n",
    "    # y_train = y_train.reshape(-1, 9)\n",
    "    # y_test = y_test.reshape(-1, 9)\n",
    "\n",
    "\n",
    "# x_train = np.reshape(x_train, (x_train.shape[0], 1, x_train.shape[1]))\n",
    "# x_test = np.reshape(x_test, (x_test.shape[0], 1, x_test.shape[1]))\n",
    "\n",
    "print(x_train.shape, x_test.shape, y_train.shape, y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# k fold cross validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(24000, 500, 1) (24000, 6)\n"
     ]
    }
   ],
   "source": [
    "#y_data onehot encoding\n",
    "y = keras.utils.to_categorical(y_data, data_classes)\n",
    "x = x_data.reshape(-1, 500,1)\n",
    "\n",
    "print(x.shape, y.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "YilogMXq358K"
   },
   "outputs": [],
   "source": [
    "\n",
    "def recall_m(y_true, y_pred):\n",
    "        true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))\n",
    "        possible_positives = K.sum(K.round(K.clip(y_true, 0, 1)))\n",
    "        recall = true_positives / (possible_positives + K.epsilon())\n",
    "        return recall\n",
    "\n",
    "def precision_m(y_true, y_pred):\n",
    "        true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))\n",
    "        predicted_positives = K.sum(K.round(K.clip(y_pred, 0, 1)))\n",
    "        precision = true_positives / (predicted_positives + K.epsilon())\n",
    "        return precision\n",
    "\n",
    "def f1_m(y_true, y_pred):\n",
    "    precision = precision_m(y_true, y_pred)\n",
    "    recall = recall_m(y_true, y_pred)\n",
    "    return 2*((precision*recall)/(precision+recall+K.epsilon()))\n",
    "\n",
    "\n",
    "from keras import backend as K\n",
    "\n",
    "# 특정 클래스에 대한 정밀도\n",
    "def single_class_precision(interesting_class_id):\n",
    "    def prec(y_true, y_pred):\n",
    "        class_id_true = K.argmax(y_true, axis=-1)\n",
    "        class_id_pred = K.argmax(y_pred, axis=-1)\n",
    "        precision_mask = K.cast(K.equal(class_id_pred, interesting_class_id), 'int32')\n",
    "        class_prec_tensor = K.cast(K.equal(class_id_true, class_id_pred), 'int32') * precision_mask\n",
    "        class_prec = K.cast(K.sum(class_prec_tensor), 'float32') / K.cast(K.maximum(K.sum(precision_mask), 1), 'float32')\n",
    "        return class_prec\n",
    "    return prec\n",
    "\n",
    "\n",
    "# 특정 클래스에 대한 재현율\n",
    "def single_class_recall(interesting_class_id):\n",
    "    def recall(y_true, y_pred):\n",
    "        class_id_true = K.argmax(y_true, axis=-1)\n",
    "        class_id_pred = K.argmax(y_pred, axis=-1)\n",
    "        recall_mask = K.cast(K.equal(class_id_true, interesting_class_id), 'int32')\n",
    "        class_recall_tensor = K.cast(K.equal(class_id_true, class_id_pred), 'int32') * recall_mask\n",
    "        class_recall = K.cast(K.sum(class_recall_tensor), 'float32') / K.cast(K.maximum(K.sum(recall_mask), 1), 'float32')\n",
    "        return class_recall\n",
    "    return recall\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "8iMdPXqMF-xu"
   },
   "source": [
    "# Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def dense_block(x, blocks, growth_rate, name):\n",
    "    for i in range(blocks):\n",
    "        x = conv_block(x, growth_rate, name=name + '_block' + str(i + 1))\n",
    "    return x\n",
    "\n",
    "def conv_block(x, growth_rate, name):\n",
    "    bn_axis = 2\n",
    "    x1 = layers.BatchNormalization(axis=bn_axis, epsilon=1.001e-5, name=name + '_0_bn')(x)\n",
    "\n",
    "    x1 = layers.Activation('relu', name=name + '_0_relu')(x1)\n",
    "\n",
    "    x1 = layers.Conv1D(4 * growth_rate, 1, use_bias=True, kernel_initializer='he_normal', name=name + '_1_conv')(x1)\n",
    "\n",
    "    x1 = layers.BatchNormalization(axis=bn_axis, epsilon=1.001e-5, name=name + '_1_bn')(x1)\n",
    "\n",
    "    x1 = layers.Activation('relu', name=name + '_1_relu')(x1)\n",
    "\n",
    "    x1 = layers.Conv1D(growth_rate, 3, padding='same', use_bias=True, kernel_initializer='he_normal', name=name + '_2_conv')(x1)\n",
    "\n",
    "    x = layers.Concatenate(axis=bn_axis, name=name + '_concat')([x, x1])\n",
    "\n",
    "    # x = layers.Conv1D(growth_rate, 1, padding='same', use_bias=False, name=name + '_3_conv')(x)\n",
    "    # x = layers.Add(name=name + '_concat')([x, x1])\n",
    "\n",
    "    \n",
    "    return x\n",
    "\n",
    "\n",
    "\n",
    "def transition_block(x, reduction, name):\n",
    "    bn_axis = 2\n",
    "    x = layers.BatchNormalization(axis=bn_axis, epsilon=1.001e-5, name=name + '_bn')(x)\n",
    "\n",
    "    x = layers.Activation('relu', name=name + '_relu')(x)\n",
    "\n",
    "    x = layers.Conv1D(int(backend.int_shape(x)[bn_axis] * reduction), 1, use_bias=False, kernel_initializer='he_normal', name=name + '_conv')(x)\n",
    "                      \n",
    "#     x = layers.AveragePooling1D(2, strides=2, name=name + '_pool')(x)\n",
    "    x = layers.MaxPooling1D(2, strides=2, name=name + '_pool')(x)\n",
    "\n",
    "    # x = spatial_pyramid_pool(x, 3, bn_axis, name)\n",
    "    \n",
    "    return x\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.engine.topology import Layer\n",
    "import keras.backend as K\n",
    "\n",
    "\n",
    "class SpatialPyramidPooling(Layer):\n",
    "    \"\"\"Spatial pyramid pooling layer for 2D inputs.\n",
    "    See Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition,\n",
    "    K. He, X. Zhang, S. Ren, J. Sun\n",
    "    # Arguments\n",
    "        pool_list: list of int\n",
    "            List of pooling regions to use. The length of the list is the number of pooling regions,\n",
    "            each int in the list is the number of regions in that pool. For example [1,2,4] would be 3\n",
    "            regions with 1, 2x2 and 4x4 max pools, so 21 outputs per feature map\n",
    "    # Input shape\n",
    "        4D tensor with shape:\n",
    "        `(samples, channels, rows, cols)` if dim_ordering='th'\n",
    "        or 4D tensor with shape:\n",
    "        `(samples, rows, cols, channels)` if dim_ordering='tf'.\n",
    "    # Output shape\n",
    "        2D tensor with shape:\n",
    "        `(samples, channels * sum([i * i for i in pool_list])`\n",
    "    \"\"\"\n",
    "\n",
    "    def __init__(self, pool_list, **kwargs):\n",
    "\n",
    "        self.pool_list = pool_list\n",
    "\n",
    "        self.num_outputs_per_channel = sum([i for i in pool_list])\n",
    "\n",
    "        super(SpatialPyramidPooling, self).__init__(**kwargs)\n",
    "\n",
    "    def build(self, input_shape):\n",
    "        self.nb_channels = input_shape[2]\n",
    "\n",
    "    def compute_output_shape(self, input_shape):\n",
    "        return (input_shape[0], self.nb_channels * self.num_outputs_per_channel)\n",
    "\n",
    "    def get_config(self):\n",
    "        config = {'pool_list': self.pool_list}\n",
    "        base_config = super(SpatialPyramidPooling, self).get_config()\n",
    "        return dict(list(base_config.items()) + list(config.items()))\n",
    "\n",
    "    def call(self, x, mask=None):\n",
    "\n",
    "        input_shape = K.shape(x)\n",
    "\n",
    "        num_rows = input_shape[1]\n",
    "        # num_cols = input_shape[2]\n",
    "\n",
    "        row_length = [K.cast(num_rows, 'float32') / i for i in self.pool_list]\n",
    "        # col_length = [K.cast(num_cols, 'float32') / i for i in self.pool_list]\n",
    "\n",
    "\n",
    "        outputs = []\n",
    "\n",
    "\n",
    "        for pool_num, num_pool_regions in enumerate(self.pool_list):\n",
    "            for jy in range(num_pool_regions):\n",
    "                # for ix in range(num_pool_regions):\n",
    "                #     x1 = ix * col_length[pool_num]\n",
    "                #     x2 = ix * col_length[pool_num] + col_length[pool_num]\n",
    "                    y1 = jy * row_length[pool_num]\n",
    "                    y2 = jy * row_length[pool_num] + row_length[pool_num]\n",
    "\n",
    "                    # x1 = K.cast(K.round(x1), 'int32')\n",
    "                    # x2 = K.cast(K.round(x2), 'int32')\n",
    "                    y1 = K.cast(K.round(y1), 'int32')\n",
    "                    y2 = K.cast(K.round(y2), 'int32')\n",
    "\n",
    "                    # new_shape = [input_shape[0], y2 - y1,\n",
    "                    #                 x2 - x1, input_shape[3]]\n",
    "\n",
    "                    new_shape = [input_shape[0], y2 - y1, input_shape[2]]\n",
    "\n",
    "\n",
    "                    x_crop = x[:, y1:y2, :]\n",
    "                    xm = K.reshape(x_crop, new_shape)\n",
    "#                     pooled_val = K.max(xm, axis=1)\n",
    "                    pooled_val = K.mean(xm, axis=1)\n",
    "                    outputs.append(pooled_val)\n",
    "\n",
    "        #outputs = K.concatenate(outputs,axis = 1)\n",
    "        outputs = K.concatenate(outputs)\n",
    "        #outputs = K.reshape(outputs,(len(self.pool_list),self.num_outputs_per_channel,input_shape[0],input_shape[1]))\n",
    "        #outputs = K.permute_dimensions(outputs,(3,1,0,2))\n",
    "        #outputs = K.reshape(outputs,(input_shape[0], self.num_outputs_per_channel * self.nb_channels))\n",
    "\n",
    "        return outputs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# version 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "def DcnnNetv3(depth=None,\n",
    "            growthRate = None,\n",
    "            include_top = None,\n",
    "            input_tensor = None,\n",
    "            input_shape = None,            \n",
    "            pooling = None,\n",
    "            classes = None,\n",
    "            **kwargs):\n",
    "    \n",
    "\n",
    "\n",
    "    if type(depth) is list:\n",
    "        if len(depth) == 4:\n",
    "            blocks = depth\n",
    "            depth = sum(blocks) * 2 + 5\n",
    "        else:\n",
    "            assert True, \"custom depth length is 4\"\n",
    "    else:\n",
    "        if depth == 121:\n",
    "            blocks = [6, 12, 24, 16]\n",
    "        elif depth == 169:\n",
    "            blocks = [6, 12, 32, 32]\n",
    "        elif depth == 201:\n",
    "            blocks = [6, 12, 48, 32]\n",
    "        elif depth == 161:\n",
    "            blocks = [6, 12, 36, 24]\n",
    "        else:\n",
    "            print(f'{depth} is not allowed depth, you will use custom_blocks parameter')\n",
    "\n",
    "\n",
    "    if input_tensor is None:\n",
    "        img_input = layers.Input(shape=input_shape)\n",
    "    else:\n",
    "        if not backend.is_keras_tensor(input_tensor):\n",
    "            img_input = layers.Input(tensor=input_tensor, shape=input_shape)\n",
    "        else:\n",
    "            img_input = input_tensor\n",
    "\n",
    "\n",
    "    nChannels = 2 * growthRate\n",
    "    bn_axis = 2\n",
    "\n",
    "\n",
    "    x = layers.ZeroPadding1D(padding=3)(img_input)\n",
    "    x = layers.Conv1D(nChannels, 7, strides=1, use_bias=True, name='conv1/conv')(x)\n",
    "\n",
    "#     x = layers.BatchNormalization(axis=bn_axis, epsilon=1.001e-5, name='conv1/bn')(x)\n",
    "#     x = layers.Activation('relu', name='conv1/relu')(x)\n",
    "\n",
    "#     x = layers.ZeroPadding1D(padding=1)(x)\n",
    "#     x = layers.MaxPooling1D(3, strides=2, name='pool1')(x)\n",
    "    \n",
    "    \n",
    "\n",
    "    # [6, 12, 48, 32]\n",
    "    x = dense_block(x, blocks[0], growthRate, name='conv2')\n",
    "    x = transition_block(x, 0.5, name='pool2')\n",
    "\n",
    "    x = dense_block(x, blocks[1], growthRate, name='conv3')\n",
    "    x = transition_block(x, 0.5, name='pool3')\n",
    "\n",
    "    x = dense_block(x, blocks[2], growthRate, name='conv4')\n",
    "    x = transition_block(x, 0.5, name='pool4')\n",
    "\n",
    "    x = dense_block(x, blocks[3], growthRate, name='conv5')\n",
    "\n",
    "    x = layers.BatchNormalization(axis=bn_axis, epsilon=1.001e-5, name='bn')(x)\n",
    "    x = layers.Activation('relu', name='relu')(x)\n",
    "\n",
    "    \n",
    "##################################################\n",
    "\n",
    "\n",
    "# FCN Classification block\n",
    "#     x = layers.MaxPooling1D(2, strides=2, name='max_pool')(x)\n",
    "#     x = layers.Conv1D(4096,6, strides=1, activation='relu',name='fc1',padding='valid')(x)\n",
    "#     x = layers.Conv1D(1440,12, strides=1, activation='relu',name='fc2',padding='valid')(x)\n",
    "#     x = layers.Conv1D(9, 1, strides=1,activation='softmax',name='predictions',padding='valid')(x)\n",
    "\n",
    "################################################\n",
    "    \n",
    "#     x = SpatialPyramidPooling([1,2,4], name='spp')(x)\n",
    "\n",
    "    if include_top:\n",
    "#         # !!!!!\n",
    "        x = layers.GlobalAveragePooling1D(name='avg_pool')(x)\n",
    "        x = layers.Dense(classes, activation='softmax', name='fc')(x)\n",
    "#     else:\n",
    "#         if pooling == 'avg':\n",
    "#             x = layers.GlobalAveragePooling1D(name='avg_pool')(x)\n",
    "#         elif pooling == 'max':\n",
    "#             x = layers.GlobalMaxPooling1D(name='max_pool')(x)\n",
    "\n",
    "\n",
    "    model = models.Model(img_input, x, name=f'densenet{depth}')\n",
    "\n",
    "    return model\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "def createModel():\n",
    "    inputs = layers.Input(shape=(500, 1))\n",
    "\n",
    "    modelv3 = DcnnNetv3(depth=201, growthRate=24, include_top=True, \n",
    "                        input_tensor=inputs, pooling='max', classes=data_classes)\n",
    "    modelv3.summary()\n",
    "    modelv3.name\n",
    "    return modelv3"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# traning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch:  0\n",
      "Learning rate:  0.01\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_3 (InputLayer)            (None, 500, 1)       0                                            \n",
      "__________________________________________________________________________________________________\n",
      "zero_padding1d_3 (ZeroPadding1D (None, 506, 1)       0           input_3[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "conv1/conv (Conv1D)             (None, 500, 48)      384         zero_padding1d_3[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block1_0_bn (BatchNormali (None, 500, 48)      192         conv1/conv[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block1_0_relu (Activation (None, 500, 48)      0           conv2_block1_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block1_1_conv (Conv1D)    (None, 500, 96)      4704        conv2_block1_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block1_1_bn (BatchNormali (None, 500, 96)      384         conv2_block1_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block1_1_relu (Activation (None, 500, 96)      0           conv2_block1_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block1_2_conv (Conv1D)    (None, 500, 24)      6936        conv2_block1_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block1_concat (Concatenat (None, 500, 72)      0           conv1/conv[0][0]                 \n",
      "                                                                 conv2_block1_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block2_0_bn (BatchNormali (None, 500, 72)      288         conv2_block1_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block2_0_relu (Activation (None, 500, 72)      0           conv2_block2_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block2_1_conv (Conv1D)    (None, 500, 96)      7008        conv2_block2_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block2_1_bn (BatchNormali (None, 500, 96)      384         conv2_block2_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block2_1_relu (Activation (None, 500, 96)      0           conv2_block2_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block2_2_conv (Conv1D)    (None, 500, 24)      6936        conv2_block2_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block2_concat (Concatenat (None, 500, 96)      0           conv2_block1_concat[0][0]        \n",
      "                                                                 conv2_block2_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block3_0_bn (BatchNormali (None, 500, 96)      384         conv2_block2_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block3_0_relu (Activation (None, 500, 96)      0           conv2_block3_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block3_1_conv (Conv1D)    (None, 500, 96)      9312        conv2_block3_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block3_1_bn (BatchNormali (None, 500, 96)      384         conv2_block3_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block3_1_relu (Activation (None, 500, 96)      0           conv2_block3_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block3_2_conv (Conv1D)    (None, 500, 24)      6936        conv2_block3_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block3_concat (Concatenat (None, 500, 120)     0           conv2_block2_concat[0][0]        \n",
      "                                                                 conv2_block3_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block4_0_bn (BatchNormali (None, 500, 120)     480         conv2_block3_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block4_0_relu (Activation (None, 500, 120)     0           conv2_block4_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block4_1_conv (Conv1D)    (None, 500, 96)      11616       conv2_block4_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block4_1_bn (BatchNormali (None, 500, 96)      384         conv2_block4_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block4_1_relu (Activation (None, 500, 96)      0           conv2_block4_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block4_2_conv (Conv1D)    (None, 500, 24)      6936        conv2_block4_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block4_concat (Concatenat (None, 500, 144)     0           conv2_block3_concat[0][0]        \n",
      "                                                                 conv2_block4_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block5_0_bn (BatchNormali (None, 500, 144)     576         conv2_block4_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block5_0_relu (Activation (None, 500, 144)     0           conv2_block5_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block5_1_conv (Conv1D)    (None, 500, 96)      13920       conv2_block5_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block5_1_bn (BatchNormali (None, 500, 96)      384         conv2_block5_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block5_1_relu (Activation (None, 500, 96)      0           conv2_block5_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block5_2_conv (Conv1D)    (None, 500, 24)      6936        conv2_block5_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block5_concat (Concatenat (None, 500, 168)     0           conv2_block4_concat[0][0]        \n",
      "                                                                 conv2_block5_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block6_0_bn (BatchNormali (None, 500, 168)     672         conv2_block5_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block6_0_relu (Activation (None, 500, 168)     0           conv2_block6_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block6_1_conv (Conv1D)    (None, 500, 96)      16224       conv2_block6_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block6_1_bn (BatchNormali (None, 500, 96)      384         conv2_block6_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block6_1_relu (Activation (None, 500, 96)      0           conv2_block6_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block6_2_conv (Conv1D)    (None, 500, 24)      6936        conv2_block6_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2_block6_concat (Concatenat (None, 500, 192)     0           conv2_block5_concat[0][0]        \n",
      "                                                                 conv2_block6_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "pool2_bn (BatchNormalization)   (None, 500, 192)     768         conv2_block6_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "pool2_relu (Activation)         (None, 500, 192)     0           pool2_bn[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "pool2_conv (Conv1D)             (None, 500, 96)      18432       pool2_relu[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "pool2_pool (MaxPooling1D)       (None, 250, 96)      0           pool2_conv[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block1_0_bn (BatchNormali (None, 250, 96)      384         pool2_pool[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block1_0_relu (Activation (None, 250, 96)      0           conv3_block1_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block1_1_conv (Conv1D)    (None, 250, 96)      9312        conv3_block1_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block1_1_bn (BatchNormali (None, 250, 96)      384         conv3_block1_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block1_1_relu (Activation (None, 250, 96)      0           conv3_block1_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block1_2_conv (Conv1D)    (None, 250, 24)      6936        conv3_block1_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block1_concat (Concatenat (None, 250, 120)     0           pool2_pool[0][0]                 \n",
