exercise3_solution.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Exercise 3\n",
    "\n",
    "Work on this before the next lecture on 19 April. We will talk about questions, comments, and solutions during the exercise after the third lecture.\n",
    "\n",
    "Please do form study groups! When you do, make sure you can explain everything in your own words, do not simply copy&paste from others.\n",
    "\n",
    "The solutions to a lot of these problems can probably be found with Google. Please don't. You will not learn a lot by copy&pasting from the internet.\n",
    "\n",
    "If you want to get credit/examination on this course please upload your work to your GitHub repository for this course before the next lecture starts and post a link to your repository in [this thread](https://github.com/wildtreetech/advanced-computing-2018/issues/7). If you worked on things together with others please add their names to the notebook so we can see who formed groups.\n",
    "\n",
    "The overall idea of this exercise is to get started and the nto freely experiment with the building blocks fo keras."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# The Dataset\n",
    "\n",
    "To get going we will use a dataset which contains images of fashion items. It was created by [Zalanod research](https://github.com/zalandoresearch/fashion-mnist) to provide an alternative to the old MNIST digits dataset. Fashion MNIST is small like MNIST (28x28 pixel images), good size (60000 examples), and significantly harder than MNIST.\n",
    "\n",
    "There are ten classes (or types) of items:\n",
    "\n",
    "| Label | Description |\n",
    "| --- | --- |\n",
    "| 0 | T-shirt/top |\n",
    "| 1 | Trouser |\n",
    "| 2 | Pullover |\n",
    "| 3 | Dress |\n",
    "| 4 | Coat |\n",
    "| 5 | Sandal |\n",
    "| 6 | Shirt |\n",
    "| 7 | Sneaker |\n",
    "| 8 | Bag |\n",
    "| 9 | Ankle boot |"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plotting imports and setup\n",
    "%matplotlib inline\n",
    "\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Keras\n",
    "\n",
    "We will use the Keras library through out this course. It is a high-level interface to tensorflow. Quoting [the keras website](https://keras.io/):\n",
    "\n",
    "> It was developed with a focus on enabling fast experimentation. Being able to go from idea to result with the least possible delay is key to doing good research.\n",
    ">\n",
    "> Use Keras if you need a deep learning library that:\n",
    ">\n",
    ">   * Allows for easy and fast prototyping (through user friendliness, modularity, and extensibility).\n",
    ">   * Supports both convolutional networks and recurrent networks, as well as combinations of the two.\n",
    ">   * Runs seamlessly on CPU and GPU.\n",
    "\n",
    "### Note\n",
    "To use keras you will have to first install it with `pip install tensorflow keras`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Fashion MNIST is built into keras\n",
    "from keras.datasets import fashion_mnist"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "(X_train, y_train), (X_test, y_test) = fashion_mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training data shape: (60000, 28, 28)\nTraining labels shape: (60000,)\n"
     ]
    }
   ],
   "source": [
    "# How is the data stored?\n",
    "print(\"Training data shape:\", X_train.shape)\n",
    "print(\"Training labels shape:\", y_train.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are 60000 examples, each of shape `28x28`. This makes sense as we are dealing with images that are 28x28 pixels big. Let's look at a few."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([9, 0, 0, 3, 0], dtype=uint8)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# classes are already shuffled\n",