      "                                                                 conv3_block1_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block2_0_bn (BatchNormali (None, 250, 120)     480         conv3_block1_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block2_0_relu (Activation (None, 250, 120)     0           conv3_block2_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block2_1_conv (Conv1D)    (None, 250, 96)      11616       conv3_block2_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block2_1_bn (BatchNormali (None, 250, 96)      384         conv3_block2_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block2_1_relu (Activation (None, 250, 96)      0           conv3_block2_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block2_2_conv (Conv1D)    (None, 250, 24)      6936        conv3_block2_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block2_concat (Concatenat (None, 250, 144)     0           conv3_block1_concat[0][0]        \n",
      "                                                                 conv3_block2_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block3_0_bn (BatchNormali (None, 250, 144)     576         conv3_block2_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block3_0_relu (Activation (None, 250, 144)     0           conv3_block3_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block3_1_conv (Conv1D)    (None, 250, 96)      13920       conv3_block3_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block3_1_bn (BatchNormali (None, 250, 96)      384         conv3_block3_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block3_1_relu (Activation (None, 250, 96)      0           conv3_block3_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block3_2_conv (Conv1D)    (None, 250, 24)      6936        conv3_block3_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block3_concat (Concatenat (None, 250, 168)     0           conv3_block2_concat[0][0]        \n",
      "                                                                 conv3_block3_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block4_0_bn (BatchNormali (None, 250, 168)     672         conv3_block3_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block4_0_relu (Activation (None, 250, 168)     0           conv3_block4_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block4_1_conv (Conv1D)    (None, 250, 96)      16224       conv3_block4_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block4_1_bn (BatchNormali (None, 250, 96)      384         conv3_block4_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block4_1_relu (Activation (None, 250, 96)      0           conv3_block4_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block4_2_conv (Conv1D)    (None, 250, 24)      6936        conv3_block4_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block4_concat (Concatenat (None, 250, 192)     0           conv3_block3_concat[0][0]        \n",
      "                                                                 conv3_block4_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block5_0_bn (BatchNormali (None, 250, 192)     768         conv3_block4_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block5_0_relu (Activation (None, 250, 192)     0           conv3_block5_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block5_1_conv (Conv1D)    (None, 250, 96)      18528       conv3_block5_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block5_1_bn (BatchNormali (None, 250, 96)      384         conv3_block5_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block5_1_relu (Activation (None, 250, 96)      0           conv3_block5_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block5_2_conv (Conv1D)    (None, 250, 24)      6936        conv3_block5_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block5_concat (Concatenat (None, 250, 216)     0           conv3_block4_concat[0][0]        \n",
      "                                                                 conv3_block5_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block6_0_bn (BatchNormali (None, 250, 216)     864         conv3_block5_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block6_0_relu (Activation (None, 250, 216)     0           conv3_block6_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block6_1_conv (Conv1D)    (None, 250, 96)      20832       conv3_block6_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block6_1_bn (BatchNormali (None, 250, 96)      384         conv3_block6_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block6_1_relu (Activation (None, 250, 96)      0           conv3_block6_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block6_2_conv (Conv1D)    (None, 250, 24)      6936        conv3_block6_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block6_concat (Concatenat (None, 250, 240)     0           conv3_block5_concat[0][0]        \n",
      "                                                                 conv3_block6_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block7_0_bn (BatchNormali (None, 250, 240)     960         conv3_block6_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block7_0_relu (Activation (None, 250, 240)     0           conv3_block7_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block7_1_conv (Conv1D)    (None, 250, 96)      23136       conv3_block7_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block7_1_bn (BatchNormali (None, 250, 96)      384         conv3_block7_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block7_1_relu (Activation (None, 250, 96)      0           conv3_block7_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block7_2_conv (Conv1D)    (None, 250, 24)      6936        conv3_block7_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block7_concat (Concatenat (None, 250, 264)     0           conv3_block6_concat[0][0]        \n",
      "                                                                 conv3_block7_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block8_0_bn (BatchNormali (None, 250, 264)     1056        conv3_block7_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block8_0_relu (Activation (None, 250, 264)     0           conv3_block8_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block8_1_conv (Conv1D)    (None, 250, 96)      25440       conv3_block8_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block8_1_bn (BatchNormali (None, 250, 96)      384         conv3_block8_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block8_1_relu (Activation (None, 250, 96)      0           conv3_block8_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block8_2_conv (Conv1D)    (None, 250, 24)      6936        conv3_block8_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block8_concat (Concatenat (None, 250, 288)     0           conv3_block7_concat[0][0]        \n",
      "                                                                 conv3_block8_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block9_0_bn (BatchNormali (None, 250, 288)     1152        conv3_block8_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block9_0_relu (Activation (None, 250, 288)     0           conv3_block9_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block9_1_conv (Conv1D)    (None, 250, 96)      27744       conv3_block9_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block9_1_bn (BatchNormali (None, 250, 96)      384         conv3_block9_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block9_1_relu (Activation (None, 250, 96)      0           conv3_block9_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block9_2_conv (Conv1D)    (None, 250, 24)      6936        conv3_block9_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block9_concat (Concatenat (None, 250, 312)     0           conv3_block8_concat[0][0]        \n",
      "                                                                 conv3_block9_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block10_0_bn (BatchNormal (None, 250, 312)     1248        conv3_block9_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block10_0_relu (Activatio (None, 250, 312)     0           conv3_block10_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block10_1_conv (Conv1D)   (None, 250, 96)      30048       conv3_block10_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block10_1_bn (BatchNormal (None, 250, 96)      384         conv3_block10_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block10_1_relu (Activatio (None, 250, 96)      0           conv3_block10_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block10_2_conv (Conv1D)   (None, 250, 24)      6936        conv3_block10_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block10_concat (Concatena (None, 250, 336)     0           conv3_block9_concat[0][0]        \n",
      "                                                                 conv3_block10_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block11_0_bn (BatchNormal (None, 250, 336)     1344        conv3_block10_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block11_0_relu (Activatio (None, 250, 336)     0           conv3_block11_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block11_1_conv (Conv1D)   (None, 250, 96)      32352       conv3_block11_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block11_1_bn (BatchNormal (None, 250, 96)      384         conv3_block11_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block11_1_relu (Activatio (None, 250, 96)      0           conv3_block11_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block11_2_conv (Conv1D)   (None, 250, 24)      6936        conv3_block11_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block11_concat (Concatena (None, 250, 360)     0           conv3_block10_concat[0][0]       \n",
      "                                                                 conv3_block11_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block12_0_bn (BatchNormal (None, 250, 360)     1440        conv3_block11_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block12_0_relu (Activatio (None, 250, 360)     0           conv3_block12_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block12_1_conv (Conv1D)   (None, 250, 96)      34656       conv3_block12_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block12_1_bn (BatchNormal (None, 250, 96)      384         conv3_block12_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block12_1_relu (Activatio (None, 250, 96)      0           conv3_block12_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block12_2_conv (Conv1D)   (None, 250, 24)      6936        conv3_block12_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv3_block12_concat (Concatena (None, 250, 384)     0           conv3_block11_concat[0][0]       \n",
      "                                                                 conv3_block12_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "pool3_bn (BatchNormalization)   (None, 250, 384)     1536        conv3_block12_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "pool3_relu (Activation)         (None, 250, 384)     0           pool3_bn[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "pool3_conv (Conv1D)             (None, 250, 192)     73728       pool3_relu[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "pool3_pool (MaxPooling1D)       (None, 125, 192)     0           pool3_conv[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block1_0_bn (BatchNormali (None, 125, 192)     768         pool3_pool[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block1_0_relu (Activation (None, 125, 192)     0           conv4_block1_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block1_1_conv (Conv1D)    (None, 125, 96)      18528       conv4_block1_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block1_1_bn (BatchNormali (None, 125, 96)      384         conv4_block1_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block1_1_relu (Activation (None, 125, 96)      0           conv4_block1_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block1_2_conv (Conv1D)    (None, 125, 24)      6936        conv4_block1_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block1_concat (Concatenat (None, 125, 216)     0           pool3_pool[0][0]                 \n",
      "                                                                 conv4_block1_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block2_0_bn (BatchNormali (None, 125, 216)     864         conv4_block1_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block2_0_relu (Activation (None, 125, 216)     0           conv4_block2_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block2_1_conv (Conv1D)    (None, 125, 96)      20832       conv4_block2_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block2_1_bn (BatchNormali (None, 125, 96)      384         conv4_block2_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block2_1_relu (Activation (None, 125, 96)      0           conv4_block2_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block2_2_conv (Conv1D)    (None, 125, 24)      6936        conv4_block2_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block2_concat (Concatenat (None, 125, 240)     0           conv4_block1_concat[0][0]        \n",
      "                                                                 conv4_block2_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block3_0_bn (BatchNormali (None, 125, 240)     960         conv4_block2_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block3_0_relu (Activation (None, 125, 240)     0           conv4_block3_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block3_1_conv (Conv1D)    (None, 125, 96)      23136       conv4_block3_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block3_1_bn (BatchNormali (None, 125, 96)      384         conv4_block3_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block3_1_relu (Activation (None, 125, 96)      0           conv4_block3_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block3_2_conv (Conv1D)    (None, 125, 24)      6936        conv4_block3_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block3_concat (Concatenat (None, 125, 264)     0           conv4_block2_concat[0][0]        \n",
      "                                                                 conv4_block3_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block4_0_bn (BatchNormali (None, 125, 264)     1056        conv4_block3_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block4_0_relu (Activation (None, 125, 264)     0           conv4_block4_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block4_1_conv (Conv1D)    (None, 125, 96)      25440       conv4_block4_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block4_1_bn (BatchNormali (None, 125, 96)      384         conv4_block4_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block4_1_relu (Activation (None, 125, 96)      0           conv4_block4_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block4_2_conv (Conv1D)    (None, 125, 24)      6936        conv4_block4_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block4_concat (Concatenat (None, 125, 288)     0           conv4_block3_concat[0][0]        \n",
      "                                                                 conv4_block4_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block5_0_bn (BatchNormali (None, 125, 288)     1152        conv4_block4_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block5_0_relu (Activation (None, 125, 288)     0           conv4_block5_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block5_1_conv (Conv1D)    (None, 125, 96)      27744       conv4_block5_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block5_1_bn (BatchNormali (None, 125, 96)      384         conv4_block5_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block5_1_relu (Activation (None, 125, 96)      0           conv4_block5_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block5_2_conv (Conv1D)    (None, 125, 24)      6936        conv4_block5_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block5_concat (Concatenat (None, 125, 312)     0           conv4_block4_concat[0][0]        \n",
      "                                                                 conv4_block5_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block6_0_bn (BatchNormali (None, 125, 312)     1248        conv4_block5_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block6_0_relu (Activation (None, 125, 312)     0           conv4_block6_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block6_1_conv (Conv1D)    (None, 125, 96)      30048       conv4_block6_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block6_1_bn (BatchNormali (None, 125, 96)      384         conv4_block6_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block6_1_relu (Activation (None, 125, 96)      0           conv4_block6_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block6_2_conv (Conv1D)    (None, 125, 24)      6936        conv4_block6_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block6_concat (Concatenat (None, 125, 336)     0           conv4_block5_concat[0][0]        \n",
      "                                                                 conv4_block6_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block7_0_bn (BatchNormali (None, 125, 336)     1344        conv4_block6_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block7_0_relu (Activation (None, 125, 336)     0           conv4_block7_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block7_1_conv (Conv1D)    (None, 125, 96)      32352       conv4_block7_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block7_1_bn (BatchNormali (None, 125, 96)      384         conv4_block7_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block7_1_relu (Activation (None, 125, 96)      0           conv4_block7_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block7_2_conv (Conv1D)    (None, 125, 24)      6936        conv4_block7_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block7_concat (Concatenat (None, 125, 360)     0           conv4_block6_concat[0][0]        \n",
      "                                                                 conv4_block7_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block8_0_bn (BatchNormali (None, 125, 360)     1440        conv4_block7_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block8_0_relu (Activation (None, 125, 360)     0           conv4_block8_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block8_1_conv (Conv1D)    (None, 125, 96)      34656       conv4_block8_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block8_1_bn (BatchNormali (None, 125, 96)      384         conv4_block8_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block8_1_relu (Activation (None, 125, 96)      0           conv4_block8_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block8_2_conv (Conv1D)    (None, 125, 24)      6936        conv4_block8_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block8_concat (Concatenat (None, 125, 384)     0           conv4_block7_concat[0][0]        \n",
      "                                                                 conv4_block8_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block9_0_bn (BatchNormali (None, 125, 384)     1536        conv4_block8_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block9_0_relu (Activation (None, 125, 384)     0           conv4_block9_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block9_1_conv (Conv1D)    (None, 125, 96)      36960       conv4_block9_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block9_1_bn (BatchNormali (None, 125, 96)      384         conv4_block9_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block9_1_relu (Activation (None, 125, 96)      0           conv4_block9_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block9_2_conv (Conv1D)    (None, 125, 24)      6936        conv4_block9_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block9_concat (Concatenat (None, 125, 408)     0           conv4_block8_concat[0][0]        \n",
      "                                                                 conv4_block9_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block10_0_bn (BatchNormal (None, 125, 408)     1632        conv4_block9_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block10_0_relu (Activatio (None, 125, 408)     0           conv4_block10_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block10_1_conv (Conv1D)   (None, 125, 96)      39264       conv4_block10_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block10_1_bn (BatchNormal (None, 125, 96)      384         conv4_block10_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block10_1_relu (Activatio (None, 125, 96)      0           conv4_block10_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block10_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block10_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block10_concat (Concatena (None, 125, 432)     0           conv4_block9_concat[0][0]        \n",