    "y_train[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n          0,   0],\n       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n          0,   0],\n       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n          0,   0],\n       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   1,\n          0,   0,  13,  73,   0,   0,   1,   4,   0,   0,   0,   0,   1,\n          1,   0],\n       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   3,\n          0,  36, 136, 127,  62,  54,   0,   0,   0,   1,   3,   4,   0,\n          0,   3],\n       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   6,\n          0, 102, 204, 176, 134, 144, 123,  23,   0,   0,   0,   0,  12,\n         10,   0],\n       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n          0, 155, 236, 207, 178, 107, 156, 161, 109,  64,  23,  77, 130,\n         72,  15],\n       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   1,   0,\n         69, 207, 223, 218, 216, 216, 163, 127, 121, 122, 146, 141,  88,\n        172,  66],\n       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   1,   1,   1,   0,\n        200, 232, 232, 233, 229, 223, 223, 215, 213, 164, 127, 123, 196,\n        229,   0],\n       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n        183, 225, 216, 223, 228, 235, 227, 224, 222, 224, 221, 223, 245,\n        173,   0],\n       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n        193, 228, 218, 213, 198, 180, 212, 210, 211, 213, 223, 220, 243,\n        202,   0],\n       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   1,   3,   0,  12,\n        219, 220, 212, 218, 192, 169, 227, 208, 218, 224, 212, 226, 197,\n        209,  52],\n       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   6,   0,  99,\n        244, 222, 220, 218, 203, 198, 221, 215, 213, 222, 220, 245, 119,\n        167,  56],\n       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   4,   0,   0,  55,\n        236, 228, 230, 228, 240, 232, 213, 218, 223, 234, 217, 217, 209,\n         92,   0],\n       [  0,   0,   1,   4,   6,   7,   2,   0,   0,   0,   0,   0, 237,\n        226, 217, 223, 222, 219, 222, 221, 216, 223, 229, 215, 218, 255,\n         77,   0],\n       [  0,   3,   0,   0,   0,   0,   0,   0,   0,  62, 145, 204, 228,\n        207, 213, 221, 218, 208, 211, 218, 224, 223, 219, 215, 224, 244,\n        159,   0],\n       [  0,   0,   0,   0,  18,  44,  82, 107, 189, 228, 220, 222, 217,\n        226, 200, 205, 211, 230, 224, 234, 176, 188, 250, 248, 233, 238,\n        215,   0],\n       [  0,  57, 187, 208, 224, 221, 224, 208, 204, 214, 208, 209, 200,\n        159, 245, 193, 206, 223, 255, 255, 221, 234, 221, 211, 220, 232,\n        246,   0],\n       [  3, 202, 228, 224, 221, 211, 211, 214, 205, 205, 205, 220, 240,\n         80, 150, 255, 229, 221, 188, 154, 191, 210, 204, 209, 222, 228,\n        225,   0],\n       [ 98, 233, 198, 210, 222, 229, 229, 234, 249, 220, 194, 215, 217,\n        241,  65,  73, 106, 117, 168, 219, 221, 215, 217, 223, 223, 224,\n        229,  29],\n       [ 75, 204, 212, 204, 193, 205, 211, 225, 216, 185, 197, 206, 198,\n        213, 240, 195, 227, 245, 239, 223, 218, 212, 209, 222, 220, 221,\n        230,  67],\n       [ 48, 203, 183, 194, 213, 197, 185, 190, 194, 192, 202, 214, 219,\n        221, 220, 236, 225, 216, 199, 206, 186, 181, 177, 172, 181, 205,\n        206, 115],\n       [  0, 122, 219, 193, 179, 171, 183, 196, 204, 210, 213, 207, 211,\n        210, 200, 196, 194, 191, 195, 191, 198, 192, 176, 156, 167, 177,\n        210,  92],\n       [  0,   0,  74, 189, 212, 191, 175, 172, 175, 181, 185, 188, 189,\n        188, 193, 198, 204, 209, 210, 210, 211, 188, 188, 194, 192, 216,\n        170,   0],\n       [  2,   0,   0,   0,  66, 200, 222, 237, 239, 242, 246, 243, 244,\n        221, 220, 193, 191, 179, 182, 182, 181, 176, 166, 168,  99,  58,\n          0,   0],\n       [  0,   0,   0,   0,   0,   0,   0,  40,  61,  44,  72,  41,  35,\n          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n          0,   0],\n       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n          0,   0],\n       [  0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n          0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n          0,   0]], dtype=uint8)"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x129d34ef0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(X_train[0], cmap='gray')\n",
    "plt.title(\"This is a %i\" % y_train[0]);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Challenge\n",
    ">\n",
    "> Make a function that plots a single example and uses a human readable label instead of an integer (replace the 9 in the previous example with \"ankleboot\"). You can find the human labels [here](https://keras.io/datasets/#fashion-mnist-database-of-fashion-articles)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# your code here"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# A first neural network\n",
    "\n",
    "Let's build a first neural network. Fit it to some toy data. In its simple form this is equivalent to performing logistic regression. Experiment with different toy datasets and adding more layers of different widths to the network. Try out different activation functions (nonlinearities)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x136955208>"
      ]
     },
     "execution_count": 185,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x154a39588>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.datasets import make_classification\n",
    "\n",
    "\n",
    "X, y = make_classification(n_features=2, n_redundant=0, n_informative=2,\n",
    "                           n_clusters_per_class=2, random_state=1)\n",
    "plt.scatter(X[:, 0], X[:, 1], marker='o', c=y,\n",
    "            s=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "\n",
    "\n",
    "def one_hot(n_classes, y):\n",
    "    return np.eye(n_classes)[y]\n",
    "\n",
    "Y_ = one_hot(2, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x159f5bf98>"
      ]
     },
     "execution_count": 190,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from keras.layers import Input, Dense, Activation\n",
    "from keras.models import Model\n",
    "\n",
    "np.random.seed(123+3)\n",
    "\n",
    "# This returns a tensor to represent the input\n",
    "inputs = Input(shape=(2,))\n",
    "\n",
    "# a layer instance is callable on a tensor, and returns a tensor\n",
    "x = Dense(10)(inputs)\n",
    "x = Activation('relu')(x)\n",
    "x = Dense(2)(inputs)\n",
    "# to find out more about activations check the keras documentation\n",
    "predictions = Activation('relu')(x)\n",
    "predictions1 = Activation('sigmoid')(x)\n",
    "predictions2 = Activation('tanh')(x)\n",
    "\n",
    "# This creates a model that includes\n",
    "# the Input layer and three Dense layers\n",
    "model = Model(inputs=inputs, outputs=predictions)\n",
    "model1 = Model(inputs=inputs, outputs=predictions1)\n",
    "model2 = Model(inputs=inputs, outputs=predictions2)\n",
    "model.compile(optimizer='sgd',\n",
    "              loss='mean_squared_error',\n",
    "              metrics=['accuracy'],\n",
    "              )\n",
    "model1.compile(optimizer='sgd',\n",
    "              loss='mean_squared_error',\n",
    "              metrics=['accuracy'],\n",
    "              )\n",
    "model2.compile(optimizer='sgd',\n",
    "              loss='mean_squared_error',\n",
    "              metrics=['accuracy'],\n",
    "              )\n",
    "# to fit the model uncomment this line, experiment with the various settings\n",
    "model.fit(X, Y_, epochs=100, verbose=False)\n",
    "model1.fit(X, Y_, epochs=100, verbose=False)\n",
    "model2.fit(X, Y_, epochs=100, verbose=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[1.8060207e-04 1.3009870e+00]\n [0.0000000e+00 1.3355684e+00]]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0.5000452  0.786001  ]\n [0.49652928 0.79176027]]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 1.80602059e-04  8.61976981e-01]\n [-1.38821015e-02  8.70603681e-01]]\n"
     ]
    }
   ],
   "source": [
    "print(model.predict(np.array([[1,.9], [1,-.9]])))\n",
    "print(model1.predict(np.array([[1,.9], [1,-.9]])))\n",
    "print(model2.predict(np.array([[1,.9], [1,-.9]])))\n",
    "\n",
    "#big change here in the prediction, why ??? "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Questions\n",
    "\n",
    "* plot the decision surface of the network\n",
    "* create a circle-in-circle dataset and try to classify it\n",
    "  * basically try to repliacte [this tensorflow playground](http://playground.tensorflow.org/#activation=tanh&batchSize=10&dataset=circle&regDataset=reg-plane&learningRate=0.03&regularizationRate=0&noise=0&networkShape=4,2&seed=0.88320&showTestData=false&discretize=false&percTrainData=50&x=true&y=true&xTimesY=false&xSquared=false&ySquared=false&cosX=false&sinX=false&cosY=false&sinY=false&collectStats=false&problem=classification&initZero=false&hideText=false) setup or something similar to it.\n",
    "  \n",
    "  \n",
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "metadata": {},
   "outputs": [],
   "source": [