      "                                                                 conv4_block10_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block11_0_bn (BatchNormal (None, 125, 432)     1728        conv4_block10_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block11_0_relu (Activatio (None, 125, 432)     0           conv4_block11_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block11_1_conv (Conv1D)   (None, 125, 96)      41568       conv4_block11_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block11_1_bn (BatchNormal (None, 125, 96)      384         conv4_block11_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block11_1_relu (Activatio (None, 125, 96)      0           conv4_block11_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block11_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block11_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block11_concat (Concatena (None, 125, 456)     0           conv4_block10_concat[0][0]       \n",
      "                                                                 conv4_block11_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block12_0_bn (BatchNormal (None, 125, 456)     1824        conv4_block11_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block12_0_relu (Activatio (None, 125, 456)     0           conv4_block12_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block12_1_conv (Conv1D)   (None, 125, 96)      43872       conv4_block12_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block12_1_bn (BatchNormal (None, 125, 96)      384         conv4_block12_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block12_1_relu (Activatio (None, 125, 96)      0           conv4_block12_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block12_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block12_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block12_concat (Concatena (None, 125, 480)     0           conv4_block11_concat[0][0]       \n",
      "                                                                 conv4_block12_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block13_0_bn (BatchNormal (None, 125, 480)     1920        conv4_block12_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block13_0_relu (Activatio (None, 125, 480)     0           conv4_block13_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block13_1_conv (Conv1D)   (None, 125, 96)      46176       conv4_block13_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block13_1_bn (BatchNormal (None, 125, 96)      384         conv4_block13_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block13_1_relu (Activatio (None, 125, 96)      0           conv4_block13_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block13_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block13_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block13_concat (Concatena (None, 125, 504)     0           conv4_block12_concat[0][0]       \n",
      "                                                                 conv4_block13_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block14_0_bn (BatchNormal (None, 125, 504)     2016        conv4_block13_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block14_0_relu (Activatio (None, 125, 504)     0           conv4_block14_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block14_1_conv (Conv1D)   (None, 125, 96)      48480       conv4_block14_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block14_1_bn (BatchNormal (None, 125, 96)      384         conv4_block14_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block14_1_relu (Activatio (None, 125, 96)      0           conv4_block14_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block14_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block14_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block14_concat (Concatena (None, 125, 528)     0           conv4_block13_concat[0][0]       \n",
      "                                                                 conv4_block14_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block15_0_bn (BatchNormal (None, 125, 528)     2112        conv4_block14_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block15_0_relu (Activatio (None, 125, 528)     0           conv4_block15_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block15_1_conv (Conv1D)   (None, 125, 96)      50784       conv4_block15_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block15_1_bn (BatchNormal (None, 125, 96)      384         conv4_block15_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block15_1_relu (Activatio (None, 125, 96)      0           conv4_block15_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block15_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block15_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block15_concat (Concatena (None, 125, 552)     0           conv4_block14_concat[0][0]       \n",
      "                                                                 conv4_block15_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block16_0_bn (BatchNormal (None, 125, 552)     2208        conv4_block15_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block16_0_relu (Activatio (None, 125, 552)     0           conv4_block16_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block16_1_conv (Conv1D)   (None, 125, 96)      53088       conv4_block16_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block16_1_bn (BatchNormal (None, 125, 96)      384         conv4_block16_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block16_1_relu (Activatio (None, 125, 96)      0           conv4_block16_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block16_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block16_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block16_concat (Concatena (None, 125, 576)     0           conv4_block15_concat[0][0]       \n",
      "                                                                 conv4_block16_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block17_0_bn (BatchNormal (None, 125, 576)     2304        conv4_block16_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block17_0_relu (Activatio (None, 125, 576)     0           conv4_block17_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block17_1_conv (Conv1D)   (None, 125, 96)      55392       conv4_block17_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block17_1_bn (BatchNormal (None, 125, 96)      384         conv4_block17_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block17_1_relu (Activatio (None, 125, 96)      0           conv4_block17_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block17_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block17_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block17_concat (Concatena (None, 125, 600)     0           conv4_block16_concat[0][0]       \n",
      "                                                                 conv4_block17_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block18_0_bn (BatchNormal (None, 125, 600)     2400        conv4_block17_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block18_0_relu (Activatio (None, 125, 600)     0           conv4_block18_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block18_1_conv (Conv1D)   (None, 125, 96)      57696       conv4_block18_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block18_1_bn (BatchNormal (None, 125, 96)      384         conv4_block18_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block18_1_relu (Activatio (None, 125, 96)      0           conv4_block18_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block18_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block18_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block18_concat (Concatena (None, 125, 624)     0           conv4_block17_concat[0][0]       \n",
      "                                                                 conv4_block18_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block19_0_bn (BatchNormal (None, 125, 624)     2496        conv4_block18_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block19_0_relu (Activatio (None, 125, 624)     0           conv4_block19_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block19_1_conv (Conv1D)   (None, 125, 96)      60000       conv4_block19_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block19_1_bn (BatchNormal (None, 125, 96)      384         conv4_block19_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block19_1_relu (Activatio (None, 125, 96)      0           conv4_block19_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block19_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block19_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block19_concat (Concatena (None, 125, 648)     0           conv4_block18_concat[0][0]       \n",
      "                                                                 conv4_block19_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block20_0_bn (BatchNormal (None, 125, 648)     2592        conv4_block19_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block20_0_relu (Activatio (None, 125, 648)     0           conv4_block20_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block20_1_conv (Conv1D)   (None, 125, 96)      62304       conv4_block20_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block20_1_bn (BatchNormal (None, 125, 96)      384         conv4_block20_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block20_1_relu (Activatio (None, 125, 96)      0           conv4_block20_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block20_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block20_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block20_concat (Concatena (None, 125, 672)     0           conv4_block19_concat[0][0]       \n",
      "                                                                 conv4_block20_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block21_0_bn (BatchNormal (None, 125, 672)     2688        conv4_block20_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block21_0_relu (Activatio (None, 125, 672)     0           conv4_block21_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block21_1_conv (Conv1D)   (None, 125, 96)      64608       conv4_block21_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block21_1_bn (BatchNormal (None, 125, 96)      384         conv4_block21_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block21_1_relu (Activatio (None, 125, 96)      0           conv4_block21_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block21_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block21_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block21_concat (Concatena (None, 125, 696)     0           conv4_block20_concat[0][0]       \n",
      "                                                                 conv4_block21_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block22_0_bn (BatchNormal (None, 125, 696)     2784        conv4_block21_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block22_0_relu (Activatio (None, 125, 696)     0           conv4_block22_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block22_1_conv (Conv1D)   (None, 125, 96)      66912       conv4_block22_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block22_1_bn (BatchNormal (None, 125, 96)      384         conv4_block22_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block22_1_relu (Activatio (None, 125, 96)      0           conv4_block22_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block22_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block22_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block22_concat (Concatena (None, 125, 720)     0           conv4_block21_concat[0][0]       \n",
      "                                                                 conv4_block22_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block23_0_bn (BatchNormal (None, 125, 720)     2880        conv4_block22_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block23_0_relu (Activatio (None, 125, 720)     0           conv4_block23_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block23_1_conv (Conv1D)   (None, 125, 96)      69216       conv4_block23_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block23_1_bn (BatchNormal (None, 125, 96)      384         conv4_block23_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block23_1_relu (Activatio (None, 125, 96)      0           conv4_block23_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block23_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block23_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block23_concat (Concatena (None, 125, 744)     0           conv4_block22_concat[0][0]       \n",
      "                                                                 conv4_block23_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block24_0_bn (BatchNormal (None, 125, 744)     2976        conv4_block23_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block24_0_relu (Activatio (None, 125, 744)     0           conv4_block24_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block24_1_conv (Conv1D)   (None, 125, 96)      71520       conv4_block24_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block24_1_bn (BatchNormal (None, 125, 96)      384         conv4_block24_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block24_1_relu (Activatio (None, 125, 96)      0           conv4_block24_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block24_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block24_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block24_concat (Concatena (None, 125, 768)     0           conv4_block23_concat[0][0]       \n",
      "                                                                 conv4_block24_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block25_0_bn (BatchNormal (None, 125, 768)     3072        conv4_block24_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block25_0_relu (Activatio (None, 125, 768)     0           conv4_block25_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block25_1_conv (Conv1D)   (None, 125, 96)      73824       conv4_block25_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block25_1_bn (BatchNormal (None, 125, 96)      384         conv4_block25_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block25_1_relu (Activatio (None, 125, 96)      0           conv4_block25_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block25_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block25_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block25_concat (Concatena (None, 125, 792)     0           conv4_block24_concat[0][0]       \n",
      "                                                                 conv4_block25_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block26_0_bn (BatchNormal (None, 125, 792)     3168        conv4_block25_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block26_0_relu (Activatio (None, 125, 792)     0           conv4_block26_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block26_1_conv (Conv1D)   (None, 125, 96)      76128       conv4_block26_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block26_1_bn (BatchNormal (None, 125, 96)      384         conv4_block26_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block26_1_relu (Activatio (None, 125, 96)      0           conv4_block26_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block26_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block26_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block26_concat (Concatena (None, 125, 816)     0           conv4_block25_concat[0][0]       \n",
      "                                                                 conv4_block26_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block27_0_bn (BatchNormal (None, 125, 816)     3264        conv4_block26_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block27_0_relu (Activatio (None, 125, 816)     0           conv4_block27_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block27_1_conv (Conv1D)   (None, 125, 96)      78432       conv4_block27_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block27_1_bn (BatchNormal (None, 125, 96)      384         conv4_block27_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block27_1_relu (Activatio (None, 125, 96)      0           conv4_block27_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block27_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block27_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block27_concat (Concatena (None, 125, 840)     0           conv4_block26_concat[0][0]       \n",
      "                                                                 conv4_block27_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block28_0_bn (BatchNormal (None, 125, 840)     3360        conv4_block27_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block28_0_relu (Activatio (None, 125, 840)     0           conv4_block28_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block28_1_conv (Conv1D)   (None, 125, 96)      80736       conv4_block28_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block28_1_bn (BatchNormal (None, 125, 96)      384         conv4_block28_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block28_1_relu (Activatio (None, 125, 96)      0           conv4_block28_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block28_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block28_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block28_concat (Concatena (None, 125, 864)     0           conv4_block27_concat[0][0]       \n",
      "                                                                 conv4_block28_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block29_0_bn (BatchNormal (None, 125, 864)     3456        conv4_block28_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block29_0_relu (Activatio (None, 125, 864)     0           conv4_block29_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block29_1_conv (Conv1D)   (None, 125, 96)      83040       conv4_block29_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block29_1_bn (BatchNormal (None, 125, 96)      384         conv4_block29_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block29_1_relu (Activatio (None, 125, 96)      0           conv4_block29_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block29_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block29_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block29_concat (Concatena (None, 125, 888)     0           conv4_block28_concat[0][0]       \n",
      "                                                                 conv4_block29_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block30_0_bn (BatchNormal (None, 125, 888)     3552        conv4_block29_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block30_0_relu (Activatio (None, 125, 888)     0           conv4_block30_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block30_1_conv (Conv1D)   (None, 125, 96)      85344       conv4_block30_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block30_1_bn (BatchNormal (None, 125, 96)      384         conv4_block30_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block30_1_relu (Activatio (None, 125, 96)      0           conv4_block30_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block30_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block30_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block30_concat (Concatena (None, 125, 912)     0           conv4_block29_concat[0][0]       \n",
      "                                                                 conv4_block30_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block31_0_bn (BatchNormal (None, 125, 912)     3648        conv4_block30_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block31_0_relu (Activatio (None, 125, 912)     0           conv4_block31_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block31_1_conv (Conv1D)   (None, 125, 96)      87648       conv4_block31_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block31_1_bn (BatchNormal (None, 125, 96)      384         conv4_block31_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block31_1_relu (Activatio (None, 125, 96)      0           conv4_block31_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block31_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block31_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block31_concat (Concatena (None, 125, 936)     0           conv4_block30_concat[0][0]       \n",
      "                                                                 conv4_block31_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block32_0_bn (BatchNormal (None, 125, 936)     3744        conv4_block31_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block32_0_relu (Activatio (None, 125, 936)     0           conv4_block32_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block32_1_conv (Conv1D)   (None, 125, 96)      89952       conv4_block32_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block32_1_bn (BatchNormal (None, 125, 96)      384         conv4_block32_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block32_1_relu (Activatio (None, 125, 96)      0           conv4_block32_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block32_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block32_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block32_concat (Concatena (None, 125, 960)     0           conv4_block31_concat[0][0]       \n",
      "                                                                 conv4_block32_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block33_0_bn (BatchNormal (None, 125, 960)     3840        conv4_block32_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block33_0_relu (Activatio (None, 125, 960)     0           conv4_block33_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block33_1_conv (Conv1D)   (None, 125, 96)      92256       conv4_block33_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block33_1_bn (BatchNormal (None, 125, 96)      384         conv4_block33_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block33_1_relu (Activatio (None, 125, 96)      0           conv4_block33_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block33_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block33_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block33_concat (Concatena (None, 125, 984)     0           conv4_block32_concat[0][0]       \n",
      "                                                                 conv4_block33_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block34_0_bn (BatchNormal (None, 125, 984)     3936        conv4_block33_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block34_0_relu (Activatio (None, 125, 984)     0           conv4_block34_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block34_1_conv (Conv1D)   (None, 125, 96)      94560       conv4_block34_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block34_1_bn (BatchNormal (None, 125, 96)      384         conv4_block34_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block34_1_relu (Activatio (None, 125, 96)      0           conv4_block34_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block34_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block34_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block34_concat (Concatena (None, 125, 1008)    0           conv4_block33_concat[0][0]       \n",
      "                                                                 conv4_block34_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block35_0_bn (BatchNormal (None, 125, 1008)    4032        conv4_block34_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block35_0_relu (Activatio (None, 125, 1008)    0           conv4_block35_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block35_1_conv (Conv1D)   (None, 125, 96)      96864       conv4_block35_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block35_1_bn (BatchNormal (None, 125, 96)      384         conv4_block35_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block35_1_relu (Activatio (None, 125, 96)      0           conv4_block35_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block35_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block35_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block35_concat (Concatena (None, 125, 1032)    0           conv4_block34_concat[0][0]       \n",
      "                                                                 conv4_block35_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block36_0_bn (BatchNormal (None, 125, 1032)    4128        conv4_block35_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block36_0_relu (Activatio (None, 125, 1032)    0           conv4_block36_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block36_1_conv (Conv1D)   (None, 125, 96)      99168       conv4_block36_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block36_1_bn (BatchNormal (None, 125, 96)      384         conv4_block36_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block36_1_relu (Activatio (None, 125, 96)      0           conv4_block36_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block36_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block36_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block36_concat (Concatena (None, 125, 1056)    0           conv4_block35_concat[0][0]       \n",
      "                                                                 conv4_block36_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block37_0_bn (BatchNormal (None, 125, 1056)    4224        conv4_block36_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block37_0_relu (Activatio (None, 125, 1056)    0           conv4_block37_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block37_1_conv (Conv1D)   (None, 125, 96)      101472      conv4_block37_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block37_1_bn (BatchNormal (None, 125, 96)      384         conv4_block37_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block37_1_relu (Activatio (None, 125, 96)      0           conv4_block37_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block37_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block37_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block37_concat (Concatena (None, 125, 1080)    0           conv4_block36_concat[0][0]       \n",
      "                                                                 conv4_block37_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block38_0_bn (BatchNormal (None, 125, 1080)    4320        conv4_block37_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block38_0_relu (Activatio (None, 125, 1080)    0           conv4_block38_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block38_1_conv (Conv1D)   (None, 125, 96)      103776      conv4_block38_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block38_1_bn (BatchNormal (None, 125, 96)      384         conv4_block38_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block38_1_relu (Activatio (None, 125, 96)      0           conv4_block38_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block38_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block38_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block38_concat (Concatena (None, 125, 1104)    0           conv4_block37_concat[0][0]       \n",
      "                                                                 conv4_block38_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block39_0_bn (BatchNormal (None, 125, 1104)    4416        conv4_block38_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block39_0_relu (Activatio (None, 125, 1104)    0           conv4_block39_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block39_1_conv (Conv1D)   (None, 125, 96)      106080      conv4_block39_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block39_1_bn (BatchNormal (None, 125, 96)      384         conv4_block39_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block39_1_relu (Activatio (None, 125, 96)      0           conv4_block39_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block39_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block39_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block39_concat (Concatena (None, 125, 1128)    0           conv4_block38_concat[0][0]       \n",
      "                                                                 conv4_block39_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block40_0_bn (BatchNormal (None, 125, 1128)    4512        conv4_block39_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block40_0_relu (Activatio (None, 125, 1128)    0           conv4_block40_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block40_1_conv (Conv1D)   (None, 125, 96)      108384      conv4_block40_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block40_1_bn (BatchNormal (None, 125, 96)      384         conv4_block40_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block40_1_relu (Activatio (None, 125, 96)      0           conv4_block40_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block40_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block40_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block40_concat (Concatena (None, 125, 1152)    0           conv4_block39_concat[0][0]       \n",
      "                                                                 conv4_block40_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block41_0_bn (BatchNormal (None, 125, 1152)    4608        conv4_block40_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block41_0_relu (Activatio (None, 125, 1152)    0           conv4_block41_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block41_1_conv (Conv1D)   (None, 125, 96)      110688      conv4_block41_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block41_1_bn (BatchNormal (None, 125, 96)      384         conv4_block41_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block41_1_relu (Activatio (None, 125, 96)      0           conv4_block41_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block41_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block41_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block41_concat (Concatena (None, 125, 1176)    0           conv4_block40_concat[0][0]       \n",
      "                                                                 conv4_block41_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block42_0_bn (BatchNormal (None, 125, 1176)    4704        conv4_block41_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block42_0_relu (Activatio (None, 125, 1176)    0           conv4_block42_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block42_1_conv (Conv1D)   (None, 125, 96)      112992      conv4_block42_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block42_1_bn (BatchNormal (None, 125, 96)      384         conv4_block42_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block42_1_relu (Activatio (None, 125, 96)      0           conv4_block42_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block42_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block42_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block42_concat (Concatena (None, 125, 1200)    0           conv4_block41_concat[0][0]       \n",
      "                                                                 conv4_block42_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block43_0_bn (BatchNormal (None, 125, 1200)    4800        conv4_block42_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block43_0_relu (Activatio (None, 125, 1200)    0           conv4_block43_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block43_1_conv (Conv1D)   (None, 125, 96)      115296      conv4_block43_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block43_1_bn (BatchNormal (None, 125, 96)      384         conv4_block43_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block43_1_relu (Activatio (None, 125, 96)      0           conv4_block43_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block43_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block43_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block43_concat (Concatena (None, 125, 1224)    0           conv4_block42_concat[0][0]       \n",
      "                                                                 conv4_block43_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block44_0_bn (BatchNormal (None, 125, 1224)    4896        conv4_block43_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block44_0_relu (Activatio (None, 125, 1224)    0           conv4_block44_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block44_1_conv (Conv1D)   (None, 125, 96)      117600      conv4_block44_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block44_1_bn (BatchNormal (None, 125, 96)      384         conv4_block44_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block44_1_relu (Activatio (None, 125, 96)      0           conv4_block44_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block44_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block44_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block44_concat (Concatena (None, 125, 1248)    0           conv4_block43_concat[0][0]       \n",
      "                                                                 conv4_block44_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block45_0_bn (BatchNormal (None, 125, 1248)    4992        conv4_block44_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block45_0_relu (Activatio (None, 125, 1248)    0           conv4_block45_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block45_1_conv (Conv1D)   (None, 125, 96)      119904      conv4_block45_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block45_1_bn (BatchNormal (None, 125, 96)      384         conv4_block45_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block45_1_relu (Activatio (None, 125, 96)      0           conv4_block45_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block45_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block45_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block45_concat (Concatena (None, 125, 1272)    0           conv4_block44_concat[0][0]       \n",
      "                                                                 conv4_block45_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block46_0_bn (BatchNormal (None, 125, 1272)    5088        conv4_block45_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block46_0_relu (Activatio (None, 125, 1272)    0           conv4_block46_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block46_1_conv (Conv1D)   (None, 125, 96)      122208      conv4_block46_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block46_1_bn (BatchNormal (None, 125, 96)      384         conv4_block46_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block46_1_relu (Activatio (None, 125, 96)      0           conv4_block46_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block46_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block46_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block46_concat (Concatena (None, 125, 1296)    0           conv4_block45_concat[0][0]       \n",
      "                                                                 conv4_block46_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block47_0_bn (BatchNormal (None, 125, 1296)    5184        conv4_block46_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block47_0_relu (Activatio (None, 125, 1296)    0           conv4_block47_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block47_1_conv (Conv1D)   (None, 125, 96)      124512      conv4_block47_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block47_1_bn (BatchNormal (None, 125, 96)      384         conv4_block47_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block47_1_relu (Activatio (None, 125, 96)      0           conv4_block47_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block47_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block47_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block47_concat (Concatena (None, 125, 1320)    0           conv4_block46_concat[0][0]       \n",
      "                                                                 conv4_block47_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block48_0_bn (BatchNormal (None, 125, 1320)    5280        conv4_block47_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block48_0_relu (Activatio (None, 125, 1320)    0           conv4_block48_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block48_1_conv (Conv1D)   (None, 125, 96)      126816      conv4_block48_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block48_1_bn (BatchNormal (None, 125, 96)      384         conv4_block48_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block48_1_relu (Activatio (None, 125, 96)      0           conv4_block48_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block48_2_conv (Conv1D)   (None, 125, 24)      6936        conv4_block48_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv4_block48_concat (Concatena (None, 125, 1344)    0           conv4_block47_concat[0][0]       \n",
      "                                                                 conv4_block48_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "pool4_bn (BatchNormalization)   (None, 125, 1344)    5376        conv4_block48_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "pool4_relu (Activation)         (None, 125, 1344)    0           pool4_bn[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "pool4_conv (Conv1D)             (None, 125, 672)     903168      pool4_relu[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "pool4_pool (MaxPooling1D)       (None, 62, 672)      0           pool4_conv[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block1_0_bn (BatchNormali (None, 62, 672)      2688        pool4_pool[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block1_0_relu (Activation (None, 62, 672)      0           conv5_block1_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block1_1_conv (Conv1D)    (None, 62, 96)       64608       conv5_block1_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block1_1_bn (BatchNormali (None, 62, 96)       384         conv5_block1_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block1_1_relu (Activation (None, 62, 96)       0           conv5_block1_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block1_2_conv (Conv1D)    (None, 62, 24)       6936        conv5_block1_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block1_concat (Concatenat (None, 62, 696)      0           pool4_pool[0][0]                 \n",
      "                                                                 conv5_block1_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block2_0_bn (BatchNormali (None, 62, 696)      2784        conv5_block1_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block2_0_relu (Activation (None, 62, 696)      0           conv5_block2_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block2_1_conv (Conv1D)    (None, 62, 96)       66912       conv5_block2_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block2_1_bn (BatchNormali (None, 62, 96)       384         conv5_block2_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block2_1_relu (Activation (None, 62, 96)       0           conv5_block2_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block2_2_conv (Conv1D)    (None, 62, 24)       6936        conv5_block2_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block2_concat (Concatenat (None, 62, 720)      0           conv5_block1_concat[0][0]        \n",
      "                                                                 conv5_block2_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block3_0_bn (BatchNormali (None, 62, 720)      2880        conv5_block2_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block3_0_relu (Activation (None, 62, 720)      0           conv5_block3_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block3_1_conv (Conv1D)    (None, 62, 96)       69216       conv5_block3_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block3_1_bn (BatchNormali (None, 62, 96)       384         conv5_block3_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block3_1_relu (Activation (None, 62, 96)       0           conv5_block3_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block3_2_conv (Conv1D)    (None, 62, 24)       6936        conv5_block3_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block3_concat (Concatenat (None, 62, 744)      0           conv5_block2_concat[0][0]        \n",
      "                                                                 conv5_block3_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block4_0_bn (BatchNormali (None, 62, 744)      2976        conv5_block3_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block4_0_relu (Activation (None, 62, 744)      0           conv5_block4_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block4_1_conv (Conv1D)    (None, 62, 96)       71520       conv5_block4_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block4_1_bn (BatchNormali (None, 62, 96)       384         conv5_block4_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block4_1_relu (Activation (None, 62, 96)       0           conv5_block4_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block4_2_conv (Conv1D)    (None, 62, 24)       6936        conv5_block4_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block4_concat (Concatenat (None, 62, 768)      0           conv5_block3_concat[0][0]        \n",
      "                                                                 conv5_block4_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block5_0_bn (BatchNormali (None, 62, 768)      3072        conv5_block4_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block5_0_relu (Activation (None, 62, 768)      0           conv5_block5_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block5_1_conv (Conv1D)    (None, 62, 96)       73824       conv5_block5_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block5_1_bn (BatchNormali (None, 62, 96)       384         conv5_block5_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block5_1_relu (Activation (None, 62, 96)       0           conv5_block5_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block5_2_conv (Conv1D)    (None, 62, 24)       6936        conv5_block5_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block5_concat (Concatenat (None, 62, 792)      0           conv5_block4_concat[0][0]        \n",