    ""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(100, 100)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x136955550>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(100, 100)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWgAAAD8CAYAAABaZT40AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4xLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvAOZPmwAAIABJREFUeJzs3Xd8XFeZ8PHfc++doi7Zkrvlkrg7sRO3JE7vQCAJISGhbXYXwi6EtsACuxBCgHcLy7KwwEIWQughCYSEjVNJQqoddzvuvchVlqw+M7c87x8j2yozqqMy0vl+Pko0996554wkP3PmlOeIqmIYhmEMPtZAV8AwDMNIzQRowzCMQcoEaMMwjEHKBGjDMIxBygRowzCMQcoEaMMwjEHKBGjDMIxuEJHrRWSbiOwUkS+mOF8uIi+KyFoR2SAib28+HhKRn4vIRhHZIiJf6qwsE6ANwzC6SERs4AfA24DZwB0iMrvNZV8GHlbV84DbgR82H78ViKjqOcAC4KMiMrmj8kyANgzD6LrFwE5V3a2qCeAh4MY21yhQ2Px9EXCoxfE8EXGAHCAB1HZUmJOpWndHaWmpTp48eSCKNgwjy6xevbpSVct6c4+LRpboSdft9LotdQ2bgFiLQ/er6v0tHo8HDrR4fBBY0uY29wLPisgngDzg6ubjj5IM5oeBXOAzqlrVUX0GJEBPnjyZVatWDUTRhmFkGRHZ19t7nHRdfrVofqfXLXjhtZiqLuxlcXcAD6rqt0XkQuCXIjKXZOvbB8YBJcArIvK8qu5OdyPTxWEYhtF1FcDEFo8nNB9r6W+BhwFU9Q0gCpQC7wOeVlVXVY8BrwEdvhmYAG0YhtF1K4FpIjJFRMIkBwGfaHPNfuAqABGZRTJAH28+fmXz8TzgAmBrR4WZAG0YhtFFquoBdwPPAFtIztbYJCL3ici7mi/7LPAREVkP/Ba4U5NpQ38A5IvIJpKB/mequqGj8gakD9owDCNbqeoyYFmbY/e0+H4zsDTF8+pJTrXrMtOCNgzDGKRMgDYMwxikTIA2DMMYpEwf9BCiGoOmJ9HYMyBhJOcmiFxBcnWqYRjZxgToIUKDKvTEe8A/ATQljyVeBWcOjPgZyRlBhmFkE9PFMURozVfBP8Kp4Jw82AjuRrThpwNWL8Mwes4E6CFAgwaIvwh4Kc7GoPE3/V0lwzAyIOu6ONZcdXGf3Pes68/rk/v2h/CFcwhPVSTN2616J2h66bf9W6lmhycuHpByU9l0ctJAVwGANVt0oKsAwLrle/utrMe+P63fyhpKTAt6KPDyQdP/KjVR2o+VMQwjU0yAHhJs/KoL0SDU7owGIfzKy/u/SoZh9FrWdXEYqfknLkOcWqyiDWda0xLgH7+CoH7WwFbOMIwe6XWAFpEo8DIQab7fo6r61d7e1+guC+/ojXDicqzcvaAWQcPZEOQMdMUMw+ihTLSg48CVqlovIiHgVRF5SlWXZ+DeRnd5RQS18wa6FoZhZECvA3RzGr365oeh5q/BMUxtGIaRxTIySCgitoisA44Bz6nqihTX3CUiq0Rk1fHjxzNRrGEYxpCWkQCtqr6qzie5/cvi5v232l5zv6ouVNWFZWW92v/RMAxjWMjoNDtVPQm8CFyfyfsa2WkwLVIxjGzU6wAtImUiUtz8fQ5wDZ3ss2UYhmF0LhOzOMYCP5dkTkuL5B5d/5eB+xqGYQxrmZjFsQHI3kQWhtGPBkseDiM7mKXehmEYg5QJ0FkuvOiiga6CYRh9xARowzCMbhCR60Vkm4jsFJEvpjhfLiIvishaEdkgIm9vce5Lzc/bJiLXdVaWSZZkGIbRRc2TIX5AcrbaQWCliDyhqptbXPZlkpMl/kdEZgPLgMnN398OzAHGAc+LyHRV9dOVZ1rQhmEYXbcY2Kmqu1U1ATwE3NjmGgUKm78vAg41f38j8JCqxlV1D7Cz+X5pmQBtGIZxRumplBTNX3e1OT8eONDi8cHmYy3dC3xARA6SbD1/ohvPbcV0cRjDwmDZ7soYGJHC3K5ta/fCa5WqurCXxd0BPKiq3xaRC4Ffpkp/0RUmQBuGYXRdBTCxxeMJzcda+lua012o6hvNOfNLu/jcVkwXh2EYRtetBKaJyBQRCZMc9HuizTX7gasARGQWEAWON193u4hERGQKMA14s6PCTAvaMAyji1TVE5G7gWcAG3hAVTeJyH3AKlV9Avgs8L8i8hmSA4Z3NufN3yQiDwObAQ/4eEczOMAEaMMwjG5R1WUkB/9aHrunxfebgaVpnvtN4JtdLct0cRiGYQxSJkAbhmEMUiZAG8YwtG753oGugtEFJkAbhmEMUiZAG4ZhDFImQBuGYQxSJkAbhmEMUiZAG4ZhDFImQBuGYQxSJkAbhmEMUiZAQ9fSEBqGYfQzE6ANwzAGKROgDcMwBikToA3DMAYpE6ANwzAGKZMPeshQJGc/4tSiiVI0PnagK2QYRi+ZAD0ESOQooQm/Bqup+YCiiVLcA+8Hv2BgK2cYRo+ZLo5sZ8UJlT8ATg1iJ5JflotEjhAqf5DkjjuGYWSjXgdoEZkoIi+KyGYR2SQin8pExYyusQrXg/iItD4uoohTi+TuHZB6GYbRe5no4vCAz6rqGhEpAFaLyHPN+3IZfczKOYBYbuqT4qPOIVavyOFQRT1lo3NZuGQMjmM+OBlGNuh1gFbVw8Dh5u/rRGQLMJ7kzrVGO4o9pgorv5GgPhf/yAhAOn1W2rt5BahaiATtzu3fnsMX31tJU/0JPC/ACVk4jsW9/7aUaTNKevEaDMPoDxkdJBSRycB5wIpM3neosArrybl6NRJ2QRRU0ESIpucXENTm9+iewcnzsUtWQJsA7bnwj7dNouaEjzZ3Q7tu8pqvfO5VHvjd9eTmhnr1egzD6FsZ+6wrIvnA74FPq2ptivN3icgqEVl1/PjxTBWbPayAnGtXIjkxJOQjTpD8f06MnGtXgtW+BdwV6pbiV16JBiFUky1xDSzeeHoE8cbQ6eDcku8HvPLCgd68GsMYtkTkehHZJiI7ReSLKc5/R0TWNX9tF5GTzcfni8gbzWN1G0TkvZ2VlZEWtIiESAbnX6vqH1Jdo6r3A/cDLFy4cNhNLXDKjyKOj7R5SxQLcHyc8qN4e3s2d9mvWkrQMAW7ZAUSqiaIj2HPyjHEmvamvD4e89mzq917qGEYnRARG/gBcA1wEFgpIk+0HHNT1c+0uP4TJHsVABqBD6nqDhEZR3K87hlVPZmuvF4HaBER4KfAFlX9z97eb6iySmqRkJ/ynIR8rOI6oOeLSzQ+Du/IzacfjyzZSyRygFisfZmhsMWo0bk9LsswhrHFwE5V3Q0gIg8BN5J+zO0O4KsAqrr91EFVPSQix4AyIG2AzkQXx1Lgg8CVLZr1b8/AfYcUbYyiXuoft3oW2hTJaHkXXTY+7TkBrri2PKPlGcYQUXqqK7b5664258cDLfsHDzYfa0dEJgFTgBdSnFsMhIFdHVUmE7M4XqU30xCGCW/vWCLnb0973t2T2aXZubkhvvi1C/iXry4nCBQ3ERAKWYglfPoLCygZEc1oeYYxmElePuFFF3Xhyu9XqurCDBV7O/Coqrb6GCsiY4FfAn+lqh0OPpml3v1E42Fir80luvQtEEVsRX0BFWKvzYVEOONlnr9oNPf/6jqeW7aXvXtqGD+hgGvfMZmyUaZ7wzB6qAKY2OLxhOZjqdwOfLzlAREpBJ4E/llVl3dWmAnQ3aZYZSexy06C6+DtH43GuxZcvf1jaagqIjTtAFZRA0FNHu6OiWh93wXMkhFRbvvAzBRnFIlWYOXtBLUJ6meiibI+q4dhDBErgWkiMoVkYL4deF/bi0RkJlACvNHiWBh4DPiFqj7alcJMgO6OkEvu1auwihqS0+ICIbJwK/HVM3C3d61PV+tzSaydceoRztQKwleuxsqLoY0REpun4O6YQJ/2GomHM+HXWDkHQDxAsEtfIqg5F+/ou/q2bMPIYqrqicjdwDOADTygqptE5D5glao+0Xzp7cBDqq0mut4GXAqMFJE7m4/dqarr0pVnAnQ3RC/eiFVSh9jNP/Pm/0cWbCOoLiCoy8UaUYsmQgSVRaQNdFaAXVZNaOY+nLEnkFCyG0oKm4gs2IZddpLY6+f02euwy57DytmPWF7zEQUJsIo2YsXGEdQs6rOyja5RVep37+LY66/iN8UoOfdcRi5cjOWYf7IDTVWXAcvaHLunzeN7UzzvV8CvulOW+W13keTEcMacOBOcW7ICopesR6Iu+FYy3afrEHvlXPxjI1pd6kytILpoa3IloZMiyVHIx5l0BGvLJILqwj54JR528eoWwblF2ZaLPfI1E6AHmKqy439/TOWbywlcF1SpXLmCvY88zLx7vka4qGigq2j0E5M1p4usgkYIUv+4xALJjSN2gIS95Lzm3Dg5V65B8htPX2ePrSS6ZPPpa9oG5zOFBTjlR/rgVQB2jI5SkIqTmQUshycuzsh9OrJr7T6+97cP8NklX+eea/+DF37xGl6i/RtPtjn++mtUvrmCIJHg1FLQIB4nXnWC7ff/zwDXzuhPJkB3UdAY7XA5dspgawWEZ+09/TAyfwfidGFJt3C6+yTj/Cgd9jF7PcsJ0t9WLdvAf33oJ2x6eTv1VQ0c3VPJY996mu/+9U/xvdQLgrJFxVNPEiTi7U/4PjVbtuDWmlWgw4UJ0F2k9bkE1QVo0D64pcp3ASC2Yo+uPv3YKqnrYmHgHyrtSTW7wMGvmY8G7Xu3NAjhVXVlrujAcuMev/yn35OItU6zmoi57HurgrXPbhqgmmVG4mR12nOW45CoSbvwzBhiTIDuhqaX56NNEdS1AdAA1Legg0axxs9kjFO3i13+Ahrv5vCAFQNJ0epKwT92LRobhwZhVEFV0CBEUDeT4GTfd0301o43d6f9EBBvTPDqw2/2b4W6aM2Wrn0qyhk7Lu25wPeIjOyrN29jsDGDhN2gjVEa/ngJzqQj2KOr0HgIb8/YZF9zXvvgqK7davqdu2s84Zn7EbuTbo5AsMdWEVR3PhgkubtxRj2FRCqTZcbG4h19Oxqb0MELCePu/xskZx9W3nbAJqibnTUbzSbiaTYoaBZvTPRTTfrGxHfeyJY9u5N90C1IKETp4gtwcs1Co+HCtKC7K7Dw9owjvnwuibUzCE4WJlvWrp1cGdhMXRvv8Ei8faNPH0tsPIugLgd1O/mxqyRng3QictlYQhN+gxU9hkiASICVU0Go/EEkcriTZwvaNBm/8lr8yquyJjgDTJ1fjpdI3c8cijqce8Wsfq5RZpWcO4/ym29BQiGsSARxHKxwmMJpMzj7zr8Z6OoZ/ci0oDMgqCym4U9Lk/OaxyRb1u72iXj7R9Pqs7jr0LjsQkJTKwjNOIBVVN8u/egp3oFRnZbrjHoq9XZX4mKXPYd38EM9e0GDXGFpARe+ewEr/rimVT+0WEI4J8wltw/+bprOTHjHOxl1yaVUrV6Nn4hTNGMW+ZMnD3S1jH5mAnSGaEMOidUz6fTDtW/j7ijH3VFOdOl6nInHTi9UgWTLO7G1HG3M6fg+jodEUm98IAJW3p7uvYAs8757byS3IMqLv3ody7LwXJ/yOeP462/dRn5J3kBXLyPChUWMueLKga6GMYBMgB5AsdfOJXT2QUJz9mLlxAnqc0hsPKtVt0ha2sly7M7OZznLtnj3P76NGz5xFccPVJFXlEvx6L5Y2GMYA8cE6AEluDsn4u6c2Pmlbfk22jQeyW2/dZWqENRPz0D9Br9wTpjx08cMdDUMo0+YQcIs5h19e/NehGeOqQoEYfzj1wxcxQaZTScnDXQVDKNHTAs6i2l8HO6+D2OXPY+VuxtItpz941ej7siBrl6P1Vc18MT3nuPNJ9bhxj0mnzuBm/7hOqYtmjLQVTOMfmUCdJbT+Bi8gx8Y6GpkTGNtE9+8+b+pOV6H7yan0u1ctZfv/s0DfOS/7mDeVbMHuIaG0X9MF4cxqLzwi9eoO1F/Ojif4sZcfvWVxwiCLuQyMYwhwrSgeyyZLhTPRnLjhGfvwRl/HPVt3F3jk0n3PfPj7a4Vjye7NVKJN8Q5tP0oE2Z2b1FNrLae9Q89zdanX0N9n7MuX8R5H3gH+WUjOn+yYQwgE0G6ywoIn7uT8IwDzQHagual26dyRVvzdxCadoDGpy6Erubf6IwE2GOqkEiCoKqQoDY7ss51V0eZ6ESk25nq6k7U85v3fYum6lr8RHJRy/qHn2Hzn/7Ce3/+DYonmhkgxuBluji6KXrZWsKz9iVzOluKhH2wtFUif3ECrLwY4dmZWSxijz5B3q0vkXPpOqJLNpP7jjfIuXplMkHSEDP/6tnYjp3ynFjChBndaz3/6XvP03ji5OngDBB4Pon6Rl7695/1qq6G0ddMgO4Ga2QNzuiqdjmdU+WCFicgdHa6zX67TvKayLliLVbERcJ+8ssJsEdV44x7uNf3H2yu+dtLieSGkTY/1HA0xM2fvQ47lDp4p7PiibUEKVrdqsrBVZtwm4bem1xn1i3fO9BVMLrIBOhucMYfP92d0RXi9D5xfGjmvpQbBYitWLn7IHSi12UMJiVjivji7z/OjAumYjs2TtimeHQh77vvJi57/4Xdvl+6/mwARPA6yYxnGAPJ9EF3Rzc2OdEA/OO93zvOLjuZeh9EALWxIkcJBmDOczzuc+J4E4XFYfLzwxm99+jJpXzmFx8h1hDHjbnkj8hr16LuqsnnTGDXmn0pz+WOKCJa1P2+/OPb9nJ44w4iBblMufh8wnmd5E0xjB4yAbobvIOjCM/d0+HWV6cFFvH1Z/e6TG2MoJpmSy0U9fs3MZDrBjz44408++RexALfU85bOIq7P3c+xSXRjJYVzYsQzYv06h43/cN1fPfDD+LFWqexcqJhLvr4e7sV+OP1jfzpH77FsS27UU3mA9FAuerLH2HGdUt7Vc9M8xoa2P/4Hzj2ysv48QT5kyYz6dbbKJ49Z6CrZnSD6eLohqC6EK+iDPVa/9jUBw0ETTioaxM0hmn6y3yCE8W9LtPdXg5emn7XIII29SCPRy98+5sreebJPcTjPrEmH9cNWP3mUT5/90u4aXI0D6TpS6Zy3Tc+Qe6IIkK5UUJ5OYTzc7nkMx9k5tsu6da9nrnnBxx9aydeLIEfT+A2xvBicf78jfs5tm3wZA/0YzHW3fsVDj//HF5DA+q51O3aweb//BbHVywf6OplPRG5XkS2ichOEflimmtuE5HNIrJJRH7T5lyhiBwUke93VpZpQXdT7JV5hGbtITx7H1ZOgqApTGLLJNxt5VjF9eBbBNUFdLgxazf4R0bg7hpH6KxDYPuIRfINQgW34nb68z224kAdq5YfJpFo/QnC95Wak3Fe/UsFV1xTnubZA+esyxYy9ZLzqdx5APV9Rk4rx3Y6/9Ov3neYnS+swIslGHnWBA6s2Ijvtu/T9hMua37xJ67/5if7ovrdduSlF0hUV6Fe67oGiQS7HvwpkQ9+EbG6N9hqJImIDfwAuAY4CKwUkSdUdXOLa6YBXwKWqmq1iLRN7v514OWulGcCdHep4G6eirt5KslO6TOBOKjsfYu5PSG+chbevjGEph1AchP4R4txd5QTmtvBtlZ9YOO642m7BGJNPm++fnhQBmgAsSzKpnctaZKq8sp//YqNv38O9QMCz8cOh1LOBgHQQDm+LXU/90A49urL7bbLOkX9gOBYBfaYwfl7ygKLgZ2quhtARB4CbgQ2t7jmI8APVLUaQFWPnTohIguA0cDTwMLOCjMBulf6K+ey4B8bgX9sYFe+hcJ22h1gAKLRodEq2/H8ct76w5/xW8zwaDmPOpX80YMnOVW6NxIg+ScbDL6uqCwyHmiZ4/cgsKTNNdMBROQ1wAbuVdWnRcQCvg18ALi6K4VlJECLyAPADcAxVZ2biXv2l7OuP69vbmz7OJOOYOXFCOpy8faPgqAbAczyccZXItEEfnUBQWUR/feGkNqiC8bww++knlESjdpcPkhbz921+hd/wot1bYd0ACcnwvw7ru/DGnVP6cJFHFx2tF0XB4AGAdao/v3kNRi44TwOT+zSVmilIrKqxeP7VfX+bhbnANOAy4EJwMsicg7JwLxMVQ92dXA6Uy3oB4HvA7/I0P2ymj2qmpwrVifjaXO+DhZvpvH5RQRVne/6YY+tJOfSdckHloIKQV0uTX9egMZ6N6uhNwqLInzgr2fzm59vIR470wqLRG3OPW8U555XNmB16614fSNbl73CkY07qdp9sMNrnUgYL55ALMEOh5h9w2VMXtpHb/Q9MPba6zj8wp/xGuppmSzcCocpv+ndVDqhAazdoFepqh11PVQALUfmJzQfa+kgsEJVXWCPiGwnGbAvBC4RkY8B+UBYROpVNeVAI2QoQKvqyyIyORP3ynohl5wrVyOhFh8jm7/PvXoV9Y9eDkH6fgLJayLn8rXtVitaxfXkXLEmmd9jAN383ulMKC/goV9upeJAHUXFEW58z9lcd8PUHs9VHmiVO/bx+49+Hd/z8Jo6bjlbIYcL/v5WDq/fQU5JAbPfdTlj5vR+OmUmhQuLmP+1r7PzZw9Qs2UzWIKTk0v5u9/D2CuvotKsJOyNlcA0EZlCMjDfDryvzTV/BO4AfiYipSS7PHar6vtPXSAidwILOwrO0I990CJyF3AXQHn50PgonEpoyuH0J60Ap/wo3t4z+SQkJ0Zo+n7sUSfRpuScZyTFykFLsYrrsUpqCaoHdu+9RReOZdGF3cuJMVhpEPD4p/+deF1Dp9faIYeply3k/PffAO/v9PIBFS0bxdx//CJ+LIYfjxMqKEAsM6u2t1TVE5G7gWdI9i8/oKqbROQ+YJWqPtF87loR2Qz4wOdVtUdLfvstQDf349wPsHDhwm6sycsekhPDHn+sdeu5JcfHKjgTCKzSk+RevQpEESdAm+Ny2oE4Fayi+gEP0ENJxdqtJOob018gAqqEciIUTxzDeR94B43VteSWZMfvwI5GsaOZXUA03KnqMmBZm2P3tPhegX9o/kp3jwdJdg13yMziyAQrIHrhWzjlRwHSr/zzbIKGU8uCNdmV0SKYnwrMaZ+voE0D1wc9FNUf7bhhUzi2lGlXX0DNwWPsfX0tj33smwSez9hzp3PNV/+egjGDZ/aGMfSYzzwZEFm0Baf8KOIEya90XbEK3r5k/mF7VDVid2+6k3oO/lGTZD6TRkwZj2rqD3SW43DW5YtoqDzJ3tfX4sWSqwf9hEvF2i387q+/TKKhqZ9rbAwnGQnQIvJb4A1gRvMSxr/NxH2zQsglNPVQu0E9ODOArq6NujZNLy4APznVTqKpFxJA86fqgNNdHupaaMKh6cXzGOipdkNN2cwpFE8Yjdjt/ylYjsVZVyxix/PL2+XyUD8gUd/ElmWv9FdVjWEoU7M47sjEfbKRVdTQPCsjxcCegCZs4mun4+4ZC+6Z6U1+dUFyCl0K6kvyes9BcmP4x4twd06ARGazxhnJXVre9d0v8Ie//wYNx6rxPQ/LtlHfZ86NV7DpiZfSLvzwYnH2vLyaebde28+1NoYL0wfdSxoPdZjdLqjLSyY8avu8ujz8Y8XYo08ibXNMBxaJ9WejjSaNZX/ILxvBBx/5NgdWvsUbP/wdx7ftRWybDY88iwYdj2eHcswAnNF3TB90L2ldHkFt7unuiFbnXJvE1vRTCpteno9/rBj1LIKEjSZsgliIphfON8G5n4kIhzds58SuAwSejx9PdCk4z37nZf1UQ2M4Mi3oDIi9Oo/c61agdnC6L1pdG+9oCd6ecemf6IZoen4RVmE91og6NB7CPzIC1Lxv9qWTB46w8md/ZP8bG7DDDrPeeTnzbr2Gtb9e1q6vOR07EmLsvOlMXjq/j2trDGcmQGdAUJNPw+OXEJp2AHt8JZpw8HZOwDswiq4M6gW1+UN2l+7B5ti2Pfz+rvvw4gnUT76ZrvrZH9n8xEt48a4FZ0SYd+t1XPTx283iD6NPmQCdIRoPk3jrLHjrrIGuitGB5+/7MW5j641i/YRLY2U1gd+1aY922GHxh9+NlWb3ccPIlGH89q/kz0oQnrM7mWc50sXWk5G16o9XUb33UMpzvuvhhEJYnSTyd6Jhznn31WYfQqNfDMsWtJ0fcPY/VxMuC7AiJ8G3iCzaSuzNmXg7+3cLKaP/eE1xLNsiXTvZDjsUjCim/tgJ3BZJk8SyCOVE8F2PmW+/lIs/mToRx4mdB1j188c5vHEHOUX5nHvbdcy4/mKsFHOsDaMrhmWAnvypGiJjfaxT05Kbp8lFF22lsbqQ4ETvd+M2Bp/CcaOwQiFIkbFORJiwYA7X/79PsvOFFex84U1sx2batRcRyc/Fi8UZPecscopT5+DY+/o6ln3hO/gJDw0CaiuO8eK/PcCO55Zzw7c/Z4K00SPDLkCHy3zyznbPBOeWrIDw7D3EXjEj80OR5dhc8He38tr3ftMuIb8dCbHkrluwQw4zrlvarV26A8/n2Xt+0G4GiNcUp2LNZnb/ZSVnX9l20w3D6NywC9CRMR6BK1jh9nNcxUrmXTaGrnm3Xov6Act//AgaBGgQkDeymKvv+TtKp3Vtz8K2Dq3fhp9mtaHbFGfDI8/TVFPPhj+8BkDZBRcxaunF2BGT+MroWFYFaA3qcYp9vJrkrtY9kai0sZw0S6wVgrrc3lTRyALzb7+ec265muq9FdjhMMXlY3q12UCivrHDyZSHN2zjyFs7Trfa6/fspuKp/2P+vV/HyTPTK430sqJjTL2DBFV/jR5bwuzvnGDu/1RSem0jyV21uyd+2CF2yEm9b6Zn4W7pWSvKyC52yKF02iRKJo3t9U4wo+eche+23/8PAEsIPK9Vl0oQjxM/cYI9D/22V+UaQ9+gD9AaVKEnboHEG4CLFQanQBn73npG39xBovUO7P2vIrwaC78p+Q9TfUE9i8TmKfhHTX7f3uri5pxDRl5pCdOuvgAnkiKZVaApl4yr53Hs9VfTpjo1DMiCLg5t+DVoI22zxdlRGP3OBo4vyyGId/w+Y+UEFMxxwVLqt4RJVNps+cxIihfHGXtNIRoL4+4Zh9bmpapBcpZHYGFSfRrpXPXlu7DDIbYuewU7HCLwfPJHj+DkvvRboKnnob6PdDL32hjSk1fNAAAgAElEQVS+Bv9fRvx5IPVGnuoLuWd71G9Kn4az7PpGxt5Wj/rJ4CqOUvlcDod+k0/161FGFM5J80wlNP0A4XN2I9E4BBbuzvHE104Hb/D/2Iz+ZYccrvrnj7D0E3dQtfsg0aICRkwZz4M3fpLaQ8dTPicycmSnC2OM4W3Qd3F09h6iHazOLVoYY8yt9VgRsHMVO1exwjDyqibK3t5x90j4/O1Ezt+GlRtHLBAnIDTtILnXrUi5qasx/MRq6tn7+joq1m49nTM6WpjPuPkzGTFlPABLPnorTrT9bA0rHKH83e/p1/oa2Wfwv33nvBvqdgCxdqdUoWFHqgnNSWNuacBOka7XjsLodzVy/KnUMzYkGic8c3+7PM1iK1ZBU3Jn7n1DY1dro/sCP+CV7/yCtx57ATvkoIBt21z91b9j6qULWl076+2X0HCsiuX/+wfETubuUM9j4o03MfriSweg9kY2GfQBWnLfjTb+Gvz9wJmFAEEcDj6YD376fuHIuPTNayui2HmpB2jssSeS902RC0dCPs7kIwMeoMOLLhrQ8rOJl3CpWLMFP55g7LzpaVcDdtUbP3qYTY+/hJ9w8RMuAC7w9D9/j/fc/1VGzZra6vqFd95IcO411GzdDEDRzFk4Of0/ndOtraVy5QoSWyuwRk/AHn9Wr2ewGH1r8AdoyYGRD6MNP4LGRwliVTTuDXHk0TzqN3e8BVQQl7RznpHk+VaH8pqw8pqQcKLj8cBBPvC+f08tf/jddrZtqaJkRJQbbj6LCy8ZNyz/Mb75p3X84itfQ0j+2gLXY85NV3LZZz/Uo1ShbizO+oeebrcSEcCLu7z508e44T8+2+6ck5PDyPMWtDveX468+Gd2/eoXiAiB64ITQvKLyb3pw0iOmYvdHSJyPfBdkk24n6jqv7Y5fyfwLaCi+dD3VfUnzefKgZ8AE0n+Sb5dVfemK2vQB2gAsfKRgs9BwedYd9XFXXpO7lQXK6Ko0m6X7cCDmtUR1G0eOIzGiV6yHru0Jjlbw/LBTrOYxbXx9g7e7o2Vbxzm3+97E9cNCAKl4kA9O7dVs/zVcXzmSwuHVZDesXIPv/yn3+PG3FbHNz/xIpGCXC78u9u6fc+aA0dSbjALgCqHN+7oSVX7VN3uXez+9S9R1z3TtnATaE0lTc/8ltybPjKQ1csqImIDPwCuAQ4CK0XkCVXd3ObS36nq3Slu8Qvgm6r6nIjkk2oz0xayYJCwZyZ8uA7LaR+cVUETwsEHC5IHRMm57k3sUScRJ0DCHuIo6JlduU8/17MI6nKbE/EPPq4b8O3/t5J43CdoMfc2FvNZ/uoh1q0+NoC163//99/Pk2gTnAG8WIJ1v32q6wn6W4gU5BOkW5QCRAoG30rUimVPJlvNbQUBwdEDBDVV/V+p7LUY2Kmqu1U1ATwE3NiVJ4rIbMBR1ecAVLVeVTucrTAkA7RTGBAdl/ofkQiop/h1yZdujz+OlRNH2uywLRbgC0EslAzqrk1i+0Qan1k8aLekWr/6WLs3lVNiMZ9nn9zbr/UZaPs3p879DIBC3ZET3b5nwZiRjJg6IWUXmBMJM/ac6Wx/9g0aq2u7fe/uCjyPk29tpHLlm8Sr0r+WxoqD7Vsbp9gOQU1lH9VwSBoPHGjx+GDzsbZuEZENIvKoiJzKYTwdOCkifxCRtSLyreYWeVpZ0cXRXRLSDj84iHPmX5ddVo2E0gwm2oq7aSKJjWc15/4Y3N0DDQ1uh/3jtTWp55MPVbmFOTTWNKU8F3gekcJUC5M6d+3XPsYjH/4qfjyBn0g2BCzHxku47HhhOTtfeJPA85j33utY+on39Um3UvWG9Wz7n+8nt+2S5OsZef4Cpt/198mUqi1ER41KBulUAh+roDjj9RtsYn6YTSe7lMahVERWtXh8v6re383i/gT8VlXjIvJR4OfAlSTj7SXAecB+4HfAncBP091ocDYFe8mtsvAaUr809aHurRZ/wK5zehFLO4Ggbqi5xTy4gzPA9Jkl+H7qd6ZwxGbegsHZNdNXLrtjScrl12IJo+eeTW5Jz2ZzjJw6gQ8+/G3Oe/8NlE2fTPGksc0fzRSvMY7b2ISfcNnwyHNsePiZ3r6MdhoOHmTL976D19CAH2vCb2pCXZeqNWvY+bOftLt+/NvegRVOMaAuglVchlUyvP4uOlGpqgtbfLUNzhUkB/hOmcCZwUAAVPWEqp5qDf0EODU6fBBY19w94gF/BM7vqDJDMkCjwqHf5BGkaDAGLhz5/ZlRa3fv2PSZ8VTw9o3uo0pm3tjx+cxfMIpQuP2v1XEsrnvHlAGo1cC54