      "                                                                 conv5_block5_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block6_0_bn (BatchNormali (None, 62, 792)      3168        conv5_block5_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block6_0_relu (Activation (None, 62, 792)      0           conv5_block6_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block6_1_conv (Conv1D)    (None, 62, 96)       76128       conv5_block6_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block6_1_bn (BatchNormali (None, 62, 96)       384         conv5_block6_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block6_1_relu (Activation (None, 62, 96)       0           conv5_block6_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block6_2_conv (Conv1D)    (None, 62, 24)       6936        conv5_block6_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block6_concat (Concatenat (None, 62, 816)      0           conv5_block5_concat[0][0]        \n",
      "                                                                 conv5_block6_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block7_0_bn (BatchNormali (None, 62, 816)      3264        conv5_block6_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block7_0_relu (Activation (None, 62, 816)      0           conv5_block7_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block7_1_conv (Conv1D)    (None, 62, 96)       78432       conv5_block7_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block7_1_bn (BatchNormali (None, 62, 96)       384         conv5_block7_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block7_1_relu (Activation (None, 62, 96)       0           conv5_block7_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block7_2_conv (Conv1D)    (None, 62, 24)       6936        conv5_block7_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block7_concat (Concatenat (None, 62, 840)      0           conv5_block6_concat[0][0]        \n",
      "                                                                 conv5_block7_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block8_0_bn (BatchNormali (None, 62, 840)      3360        conv5_block7_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block8_0_relu (Activation (None, 62, 840)      0           conv5_block8_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block8_1_conv (Conv1D)    (None, 62, 96)       80736       conv5_block8_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block8_1_bn (BatchNormali (None, 62, 96)       384         conv5_block8_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block8_1_relu (Activation (None, 62, 96)       0           conv5_block8_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block8_2_conv (Conv1D)    (None, 62, 24)       6936        conv5_block8_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block8_concat (Concatenat (None, 62, 864)      0           conv5_block7_concat[0][0]        \n",
      "                                                                 conv5_block8_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block9_0_bn (BatchNormali (None, 62, 864)      3456        conv5_block8_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block9_0_relu (Activation (None, 62, 864)      0           conv5_block9_0_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block9_1_conv (Conv1D)    (None, 62, 96)       83040       conv5_block9_0_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block9_1_bn (BatchNormali (None, 62, 96)       384         conv5_block9_1_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block9_1_relu (Activation (None, 62, 96)       0           conv5_block9_1_bn[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block9_2_conv (Conv1D)    (None, 62, 24)       6936        conv5_block9_1_relu[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block9_concat (Concatenat (None, 62, 888)      0           conv5_block8_concat[0][0]        \n",
      "                                                                 conv5_block9_2_conv[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block10_0_bn (BatchNormal (None, 62, 888)      3552        conv5_block9_concat[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block10_0_relu (Activatio (None, 62, 888)      0           conv5_block10_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block10_1_conv (Conv1D)   (None, 62, 96)       85344       conv5_block10_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block10_1_bn (BatchNormal (None, 62, 96)       384         conv5_block10_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block10_1_relu (Activatio (None, 62, 96)       0           conv5_block10_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block10_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block10_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block10_concat (Concatena (None, 62, 912)      0           conv5_block9_concat[0][0]        \n",
      "                                                                 conv5_block10_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block11_0_bn (BatchNormal (None, 62, 912)      3648        conv5_block10_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block11_0_relu (Activatio (None, 62, 912)      0           conv5_block11_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block11_1_conv (Conv1D)   (None, 62, 96)       87648       conv5_block11_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block11_1_bn (BatchNormal (None, 62, 96)       384         conv5_block11_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block11_1_relu (Activatio (None, 62, 96)       0           conv5_block11_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block11_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block11_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block11_concat (Concatena (None, 62, 936)      0           conv5_block10_concat[0][0]       \n",
      "                                                                 conv5_block11_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block12_0_bn (BatchNormal (None, 62, 936)      3744        conv5_block11_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block12_0_relu (Activatio (None, 62, 936)      0           conv5_block12_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block12_1_conv (Conv1D)   (None, 62, 96)       89952       conv5_block12_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block12_1_bn (BatchNormal (None, 62, 96)       384         conv5_block12_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block12_1_relu (Activatio (None, 62, 96)       0           conv5_block12_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block12_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block12_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block12_concat (Concatena (None, 62, 960)      0           conv5_block11_concat[0][0]       \n",
      "                                                                 conv5_block12_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block13_0_bn (BatchNormal (None, 62, 960)      3840        conv5_block12_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block13_0_relu (Activatio (None, 62, 960)      0           conv5_block13_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block13_1_conv (Conv1D)   (None, 62, 96)       92256       conv5_block13_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block13_1_bn (BatchNormal (None, 62, 96)       384         conv5_block13_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block13_1_relu (Activatio (None, 62, 96)       0           conv5_block13_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block13_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block13_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block13_concat (Concatena (None, 62, 984)      0           conv5_block12_concat[0][0]       \n",
      "                                                                 conv5_block13_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block14_0_bn (BatchNormal (None, 62, 984)      3936        conv5_block13_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block14_0_relu (Activatio (None, 62, 984)      0           conv5_block14_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block14_1_conv (Conv1D)   (None, 62, 96)       94560       conv5_block14_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block14_1_bn (BatchNormal (None, 62, 96)       384         conv5_block14_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block14_1_relu (Activatio (None, 62, 96)       0           conv5_block14_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block14_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block14_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block14_concat (Concatena (None, 62, 1008)     0           conv5_block13_concat[0][0]       \n",
      "                                                                 conv5_block14_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block15_0_bn (BatchNormal (None, 62, 1008)     4032        conv5_block14_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block15_0_relu (Activatio (None, 62, 1008)     0           conv5_block15_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block15_1_conv (Conv1D)   (None, 62, 96)       96864       conv5_block15_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block15_1_bn (BatchNormal (None, 62, 96)       384         conv5_block15_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block15_1_relu (Activatio (None, 62, 96)       0           conv5_block15_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block15_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block15_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block15_concat (Concatena (None, 62, 1032)     0           conv5_block14_concat[0][0]       \n",
      "                                                                 conv5_block15_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block16_0_bn (BatchNormal (None, 62, 1032)     4128        conv5_block15_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block16_0_relu (Activatio (None, 62, 1032)     0           conv5_block16_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block16_1_conv (Conv1D)   (None, 62, 96)       99168       conv5_block16_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block16_1_bn (BatchNormal (None, 62, 96)       384         conv5_block16_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block16_1_relu (Activatio (None, 62, 96)       0           conv5_block16_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block16_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block16_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block16_concat (Concatena (None, 62, 1056)     0           conv5_block15_concat[0][0]       \n",
      "                                                                 conv5_block16_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block17_0_bn (BatchNormal (None, 62, 1056)     4224        conv5_block16_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block17_0_relu (Activatio (None, 62, 1056)     0           conv5_block17_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block17_1_conv (Conv1D)   (None, 62, 96)       101472      conv5_block17_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block17_1_bn (BatchNormal (None, 62, 96)       384         conv5_block17_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block17_1_relu (Activatio (None, 62, 96)       0           conv5_block17_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block17_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block17_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block17_concat (Concatena (None, 62, 1080)     0           conv5_block16_concat[0][0]       \n",
      "                                                                 conv5_block17_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block18_0_bn (BatchNormal (None, 62, 1080)     4320        conv5_block17_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block18_0_relu (Activatio (None, 62, 1080)     0           conv5_block18_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block18_1_conv (Conv1D)   (None, 62, 96)       103776      conv5_block18_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block18_1_bn (BatchNormal (None, 62, 96)       384         conv5_block18_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block18_1_relu (Activatio (None, 62, 96)       0           conv5_block18_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block18_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block18_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block18_concat (Concatena (None, 62, 1104)     0           conv5_block17_concat[0][0]       \n",
      "                                                                 conv5_block18_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block19_0_bn (BatchNormal (None, 62, 1104)     4416        conv5_block18_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block19_0_relu (Activatio (None, 62, 1104)     0           conv5_block19_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block19_1_conv (Conv1D)   (None, 62, 96)       106080      conv5_block19_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block19_1_bn (BatchNormal (None, 62, 96)       384         conv5_block19_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block19_1_relu (Activatio (None, 62, 96)       0           conv5_block19_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block19_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block19_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block19_concat (Concatena (None, 62, 1128)     0           conv5_block18_concat[0][0]       \n",
      "                                                                 conv5_block19_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block20_0_bn (BatchNormal (None, 62, 1128)     4512        conv5_block19_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block20_0_relu (Activatio (None, 62, 1128)     0           conv5_block20_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block20_1_conv (Conv1D)   (None, 62, 96)       108384      conv5_block20_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block20_1_bn (BatchNormal (None, 62, 96)       384         conv5_block20_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block20_1_relu (Activatio (None, 62, 96)       0           conv5_block20_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block20_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block20_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block20_concat (Concatena (None, 62, 1152)     0           conv5_block19_concat[0][0]       \n",
      "                                                                 conv5_block20_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block21_0_bn (BatchNormal (None, 62, 1152)     4608        conv5_block20_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block21_0_relu (Activatio (None, 62, 1152)     0           conv5_block21_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block21_1_conv (Conv1D)   (None, 62, 96)       110688      conv5_block21_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block21_1_bn (BatchNormal (None, 62, 96)       384         conv5_block21_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block21_1_relu (Activatio (None, 62, 96)       0           conv5_block21_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block21_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block21_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block21_concat (Concatena (None, 62, 1176)     0           conv5_block20_concat[0][0]       \n",
      "                                                                 conv5_block21_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block22_0_bn (BatchNormal (None, 62, 1176)     4704        conv5_block21_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block22_0_relu (Activatio (None, 62, 1176)     0           conv5_block22_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block22_1_conv (Conv1D)   (None, 62, 96)       112992      conv5_block22_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block22_1_bn (BatchNormal (None, 62, 96)       384         conv5_block22_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block22_1_relu (Activatio (None, 62, 96)       0           conv5_block22_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block22_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block22_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block22_concat (Concatena (None, 62, 1200)     0           conv5_block21_concat[0][0]       \n",
      "                                                                 conv5_block22_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block23_0_bn (BatchNormal (None, 62, 1200)     4800        conv5_block22_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block23_0_relu (Activatio (None, 62, 1200)     0           conv5_block23_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block23_1_conv (Conv1D)   (None, 62, 96)       115296      conv5_block23_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block23_1_bn (BatchNormal (None, 62, 96)       384         conv5_block23_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block23_1_relu (Activatio (None, 62, 96)       0           conv5_block23_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block23_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block23_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block23_concat (Concatena (None, 62, 1224)     0           conv5_block22_concat[0][0]       \n",
      "                                                                 conv5_block23_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block24_0_bn (BatchNormal (None, 62, 1224)     4896        conv5_block23_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block24_0_relu (Activatio (None, 62, 1224)     0           conv5_block24_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block24_1_conv (Conv1D)   (None, 62, 96)       117600      conv5_block24_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block24_1_bn (BatchNormal (None, 62, 96)       384         conv5_block24_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block24_1_relu (Activatio (None, 62, 96)       0           conv5_block24_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block24_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block24_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block24_concat (Concatena (None, 62, 1248)     0           conv5_block23_concat[0][0]       \n",
      "                                                                 conv5_block24_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block25_0_bn (BatchNormal (None, 62, 1248)     4992        conv5_block24_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block25_0_relu (Activatio (None, 62, 1248)     0           conv5_block25_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block25_1_conv (Conv1D)   (None, 62, 96)       119904      conv5_block25_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block25_1_bn (BatchNormal (None, 62, 96)       384         conv5_block25_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block25_1_relu (Activatio (None, 62, 96)       0           conv5_block25_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block25_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block25_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block25_concat (Concatena (None, 62, 1272)     0           conv5_block24_concat[0][0]       \n",
      "                                                                 conv5_block25_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block26_0_bn (BatchNormal (None, 62, 1272)     5088        conv5_block25_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block26_0_relu (Activatio (None, 62, 1272)     0           conv5_block26_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block26_1_conv (Conv1D)   (None, 62, 96)       122208      conv5_block26_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block26_1_bn (BatchNormal (None, 62, 96)       384         conv5_block26_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block26_1_relu (Activatio (None, 62, 96)       0           conv5_block26_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block26_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block26_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block26_concat (Concatena (None, 62, 1296)     0           conv5_block25_concat[0][0]       \n",
      "                                                                 conv5_block26_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block27_0_bn (BatchNormal (None, 62, 1296)     5184        conv5_block26_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block27_0_relu (Activatio (None, 62, 1296)     0           conv5_block27_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block27_1_conv (Conv1D)   (None, 62, 96)       124512      conv5_block27_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block27_1_bn (BatchNormal (None, 62, 96)       384         conv5_block27_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block27_1_relu (Activatio (None, 62, 96)       0           conv5_block27_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block27_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block27_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block27_concat (Concatena (None, 62, 1320)     0           conv5_block26_concat[0][0]       \n",
      "                                                                 conv5_block27_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block28_0_bn (BatchNormal (None, 62, 1320)     5280        conv5_block27_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block28_0_relu (Activatio (None, 62, 1320)     0           conv5_block28_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block28_1_conv (Conv1D)   (None, 62, 96)       126816      conv5_block28_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block28_1_bn (BatchNormal (None, 62, 96)       384         conv5_block28_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block28_1_relu (Activatio (None, 62, 96)       0           conv5_block28_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block28_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block28_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block28_concat (Concatena (None, 62, 1344)     0           conv5_block27_concat[0][0]       \n",
      "                                                                 conv5_block28_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block29_0_bn (BatchNormal (None, 62, 1344)     5376        conv5_block28_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block29_0_relu (Activatio (None, 62, 1344)     0           conv5_block29_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block29_1_conv (Conv1D)   (None, 62, 96)       129120      conv5_block29_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block29_1_bn (BatchNormal (None, 62, 96)       384         conv5_block29_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block29_1_relu (Activatio (None, 62, 96)       0           conv5_block29_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block29_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block29_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block29_concat (Concatena (None, 62, 1368)     0           conv5_block28_concat[0][0]       \n",
      "                                                                 conv5_block29_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block30_0_bn (BatchNormal (None, 62, 1368)     5472        conv5_block29_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block30_0_relu (Activatio (None, 62, 1368)     0           conv5_block30_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block30_1_conv (Conv1D)   (None, 62, 96)       131424      conv5_block30_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block30_1_bn (BatchNormal (None, 62, 96)       384         conv5_block30_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block30_1_relu (Activatio (None, 62, 96)       0           conv5_block30_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block30_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block30_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block30_concat (Concatena (None, 62, 1392)     0           conv5_block29_concat[0][0]       \n",
      "                                                                 conv5_block30_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block31_0_bn (BatchNormal (None, 62, 1392)     5568        conv5_block30_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block31_0_relu (Activatio (None, 62, 1392)     0           conv5_block31_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block31_1_conv (Conv1D)   (None, 62, 96)       133728      conv5_block31_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block31_1_bn (BatchNormal (None, 62, 96)       384         conv5_block31_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block31_1_relu (Activatio (None, 62, 96)       0           conv5_block31_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block31_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block31_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block31_concat (Concatena (None, 62, 1416)     0           conv5_block30_concat[0][0]       \n",
      "                                                                 conv5_block31_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block32_0_bn (BatchNormal (None, 62, 1416)     5664        conv5_block31_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block32_0_relu (Activatio (None, 62, 1416)     0           conv5_block32_0_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block32_1_conv (Conv1D)   (None, 62, 96)       136032      conv5_block32_0_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block32_1_bn (BatchNormal (None, 62, 96)       384         conv5_block32_1_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block32_1_relu (Activatio (None, 62, 96)       0           conv5_block32_1_bn[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block32_2_conv (Conv1D)   (None, 62, 24)       6936        conv5_block32_1_relu[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "conv5_block32_concat (Concatena (None, 62, 1440)     0           conv5_block31_concat[0][0]       \n",
      "                                                                 conv5_block32_2_conv[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "bn (BatchNormalization)         (None, 62, 1440)     5760        conv5_block32_concat[0][0]       \n",
      "__________________________________________________________________________________________________\n",
      "relu (Activation)               (None, 62, 1440)     0           bn[0][0]                         \n",
      "__________________________________________________________________________________________________\n",
      "avg_pool (GlobalAveragePooling1 (None, 1440)         0           relu[0][0]                       \n",
      "__________________________________________________________________________________________________\n",
      "fc (Dense)                      (None, 6)            8646        avg_pool[0][0]                   \n",
      "==================================================================================================\n",
      "Total params: 9,052,566\n",
      "Trainable params: 8,880,870\n",
      "Non-trainable params: 171,696\n",
      "__________________________________________________________________________________________________\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 19200 samples, validate on 4800 samples\n",
      "Epoch 1/300\n",
      "epoch:  0\n",
      "Learning rate:  0.01\n"
     ]
    },
    {
     "ename": "ResourceExhaustedError",
     "evalue": "OOM when allocating tensor with shape[64,984,1,125] and type float on /job:localhost/replica:0/task:0/device:GPU:0 by allocator GPU_0_bfc\n\t [[Node: training_1/SGD/gradients/conv4_block34_1_conv_2/convolution/Conv2D_grad/Conv2DBackpropFilter-0-TransposeNHWCToNCHW-LayoutOptimizer = Transpose[T=DT_FLOAT, Tperm=DT_INT32, _device=\"/job:localhost/replica:0/task:0/device:GPU:0\"](conv4_block34_1_conv_2/convolution/ExpandDims, PermConstNHWCToNCHW-LayoutOptimizer)]]\nHint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.\n\n\t [[Node: metrics_2/acc/Mean/_21337 = _Recv[client_terminated=false, recv_device=\"/job:localhost/replica:0/task:0/device:CPU:0\", send_device=\"/job:localhost/replica:0/task:0/device:GPU:0\", send_device_incarnation=1, tensor_name=\"edge_91159_metrics_2/acc/Mean\", tensor_type=DT_FLOAT, _device=\"/job:localhost/replica:0/task:0/device:CPU:0\"]()]]\nHint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.\n",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mResourceExhaustedError\u001b[0m                    Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-15-7b3206345ce1>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     80\u001b[0m \u001b[0;31m#                         workers=4,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     81\u001b[0m                         \u001b[0mvalidation_data\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mxval\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0myval\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 82\u001b[0;31m                         callbacks=callbacks))\n\u001b[0m\u001b[1;32m     83\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     84\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda2/envs/pyqt/lib/python3.6/site-packages/keras/engine/training.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, **kwargs)\u001b[0m\n\u001b[1;32m   1037\u001b[0m                                         \u001b[0minitial_epoch\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0minitial_epoch\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1038\u001b[0m                                         \u001b[0msteps_per_epoch\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0msteps_per_epoch\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1039\u001b[0;31m                                         validation_steps=validation_steps)\n\u001b[0m\u001b[1;32m   1040\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1041\u001b[0m     def evaluate(self, x=None, y=None,\n",
      "\u001b[0;32m~/anaconda2/envs/pyqt/lib/python3.6/site-packages/keras/engine/training_arrays.py\u001b[0m in \u001b[0;36mfit_loop\u001b[0;34m(model, f, ins, out_labels, batch_size, epochs, verbose, callbacks, val_f, val_ins, shuffle, callback_metrics, initial_epoch, steps_per_epoch, validation_steps)\u001b[0m\n\u001b[1;32m    197\u001b[0m                     \u001b[0mins_batch\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mins_batch\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtoarray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    198\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 199\u001b[0;31m                 \u001b[0mouts\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mins_batch\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    200\u001b[0m                 \u001b[0mouts\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mto_list\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mouts\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    201\u001b[0m                 \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mo\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mout_labels\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mouts\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda2/envs/pyqt/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, inputs)\u001b[0m\n\u001b[1;32m   2713\u001b[0m                 \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_legacy_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2714\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2715\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2716\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2717\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mpy_any\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mis_tensor\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda2/envs/pyqt/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py\u001b[0m in \u001b[0;36m_call\u001b[0;34m(self, inputs)\u001b[0m\n\u001b[1;32m   2673\u001b[0m             \u001b[0mfetched\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_callable_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0marray_vals\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrun_metadata\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrun_metadata\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2674\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2675\u001b[0;31m             \u001b[0mfetched\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_callable_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0marray_vals\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2676\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mfetched\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0moutputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2677\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda2/envs/pyqt/lib/python3.6/site-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1380\u001b[0m           ret = tf_session.TF_SessionRunCallable(\n\u001b[1;32m   1381\u001b[0m               \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_session\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_handle\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstatus\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1382\u001b[0;31m               run_metadata_ptr)\n\u001b[0m\u001b[1;32m   1383\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mrun_metadata\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1384\u001b[0m           \u001b[0mproto_data\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtf_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mTF_GetBuffer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrun_metadata_ptr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda2/envs/pyqt/lib/python3.6/site-packages/tensorflow/python/framework/errors_impl.py\u001b[0m in \u001b[0;36m__exit__\u001b[0;34m(self, type_arg, value_arg, traceback_arg)\u001b[0m\n\u001b[1;32m    517\u001b[0m             \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    518\u001b[0m             \u001b[0mcompat\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mas_text\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc_api\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mTF_Message\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstatus\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstatus\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 519\u001b[0;31m             c_api.TF_GetCode(self.status.status))\n\u001b[0m\u001b[1;32m    520\u001b[0m     \u001b[0;31m# Delete the underlying status object from memory otherwise it stays alive\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    521\u001b[0m     \u001b[0;31m# as there is a reference to status from this from the traceback due to\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mResourceExhaustedError\u001b[0m: OOM when allocating tensor with shape[64,984,1,125] and type float on /job:localhost/replica:0/task:0/device:GPU:0 by allocator GPU_0_bfc\n\t [[Node: training_1/SGD/gradients/conv4_block34_1_conv_2/convolution/Conv2D_grad/Conv2DBackpropFilter-0-TransposeNHWCToNCHW-LayoutOptimizer = Transpose[T=DT_FLOAT, Tperm=DT_INT32, _device=\"/job:localhost/replica:0/task:0/device:GPU:0\"](conv4_block34_1_conv_2/convolution/ExpandDims, PermConstNHWCToNCHW-LayoutOptimizer)]]\nHint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.\n\n\t [[Node: metrics_2/acc/Mean/_21337 = _Recv[client_terminated=false, recv_device=\"/job:localhost/replica:0/task:0/device:CPU:0\", send_device=\"/job:localhost/replica:0/task:0/device:GPU:0\", send_device_incarnation=1, tensor_name=\"edge_91159_metrics_2/acc/Mean\", tensor_type=DT_FLOAT, _device=\"/job:localhost/replica:0/task:0/device:CPU:0\"]()]]\nHint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.\n"
     ]
    }
   ],
   "source": [
    "if not os.path.isdir(save_path):\n",
    "    os.makedirs(save_path)\n",
    "\n",
    "batch_size = 64\n",
    "split_val = 0.3\n",
    "fit_shuffle = True\n",
    "optimizer = 'sgd'\n",
    "start_lr = 1e-2\n",
    "\n",
    "split_shuffle = False\n",
    "history = []\n",
    "\n",
    "\n",
    "def lr_schedule(epoch):\n",
    "        lr = start_lr\n",
    "        if epoch > 225:\n",
    "            lr *= 1e-2\n",
    "        elif epoch > 150:\n",
    "            lr *= 1e-1\n",
    "        print('epoch: ', epoch)\n",
    "        print('Learning rate: ', lr)\n",
    "        return lr\n",
    "\n",
    "# 3. 모델 학습과정 설정하기\n",
    "sgd = SGD(lr=lr_schedule(0), decay=1e-4, momentum=0.9, nesterov=True)\n",
    "# adam = Adam(lr=lr_schedule(0))\n",
    "# nAdam = Nadam(lr=lr_schedule(0))\n",
    "\n",
    "\n",
    "# model.compile(optimizer=sgd, loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "\n",
    "\n",
    "\n",
    "# modelv3.compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['accuracy',\n",
    "#                                                                          f1_m, recall_m, precision_m])\n",
    "\n",
    "\n",
    "\n",
    "from sklearn.model_selection import StratifiedKFold\n",
    "# Instantiate the cross validator\n",
    "\n",
    "skf = StratifiedKFold(n_splits=5, shuffle=True)\n",
    "# Loop through the indices the split() method returns\n",
    "\n",
    "for index, (train_indices, val_indices) in enumerate(skf.split(x_data, y_data)):\n",
    "\n",
    "#     print(x[train_indices], x[val_indices])\n",
    "#     print(y[train_indices], y[val_indices])\n",
    "    xtrain, xval = x[train_indices], x[val_indices]\n",
    "    ytrain, yval = y[train_indices], y[val_indices]\n",
    "    \n",
    "    model = None\n",
    "    model = createModel()\n",
    "    \n",
    "    csv_logger_path = save_path+f'/{model.name}(batch_size={batch_size}, split_val={split_val}, fit_shuffle={fit_shuffle}, split_shuffle={split_shuffle}, optimizer={optimizer}, lr=s:{start_lr}, per:1e-1, e:(150, 225))_log.csv'\n",
    "    if not os.path.isdir(os.path.dirname(csv_logger_path)):\n",
    "        os.makedirs(os.path.dirname(csv_logger_path))\n",
    "    csv_logger = CSVLogger(csv_logger_path, append=True)\n",
    "\n",
    "    lr_scheduler = LearningRateScheduler(schedule=lr_schedule)\n",
    "\n",
    "    callbacks = [lr_scheduler, csv_logger]\n",
    "\n",
    "    model.compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['accuracy',\n",
    "                                                               f1_m, recall_m, precision_m,\n",
    "                                                                single_class_precision(0), single_class_recall(0),\n",
    "                                                                single_class_precision(1), single_class_recall(1),\n",
    "                                                                single_class_precision(2), single_class_recall(2),\n",
    "                                                                single_class_precision(3), single_class_recall(3),\n",
    "                                                                single_class_precision(4), single_class_recall(4),\n",
    "                                                                single_class_precision(5), single_class_recall(5),\n",
    "                                                                single_class_precision(6), single_class_recall(6),\n",
    "                                                                single_class_precision(7), single_class_recall(7),\n",
    "                                                                single_class_precision(8), single_class_recall(8)]) \n",
    "    \n",
    "    \n",
    "    history.append(model.fit(xtrain, ytrain, \n",
    "                        epochs=300, batch_size=batch_size, \n",
    "                        shuffle=fit_shuffle,\n",
    "#                         workers=4, \n",
    "                        validation_data=(xval, yval),\n",
    "                        callbacks=callbacks))\n",
    "\n",
    "\n",
    "\n",
    "# # 4. 모델 학습시키기\n",
    "# history = modelv3.fit(x_train, y_train, \n",
    "#                     epochs=300, batch_size=batch_size, \n",
    "#                     shuffle=fit_shuffle,\n",
    "# #                     workers=1, \n",
    "#                       validation_data=(x_test, y_test),\n",
    "#                     callbacks=callbacks)\n",
    "\n",
    "# histories[model] = models[model].fit(x_train, y_train, epochs=300, batch_size=32, workers=4, validation_data=(x_train, y_train))\n",
    "\n",
    "# 6. 모델 평가하기\n",
    "# evaluate = modelv3.evaluate(x_test, y_test, batch_size=batch_size, verbose=1)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "evaluate = modelv3.evaluate(x_test, y_test, batch_size=batch_size, verbose=1)\n",
    "print(\"loss     acc      f1      recall      precision\")\n",
    "print(evaluate)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ps8QIuN_mWhs"
   },
   "source": [
    "# version 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "fiBxlLW7H-WZ"
   },
   "outputs": [],
   "source": [
    "\n",
    "# inputs = Input(shape=(100, 1))\n",
    "\n",
    "# modelv1 = DcnnNetv1(depth=100, growthRate=32, include_top=True, input_tensor=inputs, pooling='avg', classes=9)\n",
    "# modelv1.summary()\n",
    "# modelv1.name\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "RfsjvrWu9jjn"
   },
   "source": [
    "# v1 training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "colab_type": "code",
    "id": "9AK0uGhL9nwy",
    "outputId": "1914d5b7-aae1-499f-e744-f790287e3573"
   },
   "outputs": [],
   "source": [
    "if not os.path.isdir(save_path):\n",
    "    os.mkdir(save_path)\n",
    "\n",
    "batch_size = 64\n",
    "split_val = 0.3\n",
    "fit_shuffle = True\n",
    "optimizer = 'sgd'\n",
    "start_lr = 1e-2\n",
    "split_shuffle = False\n",
    "\n",
    "def lr_schedule(epoch):\n",
    "        lr = start_lr\n",
    "        if epoch > 225:\n",
    "            lr *= 1e-2\n",
    "        elif epoch > 150:\n",
    "            lr *= 1e-1\n",
    "        print('epoch: ', epoch)\n",
    "        print('Learning rate: ', lr)\n",
    "        return lr\n",
    "\n",
    "# 3. 모델 학습과정 설정하기\n",
    "sgd = SGD(lr=lr_schedule(0), decay=1e-4, momentum=0.9, nesterov=True)\n",
    "# adam = Adam(lr=lr_schedule(0))\n",
    "# nAdam = Nadam(lr=lr_schedule(0))\n",
    "\n",
    "\n",
    "# model.compile(optimizer=sgd, loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "\n",
    "\n",
    "# models[model].compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['accuracy',\n",
    "#                                                                                     single_class_precision(0), single_class_recall(0),\n",
    "#                                                                                     single_class_precision(1), single_class_recall(1),\n",
    "#                                                                                     single_class_precision(2), single_class_recall(2),\n",
    "#                                                                                     single_class_precision(3), single_class_recall(3),\n",
    "#                                                                                     single_class_precision(4), single_class_recall(4),\n",
    "#                                                                                     single_class_precision(5), single_class_recall(5),\n",
    "#                                                                                     single_class_precision(3), single_class_recall(6),\n",
    "#                                                                                     single_class_precision(4), single_class_recall(7),\n",
    "#                                                                                     single_class_precision(5), single_class_recall(8)])\n",
    "\n",
    "modelv1.compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['accuracy'])\n",
    "\n",
    "csv_logger_path = save_path+f'/{modelv1.name}(batch_size={batch_size}, split_val={split_val}, fit_shuffle={fit_shuffle}, split_shuffle={split_shuffle}, optimizer={optimizer}, lr=s:{start_lr}, per:1e-1, e:(150, 225))_log.csv'\n",
    "if not os.path.isdir(os.path.dirname(csv_logger_path)):\n",
    "    os.makedirs(os.path.dirname(csv_logger_path))\n",
    "csv_logger = CSVLogger(csv_logger_path, append=True)\n",
    "\n",
    "lr_scheduler = LearningRateScheduler(schedule=lr_schedule)\n",
    "\n",
    "callbacks = [lr_scheduler, csv_logger]\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "# 4. 모델 학습시키기\n",
    "history = modelv1.fit(x_train, y_train, \n",
    "                                        epochs=300, batch_size=batch_size, \n",
    "                                        shuffle=fit_shuffle,\n",
    "                                        workers=4, validation_data=(x_test, y_test),\n",
    "                                        callbacks=callbacks)\n",
    "\n",
    "# histories[model] = models[model].fit(x_train, y_train, epochs=300, batch_size=32, workers=4, validation_data=(x_train, y_train))\n",
    "\n",
    "# 6. 모델 평가하기\n",
    "evaluate = modelv1.evaluate(x_test, y_test, batch_size=batch_size, verbose=1)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "colab_type": "code",
    "id": "FquYdJ59SDb_",
    "outputId": "29ae5bd6-557a-4a85-a8db-67e0e53b0582"
   },
   "outputs": [],
   "source": [
    "# 6. 모델 평가하기\n",
    "evaluate = modelv1.evaluate(x_test, y_test, batch_size=batch_size, verbose=1)\n",
    "print(evaluate)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "3jyXxNNjSSnb"
   },
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ZtIw2AGf9fu6"
   },
   "source": [
    "# version2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "id": "Y_wU9z4T69A5",
    "outputId": "6ee3d579-6f69-49dc-f9d0-6b939517a8ee"
   },
   "outputs": [],
   "source": [
    "inputs = Input(shape=(100, 1))\n",
    "\n",
    "modelv2 = DcnnNetv2(depth=201, growthRate=32, include_top=True, input_tensor=inputs, pooling='avg', classes=9)\n",
    "modelv2.summary()\n",
    "modelv2.name"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "awvyPBrqLXzj"
   },
   "source": [
    "# version2 Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "id": "-jkUc-9jLTlF",
    "outputId": "d334086f-1d42-4722-9bdf-fb8489ea970e",
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "if not os.path.isdir(save_path):\n",
    "    os.makedirs(save_path)\n",
    "\n",
    "batch_size = 64\n",
    "split_val = 0.3\n",
    "fit_shuffle = True\n",
    "optimizer = 'sgd'\n",
    "start_lr = 1e-2\n",
    "\n",
    "split_shuffle = False\n",
    "\n",
    "def lr_schedule(epoch):\n",
    "        lr = start_lr\n",
    "        if epoch > 225:\n",
    "            lr *= 1e-2\n",
    "        elif epoch > 150:\n",
    "            lr *= 1e-1\n",
    "        print('epoch: ', epoch)\n",
    "        print('Learning rate: ', lr)\n",
    "        return lr\n",
    "\n",
    "# 3. 모델 학습과정 설정하기\n",
    "sgd = SGD(lr=lr_schedule(0), decay=1e-4, momentum=0.9, nesterov=True)\n",
    "# adam = Adam(lr=lr_schedule(0))\n",
    "# nAdam = Nadam(lr=lr_schedule(0))\n",
    "\n",
    "\n",
    "# model.compile(optimizer=sgd, loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "\n",
    "\n",
    "# models[model].compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['accuracy',\n",
    "#                                                                                     single_class_precision(0), single_class_recall(0),\n",
    "#                                                                                     single_class_precision(1), single_class_recall(1),\n",
    "#                                                                                     single_class_precision(2), single_class_recall(2),\n",
    "#                                                                                     single_class_precision(3), single_class_recall(3),\n",
    "#                                                                                     single_class_precision(4), single_class_recall(4),\n",
    "#                                                                                     single_class_precision(5), single_class_recall(5),\n",
    "#                                                                                     single_class_precision(3), single_class_recall(6),\n",
    "#                                                                                     single_class_precision(4), single_class_recall(7),\n",
    "#                                                                                     single_class_precision(5), single_class_recall(8)])\n",
    "\n",
    "modelv2.compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['accuracy',\n",
    "                                                                         f1_m, recall_m, precision_m])\n",
    "\n",
    "csv_logger_path = save_path+f'/{modelv2.name}(batch_size={batch_size}, split_val={split_val}, fit_shuffle={fit_shuffle}, split_shuffle={split_shuffle}, optimizer={optimizer}, lr=s:{start_lr}, per:1e-1, e:(150, 225))_log.csv'\n",
    "if not os.path.isdir(os.path.dirname(csv_logger_path)):\n",
    "    os.makedirs(os.path.dirname(csv_logger_path))\n",
    "csv_logger = CSVLogger(csv_logger_path, append=True)\n",
    "\n",
    "lr_scheduler = LearningRateScheduler(schedule=lr_schedule)\n",
    "\n",
    "callbacks = [lr_scheduler, csv_logger]\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "# 4. 모델 학습시키기\n",
    "history = modelv2.fit(x_train, y_train, \n",
    "                    epochs=300, batch_size=batch_size, \n",
    "                    shuffle=fit_shuffle,\n",
    "#                     workers=1, \n",
    "                      validation_data=(x_test, y_test),\n",
    "                    callbacks=callbacks)\n",
    "\n",
    "# histories[model] = models[model].fit(x_train, y_train, epochs=300, batch_size=32, workers=4, validation_data=(x_train, y_train))\n",
    "\n",
    "# 6. 모델 평가하기\n",
    "evaluate = modelv2.evaluate(x_test, y_test, batch_size=batch_size, verbose=1)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 72
    },
    "colab_type": "code",
    "id": "dz5gIfVHbSJW",
    "outputId": "c2a98073-fc17-4f62-fe73-3af81086b854"
   },
   "outputs": [],
   "source": [
    "evaluate = modelv2.evaluate(x_test, y_test, batch_size=batch_size, verbose=1)\n",
    "print(\"loss     acc      f1      recall      precision\")\n",
    "print(evaluate)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 170
    },
    "colab_type": "code",
    "id": "qn-aGNVuO35L",
    "outputId": "2a0d756d-5860-4e55-d839-b9ca09ca341d"
   },
   "outputs": [],
   "source": [
    "epoch:  0\n",
    "Learning rate:  0.01\n",
    "Train on 2268 samples, validate on 972 samples\n",
    "Epoch 1/300\n",
    "epoch:  0\n",
    "Learning rate:  0.01\n",
    "2268/2268 [==============================] - 110s 48ms/step - loss: 0.5774 - acc: 0.7743 - f1_m: 0.7240 - recall_m: 0.6958 - precision_m: 0.7823 - val_loss: 8.3245 - val_acc: 0.3313 - val_f1_m: 0.3313 - val_recall_m: 0.3313 - val_precision_m: 0.3313\n",
    "Epoch 2/300\n",
    "epoch:  1\n",
    "Learning rate:  0.01\n",
    "2268/2268 [==============================] - 14s 6ms/step - loss: 0.1662 - acc: 0.9374 - f1_m: 0.9380 - recall_m: 0.9361 - precision_m: 0.9399 - val_loss: 0.8003 - val_acc: 0.7840 - val_f1_m: 0.7828 - val_recall_m: 0.7809 - val_precision_m: 0.7848\n",
    "Epoch 3/300\n",
    "epoch:  2\n",
    "Learning rate:  0.01\n",
    "2268/2268 [==============================] - 15s 7ms/step - loss: 0.1239 - acc: 0.9612 - f1_m: 0.9616 - recall_m: 0.9603 - precision_m: 0.9628 - val_loss: 5.2734 - val_acc: 0.3539 - val_f1_m: 0.3538 - val_recall_m: 0.3313 - val_precision_m: 0.3803\n",
    "Epoch 4/300\n",
    "epoch:  3\n",
    "Learning rate:  0.01\n",
    " 832/2268 [==========>...................] - ETA: 8s - loss: 0.1580 - acc: 0.9363 - f1_m: 0.9380 - recall_m: 0.9303 - precision_m: 0.9462"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "_JtOOvKnfqt-"
   },
   "source": [
    "# version 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "OlMAN-jlfphr"
   },
   "outputs": [],
   "source": [
    "img_input = layers.Input(shape=input_shape)\n",
    "\n",
    "nChannels = 2 * growthRate\n",
    "bn_axis = 2\n",
    "\n",
    "x = layers.ZeroPadding1D(padding=1)(img_input)\n",
    "x = layers.Conv1D(nChannels, 3, strides=1, use_bias=False, name='conv1/conv')(x)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Bg0XYdPOSDA4"
   },
   "source": [
    "# 실험 결과\n",
    "\n",
    "## version1\n",
    "\n",
    "depth = 100\n",
    "\n",
    "### concatenate\n",
    "```\n",
    "start lr = 0.01\n",
    "    972/972 [==============================] - 1s 975us/step\n",
    "    [0.26581049285380653, 0.9259259259259259]\n",
    "\n",
    "start lr = 0.1\n",
    "    972/972 [==============================] - 1s 976us/step\n",
    "    [0.43105335702636727, 0.9176954732510288]\n",
    "```\n",
    "\n",
    "\n",
    "### add\n",
    "\n",
    "```\n",
    "start lr = 0.01\n",
    "    972/972 [==============================] - 1s 520us/step\n",
    "    [0.3993826474181909, 0.8713991769547325]\n",
    "\n",
    "start lr = 0.1\n",
    "\n",
    "학습되지 않음\n",
    "```\n",
    "\n",
    "## version 2\n",
    "\n",
    "201 layer\n",
    "\n",
    "### concatenate\n",
    "\n",
    "```\n",
    "start lr = 0.1\n",
    "    972/972 [==============================] - 1s 1ms/step\n",
    "    [0.5396262257064811, 0.8446502060066035]\n",
    "Data normalization\n",
    "    972/972 [==============================] - 1s 1ms/step\n",
    "    loss     acc      f1      recall      precision\n",
    "    [2.230289366755466, 0.6306584359687052, 0.6331204848034392, 0.6255144030468944, 0.6410546889030394]\n",
    "min-max normalization\n",
    "    972/972 [==============================] - 1s 1ms/step\n",
    "    loss     acc      f1      recall      precision\n",
    "    [2.013901627603382, 0.6707818932494019, 0.6677523871017582, 0.6615226339901426, 0.674254747329916]\n",
    "\n",
    "start lr = 0.01\n",
    "    972/972 [==============================] - 1s 1ms/step\n",
    "    loss     acc      f1      recall      precision\n",
    "    [0.5466709941502952, 0.8631687245251219, 0.8630309512095197, 0.8621399179407598, 0.8639362478943028]\n",
    "\n",
    "```\n",
    "\n",
    "\n",
    "```python \n",
    "start lr = 0.01\n",
    "x = layers.ZeroPadding1D(padding=3)(img_input)\n",
    "x = layers.Conv1D(nChannels, 7, strides=1, use_bias=False, name='conv1/conv')(x)\n",
    "\n",
    "# x = layers.BatchNormalization(\n",
    "#     axis=bn_axis, epsilon=1.001e-5, name='conv1/bn')(x)\n",
    "# x = layers.Activation('relu', name='conv1/relu')(x)\n",
    "\n",
    "# x = layers.ZeroPadding1D(padding=1)(x)\n",
    "# x = layers.MaxPooling1D(3, strides=2, name='pool1')(x)\n",
    "\n",
    "    972/972 [==============================] - 2s 2ms/step\n",
    "    loss     acc      f1      recall      precision\n",
    "    [0.22257122489547493, 0.9454732510288066, 0.9459430514049137, 0.9454732510288066, 0.9464204066084245]\n",
    "\n",
    "x = layers.ZeroPadding1D(padding=3)(img_input)\n",
    "x = layers.Conv1D(nChannels, 7, strides=2, use_bias=False, name='conv1/conv')(x)\n",
    "\n",
    "# x = layers.BatchNormalization(\n",
    "#     axis=bn_axis, epsilon=1.001e-5, name='conv1/bn')(x)\n",
    "# x = layers.Activation('relu', name='conv1/relu')(x)\n",
    "\n",
    "# x = layers.ZeroPadding1D(padding=1)(x)\n",
    "# x = layers.MaxPooling1D(3, strides=2, name='pool1')(x)\n",
    "\n",
    "    972/972 [==============================] - 1s 1ms/step\n",
    "    loss     acc      f1      recall      precision\n",
    "    [0.5289181569809623, 0.8837448559670782, 0.8851272049264162, 0.8816872427983539, 0.8886497324876824]\n",
    "\n",
    "```\n",
    "\n",
    "### add\n",
    "\n",
    "```\n",
    "start lr = 0.01\n",
    "    972/972 [==============================] - 1s 923us/step\n",
    "    [0.8464200724790125, 0.7129629632082496]\n",
    "\n",
    "\n",
    "start lr = 0.1\n",
    "    972/972 [==============================] - 1s 915us/step\n",
    "    [1.0798768680772663, 0.49074074074074076]\n",
    "```\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "BLgdTKS9XtSV"
   },
   "outputs": [],
   "source": [
    "\n",
    "if not os.path.isdir(save_path):\n",
    "    os.mkdir(save_path)\n",
    "\n",
    "params = {\n",
    "    'batch_size' : [2, 4, 8, 16, 32, 64, 128, 256],\n",
    "    'fit_shuffle' : [False, True],\n",
    "    'split_shuffle' : [False, True],\n",
    "    'start_lr' : [1e-1, 1e-2, 1e-3],\n",
    "    'optimizer' : ['sgd','adam','nAdam'],\n",
    "    'split_val' : [0.2, 0.3, 0.4]\n",
    "}\n",
    "\n",
    "batch_size = params['batch_size'][0]\n",
    "fit_shuffle = params['fit_shuffle'][0]\n",
    "split_shuffle = params['split_shuffle'][0]\n",
    "start_lr = params['start_lr'][0]\n",
    "optimizer = params['optimizer'][0]\n",
    "split_val = params['split_val'][0]\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "p_s-3f2pReqM"
   },
   "outputs": [],
   "source": [
    "from keras.models import Sequential #스택형 모델\n",
    "from keras.models import Model #함수형 모델\n",
    "\n",
    "from keras.layers import Dense, Embedding, LSTM, Input\n",
    "from keras.layers import Flatten\n",
    "from keras.layers import Conv1D, MaxPooling1D, GlobalMaxPooling1D, Dropout"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "1Gw7r_DpX-rs"
   },
   "outputs": [],
   "source": [
    "from keras import backend as K\n",
    "\n",
    "# 특정 클래스에 대한 정밀도\n",
    "def single_class_precision(interesting_class_id):\n",
    "    def prec(y_true, y_pred):\n",
    "        class_id_true = K.argmax(y_true, axis=-1)\n",
    "        class_id_pred = K.argmax(y_pred, axis=-1)\n",
    "        precision_mask = K.cast(K.equal(class_id_pred, interesting_class_id), 'int32')\n",
    "        class_prec_tensor = K.cast(K.equal(class_id_true, class_id_pred), 'int32') * precision_mask\n",
    "        class_prec = K.cast(K.sum(class_prec_tensor), 'float32') / K.cast(K.maximum(K.sum(precision_mask), 1), 'float32')\n",
    "        return class_prec\n",
    "    return prec\n",
    "\n",
    "\n",
    "# 특정 클래스에 대한 재현율\n",
    "def single_class_recall(interesting_class_id):\n",
    "    def recall(y_true, y_pred):\n",
    "        class_id_true = K.argmax(y_true, axis=-1)\n",
    "        class_id_pred = K.argmax(y_pred, axis=-1)\n",
    "        recall_mask = K.cast(K.equal(class_id_true, interesting_class_id), 'int32')\n",
    "        class_recall_tensor = K.cast(K.equal(class_id_true, class_id_pred), 'int32') * recall_mask\n",
    "        class_recall = K.cast(K.sum(class_recall_tensor), 'float32') / K.cast(K.maximum(K.sum(recall_mask), 1), 'float32')\n",
    "        return class_recall\n",
    "    return recall\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "uZa_m3V5_a-Z"
   },
   "outputs": [],
   "source": [
    "models = {}\n",
    "\n",
    "model_type = {\n",
    "    \"MLP\" : {},\n",
    "    \"LSTM\" : {},\n",
    "    \"CONV1D\" : {},\n",
    "    \"CONV1D+LSTM\" : {},\n",
    "    \"LSTM+CONV1D\" : {}\n",
    "}\n",
    "\n",
    "histories = {}\n",
    "evaluates = {}\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "OsbzcSNPubub"
   },
   "source": [
    "#MLP\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "iyC9i6cgFK1-"
   },
   "outputs": [],
   "source": [
    "\n",
    "for ii in range(6):\n",
    "    iii = 0\n",
    "    inputs = Input(shape=(1, 100))\n",
    "    x = Dense(10, activation='relu')(inputs)\n",
    "    for iii in range(ii):\n",
    "        x = Dense(10, activation='relu')(x)\n",
    "    y = Dense(9, activation='softmax')(x)\n",
    "    model = Model(inputs=inputs, outputs=y)\n",
    "    models[f\"[MLP]{ii+1}_layer\"] = model\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "5k2Zv6BeuwUe"
   },
   "source": [
    "#LSTM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "AZ0AkepIuzmR"
   },
   "outputs": [],
   "source": [
    "\n",
    "for ii in range(6):\n",
    "    iii = 0\n",
    "    inputs = Input(shape=(1,100))\n",
    "    x = LSTM(10, return_sequences=True, activation=\"relu\")(inputs)\n",
    "    for iii in range(ii):\n",
    "        if iii != ii - 1:\n",
    "            x = LSTM(10, return_sequences=True, activation=\"relu\")(x)\n",
    "        else:\n",
    "            x = LSTM(10, return_sequences=True, activation=\"relu\")(x)\n",
    "    y = Dense(6, activation=\"softmax\")(x)\n",
    "\n",
    "    models[f\"[LSTM]{ii+1}_layer\"] = Model(inputs=inputs, outputs=y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "pt0wBtJ0uf9D"
   },
   "source": [
    "#Conv1D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "GQ11U9_cufdc"
   },
   "outputs": [],
   "source": [
    "\n",
    "for ii in range(6):\n",
    "    inputs = Input(shape=(1,100))\n",
    "    x = Conv1D(10, 1, activation=\"relu\", padding=\"same\", strides=1)(inputs)\n",
    "    for iii in range(ii):\n",
    "        x = Conv1D(100, 3, activation=\"relu\", padding=\"same\", strides=1)(x)\n",
    "    y = Dense(9, activation='softmax')(x)\n",
    "\n",
    "    model = Model(inputs=inputs, outputs=y)\n",
    "    models[f\"[Conv1D]{ii+1}_layer\"] = model\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "datVAgzhul62"
   },
   "source": [
    "#Conv1D + LSTM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "x-7dCgy-ulXn"
   },
   "outputs": [],
   "source": [
    "for ii in range(6):\n",
    "    inputs = Input(shape=(1,100))\n",
    "    x = Conv1D(10, 1, activation=\"relu\", padding=\"valid\", strides=1)(inputs)\n",
    "    x = LSTM(10, activation=\"relu\", return_sequences=True)(x)\n",
    "    for iii in range(ii):\n",
    "        x = Conv1D(10, 1, activation=\"relu\", padding=\"valid\", strides=1)(x)\n",
    "        x = LSTM(10, activation=\"relu\", return_sequences=True)(x)  \n",
    "    y = Dense(9, activation='softmax')(x)\n",
    "\n",
    "    model = Model(inputs=inputs, outputs=y)\n",
    "    models[f\"[Conv1D + LSTM]{ii+1}_layer\"] = model\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "b86CS4nLulQ3"
   },
   "source": [
    "#LSTM + Conv1D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 125
    },
    "colab_type": "code",
    "id": "TP2dAIciuuBq",
    "outputId": "384fde54-57ab-460c-9912-684576ade9e4"
   },
   "outputs": [],
   "source": [
    " for ii in range(6):\n",
    "    inputs = Input(shape=(1,100))\n",
    "    x = LSTM(10, activation=\"relu\", return_sequences=True)(inputs)\n",
    "    x = Conv1D(10, 1, activation=\"relu\", padding=\"valid\", strides=1)(x)\n",
    "    for iii in range(ii):\n",
    "        x = LSTM(10, activation=\"relu\", return_sequences=True)(x)  \n",
    "        x = Conv1D(10, 1, activation=\"relu\", padding=\"valid\", strides=1)(x)\n",
    "    y = Dense(9, activation='softmax')(x)\n",
    "\n",
    "    model = Model(inputs=inputs, outputs=y)\n",
    "    models[f\"[LSTM + Conv1D]{ii+1}_layer\"] = model\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "3VWYJUHQw88a"
   },
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Cze9MLtlw845"
   },
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 125
    },
    "colab_type": "code",
    "id": "8iTEQH-kJmKa",
    "outputId": "3ea2f8b9-68d2-4766-b340-48de2186cca6"
   },
   "outputs": [],
   "source": [
    "models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "3WPzvqKLCWCb"
   },
   "outputs": [],
   "source": [
    "for model in models:\n",
    "    print(model)\n",
    "    models[model].summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "id": "lkBi8hG_Renb",
    "outputId": "ae34035d-db04-43a8-cfc7-ba99da73be7f"
   },
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "for model in models:\n",
    "\n",
    "    def lr_schedule(epoch):\n",
    "        lr = start_lr\n",
    "        if epoch > 225:\n",
    "            lr *= 1e-2\n",
    "        elif epoch > 150:\n",
    "            lr *= 1e-1\n",
    "        print('epoch: ', epoch)\n",
    "        print('Learning rate: ', lr)\n",
    "        return lr\n",
    "\n",
    "    # 3. 모델 학습과정 설정하기\n",
    "    if optimizer == 'sgd':\n",
    "        sgd = SGD(lr=lr_schedule(0), decay=1e-4, momentum=0.9, nesterov=True)\n",
    "    elif optimizer == 'adam':\n",
    "        adam = Adam(lr=lr_schedule(0))\n",
    "    elif optimizer == 'nAdam':\n",
    "        nAdam = Nadam(lr=lr_schedule(0))\n",
    "\n",
    "\n",
    "    # model.compile(optimizer=sgd, loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "\n",
    "\n",
    "    # models[model].compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['accuracy',\n",
    "    #                                                                                     single_class_precision(0), single_class_recall(0),\n",
    "    #                                                                                     single_class_precision(1), single_class_recall(1),\n",
    "    #                                                                                     single_class_precision(2), single_class_recall(2),\n",
    "    #                                                                                     single_class_precision(3), single_class_recall(3),\n",
    "    #                                                                                     single_class_precision(4), single_class_recall(4),\n",
    "    #                                                                                     single_class_precision(5), single_class_recall(5),\n",
    "    #                                                                                     single_class_precision(3), single_class_recall(6),\n",
    "    #                                                                                     single_class_precision(4), single_class_recall(7),\n",
    "    #                                                                                     single_class_precision(5), single_class_recall(8)])\n",
    "\n",
    "    models[model].compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['accuracy'])\n",
    "    \n",
    "    csv_logger_path = save_path+f'/{model}(batch_size={batch_size}, split_val={split_val}, fit_shuffle={fit_shuffle}, split_shuffle={split_shuffle}, optimizer={optimizer}, lr=s:{start_lr}, per:1e-1, e:(150, 225))_log.csv'\n",
    "    if not os.path.isdir(os.path.dirname(csv_logger_path)):\n",
    "        os.makedirs(os.path.dirname(csv_logger_path))\n",
    "    csv_logger = CSVLogger(csv_logger_path, append=True)\n",
    "\n",
    "    lr_scheduler = LearningRateScheduler(schedule=lr_schedule)\n",
    "\n",
    "    callbacks = [lr_scheduler, csv_logger]\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "    # 4. 모델 학습시키기\n",
    "    histories[model] = models[model].fit(x_train, y_train, \n",
    "                                         epochs=300, batch_size=batch_size, \n",
    "                                         shuffle=fit_shuffle,\n",
    "                                         workers=4, validation_data=(x_test, y_test),\n",
    "                                         callbacks=callbacks)\n",
    "    \n",
    "    # histories[model] = models[model].fit(x_train, y_train, epochs=300, batch_size=32, workers=4, validation_data=(x_train, y_train))\n",
    "\n",
    "    # 6. 모델 평가하기\n",
    "    evaluates[model] = models[model].evaluate(x_test, y_test, batch_size=batch_size, verbose=1)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 269
    },
    "colab_type": "code",
    "id": "TLIdaxrQaJxX",
    "outputId": "1e9b6beb-5ec5-40cc-b9c2-54b0020bcf1c"
   },
   "outputs": [],
   "source": [
    "evaluates"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "l2KkJdfuD8UZ"
   },
   "outputs": [],
   "source": [
    "# 5. 학습과정 살펴보기\n",
    "\n",
    "for history in histories:\n",
    "    fig = plt.figure(figsize=(15,5))\n",
    "\n",
    "    ax = fig.add_subplot(1,2,1)\n",
    "    ax.plot(histories[history].history['loss'], 'y', label=\"train loss\")\n",
    "    ax.plot(histories[history].history['val_loss'], 'r', label=f\"val loss({format(evaluates[history][0], '0.4f')})\")\n",
    "    ax.set_ylabel('loss')\n",
    "    ax.grid(True)\n",
    "    ax.legend()\n",
    "    ax.set_title(f\"{history}\")\n",
    "\n",
    "    ax = fig.add_subplot(1,2,2)\n",
    "    ax.plot(histories[history].history['acc'], 'y', label=\"train acc\")\n",
    "    ax.plot(histories[history].history['val_acc'], 'r', label=f\"val acc({format(evaluates[history][1],'0.4f')})\")\n",
    "    ax.set_ylabel('accuray')\n",
    "    ax.grid(True)\n",
    "    ax.legend()\n",
    "    ax.set_title(f\"{history}\")\n",
    "\n",
    "    save_type = \"history\"\n",
    "    history_path = f\"{save_path}/{save_type}\"\n",
    "\n",
    "    if not os.path.isdir(history_path):\n",
    "        os.mkdir(history_path)\n",
    "    plt.savefig(history_path + \"/\" + history)\n",
    "\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 172
    },
    "colab_type": "code",
    "id": "LNP-CW0ETOVP",
    "outputId": "a7760fac-7115-42f6-9855-3860bf986d2f"
   },
   "outputs": [],
   "source": [
    "*histories['[MLP]1_layer'].history[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "apf6Nyv-GpD2"
   },
   "outputs": [],
   "source": [
    "for history in histories:\n",
    "\n",
    "    precs = [x for x in histories[history].history.keys() if \"prec\" in x if \"val\" not in x]\n",
    "    recalls = [x for x in histories[history].history.keys() if \"recall\" in x if \"val\" not in x]\n",
    "\n",
    "    val_precs = [x for x in histories[history].history.keys() if \"prec\" in x if \"val\" in x]\n",
    "    val_recalls = [x for x in histories[history].history.keys() if \"recall\" in x if \"val\" in x]\n",
    "\n",
    "    fig = plt.figure(figsize=(30,5))\n",
    "\n",
    "    for prec in precs:\n",
    "        ax = fig.add_subplot(1,4,1)\n",
    "        ax.plot(histories[history].history[prec], label=prec)\n",
    "        ax.set_xlabel('epoch')\n",
    "        ax.set_ylabel('precision')\n",
    "        ax.grid(True)\n",
    "        ax.legend()\n",
    "        ax.set_title(f\"{history}\")\n",
    "        ax.set_ylim(0, 1.0)\n",
    "        \n",
    "    for recall in recalls:\n",
    "        ax = fig.add_subplot(1,4,2)\n",
    "        ax.plot(histories[history].history[recall], label=recall)\n",
    "        ax.set_xlabel('epoch')\n",
    "        ax.set_ylabel('recall')\n",
    "        ax.grid(True)\n",
    "        ax.legend()\n",
    "        ax.set_title(f\"{history}\")  \n",
    "        ax.set_ylim(0, 1.0)\n",
    "\n",
    "    for prec in val_precs:\n",
    "        ax = fig.add_subplot(1,4,3)\n",
    "        ax.plot(histories[history].history[prec], label=prec)\n",
    "        ax.set_xlabel('epoch')\n",
    "        ax.set_ylabel('val_precision')\n",
    "        ax.grid(True)\n",
    "        ax.legend()\n",
    "        ax.set_title(f\"{history}\")\n",
    "        ax.set_ylim(0, 1.0)\n",
    "        \n",
    "    for recall in val_recalls:\n",
    "        ax = fig.add_subplot(1,4,4)\n",
    "        ax.plot(histories[history].history[recall], label=recall)\n",
    "        ax.set_xlabel('epoch')\n",
    "        ax.set_ylabel('val_recall')\n",
    "        ax.grid(True)\n",
    "        ax.legend()\n",
    "        ax.set_title(f\"{history}\")  \n",
    "        ax.set_ylim(0, 1.0)\n",
    "\n",
    "    save_type = \"precision,recall\"\n",
    "    history_path = f\"{save_path}/{save_type}\"\n",
    "\n",
    "    if not os.path.isdir(history_path):\n",
    "        os.mkdir(history_path)\n",
    "    plt.savefig(history_path + \"/\" + history)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "nTdZoxM_YJL-"
   },
   "outputs": [],
   "source": [
    "fig = plt.figure(figsize=(15,7))\n",
    "\n",
    "for id, evaluate in enumerate(evaluates):\n",
    "    # print(evaluates[evaluate])\n",
    "    # print(evaluates[evaluate][2:])\n",
    "    metrics = np.array(evaluates[evaluate][2:])\n",
    "\n",
    "    idx = np.linspace(0, 11, 12) \n",
    "    # print(idx)\n",
    "    precision = metrics[(idx % 2) == 0]\n",
    "    recall = metrics[((idx+1) % 2) == 0]\n",
    "\n",
    "    # print(precision)\n",
    "    # print(recall)\n",
    "    # print(\"\\n\")\n",
    "    # print(\"\\n\")\n",
    "    # print(\"\\n\")\n",
    "    # print(\"\\n\")\n",
    "\n",
    "\n",
    "    N = 6\n",
    "    ind = np.arange(N)\n",
    "    width = 0.35\n",
    "    \n",
    "    ax = fig.add_subplot(2,3,id+1)\n",
    "\n",
    "    # fig, ax = plt.subplots()\n",
    "    prec_bar = ax.bar(ind, precision, width, color='r')\n",
    "    recall_bar = ax.bar(ind + width, recall, width, color='y')\n",
    "\n",
    "    ax.set_ylabel('Scores')\n",
    "    ax.set_title('Precision and Recall')\n",
    "    ax.set_xticks(ind + width / 2)\n",
    "    ax.set_xticklabels(('C1', 'C2', 'C3', 'C4', 'C5', 'C6'))\n",
    "\n",
    "    ax.legend((prec_bar[0], recall_bar[0]), ('Precision', 'Recall'))\n",
    "    # ax.grid(True)\n",
    "    ax.set_title(evaluate)\n",
    "    ax.set_ylim(0,1.0)\n",
    "\n",
    "        \n",
    "save_type = \"precision,recall_graph\"\n",
    "history_path = f\"{save_path}/{save_type}\"\n",
    "\n",
    "if not os.path.isdir(history_path):\n",
    "    os.mkdir(history_path)\n",
    "plt.savefig(history_path + \"/\" + evaluate)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "MQZwJqsBYF6j"
   },
   "outputs": [],
   "source": [
    "``"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "8L9-hm1_43vO",
    "outputId": "218eedc8-2af7-49ab-a65a-65a74831d8ac"
   },
   "outputs": [],
   "source": [
    "x_train = x_train.reshape(-1, 1, 100)\n",
    "x_test  = x_test.reshape(-1, 1, 100)\n",
    "y_train = y_train.reshape(-1,  6)\n",
    "y_test = y_test.reshape(-1, 6)\n",
    "print(x_train.shape, x_test.shape, y_train.shape, y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "UOrPqCLTDxl4"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 275
    },
    "colab_type": "code",
    "id": "m80D0FrrISs3",
    "outputId": "56f6f592-b172-49c0-aec2-d369ef68b2b3"
   },
   "outputs": [],
   "source": [
    "inputs = Input(shape=(1,100))\n",
    "x = LSTM(10, return_sequences=True, activation=\"relu\")(inputs)\n",
    "y = Dense(6, activation=\"softmax\")(x)\n",
    "model = Model(inputs=inputs, outputs=y)\n",
    "\n",
    "\n",
    "model.summary()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "id": "52oawxk5RekY",
    "outputId": "da0de6b0-c54c-47cd-8722-922d0ab5b378"
   },
   "outputs": [],
   "source": [
    "# 2. 모델 구성하기\n",
    "# model = Sequential()\n",
    "# model.add(LSTM(128))\n",
    "# # model.add(Flatten())\n",
    "# model.add(Dense(6, activation='softmax'))\n",
    "\n",
    "\n",
    "models = {}\n",
    "histories = {}\n",
    "evaluates = {}\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "for model in models:\n",
    "    print(model)\n",
    "    models[model].summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 386
    },
    "colab_type": "code",
    "id": "UMU5OAb_N6Kh",
    "outputId": "7195b0a7-d3c4-43bb-f9b7-5d91d913fef5"
   },
   "outputs": [],
   "source": [
    "inputs = Input(shape=(1,100))\n",
    "x = Conv1D(10, 1, activation=\"relu\", padding=\"valid\", strides=1)(inputs)\n",
    "x = LSTM(10, activation=\"relu\",return_sequences=True)(x)\n",
    "x = Conv1D(10, 1, activation=\"relu\", padding=\"valid\", strides=1)(x)\n",
    "x = LSTM(10, activation=\"relu\")(x)  \n",
    "y = Dense(6, activation='softmax')(x)\n",
    "model = Model(inputs=inputs, outputs=y)\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "id": "mXnija4fK0tR",
    "outputId": "d48b7d05-6a3b-4b51-a97d-b92f26bb021c"
   },
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "models = {}\n",
    "histories = {}\n",
    "evaluates = {}\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "for model in models:\n",
    "    print(\"\\n\"+model)\n",
    "    models[model].summary()\n",
    "# model.add(Conv1D(50,\n",
    "#                  1,\n",
    "#                  input_shape=(1,100),\n",
    "#                  padding='valid',\n",
    "#                  activation='relu',\n",
    "#                  strides=1))\n",
    "# # model.add(GlobalMaxPooling1D())\n",
    "# model.add(LSTM(50))\n",
    "# model.add(Dense(6, activation='softmax'))\n",
    "\n",
    "# model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "6nD8PJcbGWfC"
   },
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "# 3. 모델 학습과정 설정하기\n",
    "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "\n",
    "# 4. 모델 학습시키기\n",
    "hist = model.fit(x_train, y_train, epochs=10, batch_size=32, validation_data=(x_test, y_test))\n",
    "\n",
    "# 5. 학습과정 살펴보기\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "fig = plt.figure(figsize=(10,5))\n",
    "\n",
    "ax = fig.add_subplot(1,2,1)\n",
    "ax.plot(hist.history['loss'], 'y', label='train loss')\n",
    "ax.plot(hist.history['val_loss'], 'r', label='val loss')\n",
    "ax.set_ylabel('loss')\n",
    "ax.grid(True)\n",
    "ax.legend()\n",
    "\n",
    "ax = fig.add_subplot(1,2,2)\n",
    "ax.plot(hist.history['acc'], 'b', label='train acc')\n",
    "ax.plot(hist.history['val_acc'], 'g', label='val acc')\n",
    "ax.set_ylabel('accuray')\n",
    "ax.grid(True)\n",
    "ax.legend()\n",
    "\n",
    "\n",
    "\n",
    "# fig, loss_ax = plt.subplots()\n",
    "\n",
    "# acc_ax = loss_ax.twinx()\n",
    "\n",
    "# loss_ax.plot(hist.history['loss'], 'y', label='train loss')\n",
    "# loss_ax.plot(hist.history['val_loss'], 'r', label='val loss')\n",
    "# loss_ax.set_ylim([0.0, 3.0])\n",
    "\n",
    "# acc_ax.plot(hist.history['acc'], 'b', label='train acc')\n",
    "# acc_ax.plot(hist.history['val_acc'], 'g', label='val acc')\n",
    "# acc_ax.set_ylim([0.0, 1.0])\n",
    "\n",
    "# loss_ax.set_xlabel('epoch')\n",
    "# loss_ax.set_ylabel('loss')\n",
    "# acc_ax.set_ylabel('accuray')\n",
    "\n",
    "# loss_ax.legend(loc='upper left')\n",
    "# acc_ax.legend(loc='lower left')\n",
    "\n",
    "# plt.show()\n",
    "\n",
    "# 6. 모델 평가하기\n",
    "loss_and_metrics = model.evaluate(x_test, y_test, batch_size=32)\n",
    "print('## evaluation loss and_metrics ##')\n",
    "print(loss_and_metrics)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 247
    },
    "colab_type": "code",
    "id": "GPc317BeUxfj",
    "outputId": "adcdb115-a4a3-4860-a48d-1e6e0a2f08c9"
   },
   "outputs": [],
   "source": [
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 958
    },
    "colab_type": "code",
    "id": "ld0YvmINUxzr",
    "outputId": "04a838bc-8b34-4128-a9fd-6a36931c4e0f"
   },
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "model.add(Conv1D(50,\n",
    "                 1,\n",
    "                 input_shape=(1,100),\n",
    "                 padding='valid',\n",
    "                 activation='relu',\n",
    "                 strides=1))\n",
    "# model.add(GlobalMaxPooling1D())\n",
    "model.add(LSTM(50))\n",
    "model.add(Dense(6, activation='softmax'))\n",
    "\n",
    "model.summary()\n",
    "\n",
    "# 3. 모델 학습과정 설정하기\n",
    "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "\n",
    "# 4. 모델 학습시키기\n",
    "hist = model.fit(x_train, y_train, epochs=10, batch_size=64, validation_data=(x_test, y_test))\n",
    "\n",
    "# 5. 학습과정 살펴보기\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "fig, loss_ax = plt.subplots()\n",
    "\n",
    "acc_ax = loss_ax.twinx()\n",
    "\n",
    "loss_ax.plot(hist.history['loss'], 'y', label='train loss')\n",
    "loss_ax.plot(hist.history['val_loss'], 'r', label='val loss')\n",
    "loss_ax.set_ylim([0.0, 3.0])\n",
    "\n",
    "acc_ax.plot(hist.history['acc'], 'b', label='train acc')\n",
    "acc_ax.plot(hist.history['val_acc'], 'g', label='val acc')\n",
    "acc_ax.set_ylim([0.0, 1.0])\n",
    "\n",
    "loss_ax.set_xlabel('epoch')\n",
    "loss_ax.set_ylabel('loss')\n",
    "acc_ax.set_ylabel('accuray')\n",
    "\n",
    "loss_ax.legend(loc='upper left')\n",
    "acc_ax.legend(loc='lower left')\n",
    "\n",
    "plt.show()\n",
    "\n",
    "# 6. 모델 평가하기\n",
    "loss_and_metrics = model.evaluate(x_test, y_test, batch_size=64)\n",
    "print('## evaluation loss and_metrics ##')\n",
    "print(loss_and_metrics)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "rf0mstPzcqmF"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 407
    },
    "colab_type": "code",
    "id": "fgJET27RjWTv",
    "outputId": "56298d1c-8730-42e1-e35f-e1909b4e2a0e"
   },
   "outputs": [],
   "source": [
    "\n",
    "model = Sequential([\n",
    "                    # Input(shape=(1,100)),\n",
    "                    Conv1D(50,1, input_shape=(1,100) ,padding='valid',activation='relu',strides=1),\n",
    "                    Conv1D(50,1, padding='valid',activation='relu',strides=1),\n",
    "                    Conv1D(50,1, padding='valid',activation='relu',strides=1),\n",
    "                    Conv1D(50,1, padding='valid',activation='relu',strides=1),\n",
    "                    GlobalMaxPooling1D(),\n",
    "                    Dense(50, activation='relu'),\n",
    "                    # model.add(Dropout(0.2))\n",
    "                    Dense(6, activation='softmax')\n",
    "\n",
    "])\n",
    "\n",
    "model.summary()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "id": "CE3nu0OtjWRc",
    "outputId": "e2a4298a-d377-4bef-b534-36963170460c"
   },
   "outputs": [],
   "source": [
    "# model = Sequential()\n",
    "\n",
    "# # model.add(Embedding(10, 128, input_length=100))\n",
    "# # model.add(Dropout(0.2))\n",
    "# model.add(Input(shape=(1,100)))\n",
    "# model.add(Conv1D(100,\n",
    "#                  1,\n",
    "#                  padding='valid',\n",
    "#                  activation='relu',\n",
    "#                  strides=1))\n",
    "# model.add(GlobalMaxPooling1D())\n",
    "# model.add(Dense(128, activation='relu'))\n",
    "# # model.add(Dropout(0.2))\n",
    "# model.add(Dense(6, activation='softmax'))\n",
    "\n",
    "\n",
    "\n",
    "# 3. 모델 학습과정 설정하기\n",
    "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "\n",
    "# 4. 모델 학습시키기\n",
    "hist = model.fit(x_train, y_train, epochs=10, batch_size=64, validation_data=(x_test, y_test))\n",
    "\n",
    "# 5. 학습과정 살펴보기\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "\n",
    "fig = plt.figure(figsize=(10,5))\n",
    "\n",
    "ax = fig.add_subplot(1,2,1)\n",
    "ax.plot(hist.history['loss'], 'y', label='train loss')\n",
    "ax.plot(hist.history['val_loss'], 'r', label='val loss')\n",
    "ax.set_ylabel('loss')\n",
    "ax.grid(True)\n",
    "ax.legend()\n",
    "\n",
    "ax = fig.add_subplot(1,2,2)\n",
    "ax.plot(hist.history['acc'], 'b', label='train acc')\n",
    "ax.plot(hist.history['val_acc'], 'g', label='val acc')\n",
    "ax.set_ylabel('accuray')\n",
    "ax.grid(True)\n",
    "ax.legend()\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "# fig, loss_ax = plt.subplots()\n",
    "\n",
    "# acc_ax = loss_ax.twinx()\n",
    "\n",
    "# loss_ax.plot(hist.history['loss'], 'y', label='train loss')\n",
    "# loss_ax.plot(hist.history['val_loss'], 'r', label='val loss')\n",
    "# loss_ax.set_ylim([0.0, 3.0])\n",
    "\n",
    "# acc_ax.plot(hist.history['acc'], 'b', label='train acc')\n",
    "# acc_ax.plot(hist.history['val_acc'], 'g', label='val acc')\n",
    "# acc_ax.set_ylim([0.0, 1.0])\n",
    "\n",
    "# loss_ax.set_xlabel('epoch')\n",
    "# loss_ax.set_ylabel('loss')\n",
    "# acc_ax.set_ylabel('accuray')\n",
    "\n",
    "# loss_ax.legend(loc='upper left')\n",
    "# acc_ax.legend(loc='lower left')\n",
    "\n",
    "# plt.show()\n",
    "\n",
    "# 6. 모델 평가하기\n",
    "loss_and_metrics = model.evaluate(x_test, y_test, batch_size=64)\n",
    "print('## evaluation loss and_metrics ##')\n",
    "print(loss_and_metrics)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "B4bNYV8ijYRS"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [
    "KjCs0sUyFeeV",
    "RfsjvrWu9jjn"
   ],
   "name": "1D DenseNet.ipynb",
   "provenance": []
  },
  "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": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}