kNLKZsxGSfnzEIROxIiUpjPNff8Xa/unVdazEUfey93/PpfUD9I2S/txeKsfOCPGc+3UfHknwi89uUFboLjK5bj1rXuXimaOYsJN7wr2bJunotNKIzkFhB9W+rdYYy0VgLTRGSKiISB24EnWl4gIi1nEbwL2NLiucUiUtb8+Eqg7eBiK0OyiwPg5BvJPePGva8BJy8AC2KHHA7+tIDYwTMvW+tzSWycSnjuHnD8Uw0h8C3i689CG7Nr77nPf2Ux3/331ax47TChsIXvK6WlOXzh3iUUFQ+v/MOhiMO7f/wVdv55BZsffwk3FmfqpQuYe/NVRIsyM7VMVampOJr2fKymvnmT2hQrpnqoducOCFJ/UrIch8aKCopmtv50UH7Tuym74EKOvvoKR3Yewh4/BWfqHMQesiGgT6iqJyJ3A8+QnGb3gKpuEpH7gFWq+gTwSRF5F+ABVSS7MVBVX0Q+B/xZkv1eq4H/7ai8If3bOflGDieXRwmNDFAXvJrU/fGJt87CrywmPHsPVmEjfk0e7ubJWZnZLhJ1+Md7llBdFWP/3lqKiiNMmlI4rKbXtWQ73d8hpTtEhHBuDomG1H3dYlups9z1QqigkNjRIynPqe/j5BekPJczZiyT33MbJ5fvzWh9hhtVXQYsa3Psnhbffwn4UprnPgec29WyhmSALpgXZ/SNjUTHebgnLY4vy6XqlY5bMP6RkTQdyb6AnE7JiCglIzLXajPSm3PjFWx49NnTA4anWI7D9OsuwnJsMrm6ady117LjwD6CePs+vPCIkeSOTzWpwMhGQyJAh0d5lF7bRO5kDwkF5JT7WM2NFqfAZ/xf1ZE/N8H+HxaSDYN9RnZZ8tFbObBqEzUHjpzeDdzJiZBfNoJLPv2BjJdXuvgCKt9cQfXGDaeDtBUKIY7DzLs/OWw/LQ1FWR+gC8+PM/nuGrDACpFy5aAdhaIFcXLP9mjcmT65kmH0RDg3ynsf/AY7X1jBtqdeQ4OA6ddeyLRrLsx49waAWBYz7/4UVevXcvTFF3Dr6ymeew5jr7qGcJHZkX4oyUiA7mxtel+xogGTPl6D1WLsK13jwQpDycVNJkAbfaInu4H3hlgWI89b0O38HutM/3NW6fU0uxZr098GzAbuaF7S2OeKFsa73LUnFtjDaxKDYRhZLhPzoHu8Nr23nEJFuvgZwG+C2vWZ/7hpGIbRVzIRoLu0Nl1E7hKRVSKy6vjx1FsAdVfTPgf1Oh8QUR+8WoualaYJbRhG9ui3QcLmJZP3AyxcuDAjc47qN4dwqy2siJ9MbpSyXGjc47Dn28Xpl3QDVulJwjP2IXkx/BNFuNvK0frBl5nMMIzhIxMButO16X1GhZ3fLObsr1QTGR2kHCD0G2DHfSWnd14Jl/mMvLKJyFiP2AEHCWLJzWFn7wUrSPZVl9YQnnaApr+ch3+4tF9eimEYRluZCNCn16aTDMy3A+/LwH27xDtps/2fSpj7oxMQaj2LQxUad4VOB+fii5oo/3Dd6Sl5wbwEYr0CokiLBP1iK9hKzqXrqH/kymQSf8MwjH7W6wCdbm16r2vWDSUXx9EArDYtaBHIn+USKvFRhfKP1J1ewAI0fx+kTZUL4Iw/jncgexImGYYxdGSkDzrV2vT+VLQgnnL3bgD1JRmkR6bfQDbtwisBiXR/1w3DMIxMGBKf3YNEBzM5NJliNDTSb9V6bnVJBy1ov6p3O0AbhmH01JAI0FWv5ODHUp8TW6nbEKZpn5P2mlTUF4KaPIIqs3TWMHoqqK/BP7yPykNm38OeyPpcHAC1a8I07gyRO81ttVrQj0PFr/IJ4hYnX48y7o4G2i491AA0EUIDSW59pYCl+FWFxF46r19fh2EMFUFjPfHnf4d/eB/YDn/11APMu3wuX/zVJygckTodqtHekAjQqLD734sZeU0jZdc24RQoTQccjj6WS93GZMQO4ha7vlnM1C+cxAoBtmJZNhoP0fj8QrQuF3tUNZKTIDiZT1CTmYTuhjHcaBDQ9Mf70ZqqZAvI90gAa1/YwOeuvJcfr/0Pk3Gvi4ZGgCbZJVH5dB6VT6ffCLRpX4hNd5dScE6CcKnPyIkz8Y+M4FQKUv/YiH6qrWEMXf7+bWhDbTI4t+AlfI7sPsbaP2/k/Ku7nLN+WBsSfdDdEgh16yOc+HMu/pGRmPzQhpFZ/qE94Kae/dRUH2Pjq1tSnjPaG34B2hhyvISX8Y1ZjV4IR8FKHVpCEYfcApNCoauGTBeHMbyoKi//dgXLfvgCNcfqCEUdLrplITd/9nqi+SYp1kAKTZuHu/oloP3GtiLCZbdd2O91ylamBW1kpcf+42ke/dcnOXm0FlUl0eTy6sMr+dbt/4PXZm/AbKZBQKKmBj/ejTmiA8wqGklo/iXgtN4cI5oX4QNfeQ+jJpr8Nl1lWtBG1qk9Uc+ff/4aXrx1IPYSHscPVLH2uU0sese8AapdZqgqh555igNPPJ4MzqoUzz2Hs+/8WyIjBv9gdmTJNdhjynHXvUJQU8V5F0zkts/fyIJrsvv30t+yPkA7hQEF5yY3zqzbEMGr7eGHAitIfnlZ/yMZ8ra8ugPbsfDab2pNvDHByv9bn/UBev8fHqHi6ada7dxdvWE96776ZRb867dw8tLPVhosnEkzcCbNAODfvj9tgGuTnbI4Giljb6+n7Lqm03mexa7j+LM5HP5tPl2dnSGFDUQXbcEek1zpFNTlEl89A7+irK8qbmTCEB4T9BobqVj2JIHrtj4RBHh1taz/xtcoW3wBY664gnBxycBUchjr6h6sInIL8CiwSFVXiUgI+AlwPsnY+wtV/ZeOysraPuiRVzZRek0TVhjsHMXOUawwlF7dxMirmrp0D8lrIu9ty7HHnEAsRSzFLmog54o15N/xLJFFmyHsdn4jo1/Nungavtd+AAogkhtm0Q3Z3Xqu3b4NcVK3nTQIaKo4yIH/e5xVn/sHTr61sZ9rN7x1dQ9WESkAPgWsaHH4ViCiqucAC4CPisjkjsrL2gA9+qbGlBns7CiMubmxS/cIn7MLHK/dbiwiII4SmnaA3LctB2foDDoNBYUj87nqzqWEc1oPQjlhh7LykZx3zZwBqllmiG13eo26LkEizubvfQc/nqKvx+grXd2D9evAvwEtR3cVyBMRB8gBEkBtR4VlZxeHpYRGpG5BAThFAWJrh1tcATgTj6XdKgtAbLByY4SmVuBun9TT2hp94ObPXU/pxBE8+YMXqDlaSyjqsPQ9C7nps9fjhLPzz/qUohkz0SD933dbJ1avYtRFS/uwRtmvIQZrtnSpX6xURFa1eHx/83Z9p6Tag3VJyxuIyPnARFV9UkQ+3+LUoySD+WEgF/iMqnaYRSo7/5IDCOKCHU39Aw/c5EaxHQmP9pBI590X4gQ4Uw+bAD3IiAiX3r6ES29fgu/6WI41JPI7NFZUcOjZpwkVFhE/UQmdBGp1XRI1J7t0b7euFr/qKFZBMRIyc8XTqFTVhT19sohYwH8Cd6Y4vRjwgXFACfCKiDyvqrvT3S87AzTCiRejlF7V1C7Hc5CAqr9E6WyQcMpnarpemrR9I1CswoZkebV5nZZl9C071HmXQDY4vvwNdvzkxwSedyYwn3rTSbNSUkIh8sZP6PC+iZMn2X7/j6jZugUVCwIfZ8b5RC65AbGzNAQMnM72YC0A5gIvNTcYxgBPiMi7SG4F+LSqusAxEXkNWAgMtQANRx7JI3+mS2SMh52TPOY3QfyozeGHOs5EFxnvER7lp99JpQX1LNy9Y08/diYcJbJkCxJK9kuraxN/cxbegTE9fi2G4TU0JINzok0OC1UkFEIsq9WUOwBECOXlUzz3nLT3DRIJ1t93D/GqagjOfKz0tq1BG2vJefuHMvkyhoMO92BV1Rrg9EocEXkJ+FzzLI6rgCuBX4pIHnAB8F8dFZa1ATqIW2z/aglF58cpuTAOolS/EaVmdeT0JrHphIoD1BOIdNwnpYGg8RDuzvEA2GMriV6yAXHOfOyUkE/04o00/cXGP2Sm5hk9U7nqzfR7rwUBpRdeRNWaNajnokGAWBahwkLm/uOXkDR5LwAqV67AratrFZwB8D38AzsJqo5hjRiVwVcytKXbg1VE7gNWqeoTHTz9B8DPRGQTyY/dP1PVDR2Vl7UBGgBfqFkZpWZlmg0J04gftrFCqYPz6U+SgeDtH0181Uxwk7MFIudvbxWcTxEnILJoK42PmwBt9IxXV9d+3nMz9X3Esljy3z+keuMG4lUnyBk7jqKZszrtd69at7Z9y/s0wT+0xwTobkq1B6uq3pPm2stbfF9Pcqpdl2V3gO4ht8qm7q0whee6iNMmUPsWjc8tJDhRBNqiZSKKVVKX9p5WQSPRpeuIvTaP4dwnfXji4oGuQlby4/G0/cxWJELh2dMQ22bE/O7t8mNHO2i8iLTLl2EMLlk7D7q39v2wEL+yGPWs5FfCRj2L2OtzCSpLWgdnSM5gDNIHXpHktL3Q9P19W3FjyNn/+GMcXPZk6gAtghUOU7qkZxngRi29BCuSZsaGBjiTZ/bovkb/GJYtaICgyaLpuQVYJbXYpTVowsGrKOsgF4fgHRiFU3407dxpCSnhOXvNlDyjy9zaWg48/kfUS929ESoq4pwvfRk7XZDtROGMmZTMm0/1+nWtuzqcEOEl1yJRk5t5MBu2AfqUoLqQoLqwS9fGV8/AHnsCwl7a8RzJNau6jK6rWr8OsS00zWLVollzyB07rsf3FxFmfuwTHH3lZQ49vYzGE1VYJWWEF1yBUz69x/c1+sewD9DdoY05uNvHE567r4NrzAIAI8n3PE7s2M/WfULexPKUsy3U99P2PTffpNf1EMtizGWXM+ayy1m3fG+v72f0HxOgu0MCwjMq0raeVSHx1pQ+r0Z40UV9XobRsaObdrH8/kc5vGE7oZwIs955OePOnc6JXfsJ5+fixRK8+ZPfEwSK7yt2KMxZd/4NpYtaD6IWz56ddrsuKxJl5IJF/fFyjEHKBOhukGgimTM6HQV3x8T0540hYe/r61j2j9/BSyRAIVHfyKoHHgMRxJLkohK3dcs3iMXY/uMf4uTktFpYEh01mtJFSzix6s1Wi1TEtgmXFDNyoQnQw5kJ0N2grgPtln23OB8L036KneJMrSByzh4kvwmNhUlsLcfdPLn9TBFj0DnVuj0131iDgOe//mO8eIpdq1VRX1E/9Zt4kEiw95HfMb/Nyr/pH/ko+0aO5PBzz6BBgAbKyAULOOuv/gYrZKbBDWcmQHeH5+BVlOJMqESs1oFaPQt3e3mrYxKNEzp3B+Gph5DmhTGSGydy7i6c0dU0vXA+w3nO9GB2ZNNOXv3urzm8fhuIMOmCeVz86ffjxRO4jT3fH7B+7952x8S2mXzreym/+Rbc2lqcvFzsSPcWXxlDU68CtIjcCtwLzAIWq+qqjp+R/eIr5mCPWA4RFwkll8+qaxNU55PYNBkAq/Qk0SWbsYrqwdJ2fdbiBNijqrFHVeMfG/z7yw03h9dv57G7v4kXO9VKVva+vo5D67Zy5T99uFdZ86xw+hax5ThZsd+g0X9624J+C3g38OMM1CUraCxCw5+WEppyGKf8KOpbeLvH4R0YBWphFdeRe/Wq08E7LdvHmXzYBOgBEPhBsq84TaB96Vs/axGcm6niNsbY/uwb3crV3IptU3ZBzxacGMNTrwK0qm4BhkQe3m7xHNwdE1MOCIbn7QS7k+AMycUu1hDeWG8Q2HSy9YKhfW+s57X//g2VO/djOQ5nX7GYiz/1fkI5EeqPV5NfVsLuV9ZwfNvelPdTVXb/ZRViWYhldS9QWxbhwkIm3freXrwiY7jptz5oEbkLuAugvLy8k6uzlzPmRIe7tJyirp1sdRv9Ysfzy3nu3v85PbgXuB47nl/OrpdWoqrYIQcvnkCDzt80e9KCzhkzlnO//FVC+R2nwjWMljoN0CLyPMmk0239s6o+3tWCmreNuR9g4cKFQ7bpqIGF0HELWn0hqM016Un7SeAHyW6LNjMvNAjwE8lg23ZaXCZJKMyopReb4Gx0W6cBWlWv7o+KZBurpJbQjP1Y+U34Jwpxt5ejDTl4e8cQmnYQsdu/B6kCvoW7Z2wyjakOs66hAVK152D7PuV+JJYw+rLLB6x8I3uZaXY9EJq1l8j8HWAFiAX2qGrCM/bT9PJ8EhvPwpl0BMIu0rwTkwaALzT95Tz8oyMhMPOf+1VHS6n7kJMTwbJtZnzys4QLiwakDkZ261WkEJGbReQgcCHwpIg8k5lqDV5WYT2R+TsQJzjd1yy2IqGAnEvXo55NcKKQdvOb1UIbc0xwHgAjpk7Ecvpv30KxLUbNnsq1936MDz/zI4pmzuq3so2hpbezOB4DHstQXbKCc3ZF+tkXCpHzdmCPrm7VxSEWqOMTvWQ9jf+3tJ9qapzy0rceIF7X0C9l2ZEwOUX5vOs/P0+kKJ99r6/n8IYT5I6fQOGMmQM648kkSso+poujm6y8WLtVhGdOKs6EY0goxbZYVnLXFSloQOvy+riWw9eedfv5/beeYtfqfWBZ2OEQbkNTn5XnRMKUzZqK1xRDVZl+7UWcc8vVVO0+yBO3fZ7A8/C9AMQiXFzM3C98iWipGRw2usYE6G7yjxfhjE8dhFHpOJlSYCHRhAnQfWTbil18/8MPkog1J7/3gy7PzggX5BG4Hr7rts6lIc3/adOPbYdDIDD3lqu5+JPvx7LPdF01nazjj5/4l3ZLwmPHjvLWv/4/Fvz7tzvc6NUwTjEBupvcXeOJzNsFtA7EGkDQECWoLkAmHUk9F9oKCGpMcO4rv/nq42eCczeIbfHXj38XsSz+/I372f3yauyQg+96TFxyDjOvX8pbf3yB2opjFJeP5eyrljBi8jhGnl1OJL/9jiSb//SX1HOlVUnU1FCzZTPFc+b25CUaw4wJ0N3lhmh8bhE5V6xBHO/0WGBQl0vTCwuQaBxn4rF2LWn1ktPrSIQHoNJDX83xOioPVPXoueoH/P7vvs6N3/0Cb/uXTxGrqafuSCV5o0aQW5LcbWf6tV3PwX1sy+600/rU92msqDAB2ugS8zmrB4KqQhp+fxlNfzmP2IrZND67iMYnL0KbIgTVhTS9Mg9NOMmNaBN2Ml/HgVHE35wNKPbYSsLnbyU8bztWSe1Av5whIfCDXiUGPLHrAI997JuoKtGifMpmTD4dnLurcFxZ2lkj4tiEi4t7XlFjwInI9SKyTUR2isgXO7juFhFREVnY4tiXmp+3TUSu66ws04LuMcE/MjLlGf/gKOofuQJ7dBUS8vAri5JT7ByP3OtWYhU1gOODQnj2Prx9Y4i9PheTerT7fM/nuZ++wgs/fw0v3vPVgOoH1B2p5NC6bYw/r3c7Xc+58QrWPfQUeO1XlIoII+af16v7GwNHRGzgB8A1wEFgpYg8oaqb21xXAHwKWNHi2GzgdmAOMA54XkSmq2rapcemBd1XAgv/cCne/jHJ4AxEl27AGlGLhHxEkjM7xAlwyo8QOvsgAFZRPdGLNpJ308vkvu0NnKkVID3MnjbEqSo/+tivePL7f6bmeF3XntTBNLfAD6jcvrfX9SqeOIZLP/Mh7Ej4dEtaQiGsSIRZn/4sVth0c2WxxcBOVd2tqgngIeDGFNd9Hfg3oOVI8Y3AQ6oaV9U9wM7m+6VlWtD9xBpVhTPxeMr4IKGA0Jy9BA055Fy+NplD2lIoaCJatAV/8mGaXlxgloa3sWv1PrYu35V2YFCs5IKREZPHU1NxlIIxpWjgs+P5N5ObtbZhOTY5JZlZ8XfOLVczYeEc3nrsefZvPUbe5CmMufxKwkVmReFAaKpPdHUeeKmItMxrf39zHqFTxgMHWjw+CCxpeQMROR+YqKpPisjn2zx3eZvnju+oMsM2QJ91ff9+zIxesKmjxhtWTiPRizcgTuvWsoR87FEnccqP4O0b28e1zC6rntqAmyY4O2GbC/7+ds7/wA2tjlftqWDXS6vwUwRogKmXLshY/UomjeWST3+QNVuGbG6woahSVRd2fllqImIB/wncmYnKmC6OfiC5Maz8ThZLCEgoTUsw5BOadrAPapbdAi9IuyO2Kim7M0ZMGc8FH70VJxo+PRfZCjk40Qhv+5dP4URN94PRoQqgZSL4Cc3HTikA5gIviche4ALgieaBws6e286wbUH3qy4k8KeTVBES6f783qHu3KtmsfyPa4g3tp/SZlnCpAvOTfm8BR98J+WLz2H9I89Se+gYo2ZO4dxbr6VwrFnhZ3RqJTBNRKaQDK63A+87dVJVa4DSU49F5CXgc6q6SkSagN+IyH+SHCScBrzZUWEmQPcDrc9NJknqIFB31P2hnuAdSj1jZDibffE0xp49ioNbj+AlzszgCEVDzL10OiPPar/jzSllMyZz9Zfv6o9qGkOIqnoicjfwDMlm1QOquklE7gNWqeoTHTx3k4g8DGwGPODjHc3gABOg+4cKsTXTiC7e2mEgTvlUBQILd+ukTq/NRkf3HOepH73EtuW7iOSGufi9i7n09iWEo+k3Vz3Fsiz+4Zd38ci/PsmKx9YQBIoTsrn8Axfyrk9dw9b+yY9kDDOq+v/bu/cYucoyjuPf38zudrc3C7alQktbWqypUAIUKzHhXi1qUSQESlBRIsG0ovGCGJQ/SHfbjRcAAAb8SURBVIgQlYTEJlABTQyXNpEiCZgC1thgKOk12osFWkPoxbrlopRud2d3Hv+YWTJsd2e3nZ05Z2d+n2STnpl3z/ucnPbpyXve93mfA57r89ldA7S9pM/xPcA9Q+3LCbpGul+dTs/pB8lOeWfISToCoqOFjrXnEx2t1Q0wAbu3vMEDNz1CrjNHvqcwlvz0r9awfvVmbl/57SEl6VGjW7jx7qtZ8rOr6HjvKG3jW8nWsLSoWTX5JWENdbx4AbndpxYSb+9Pvkw9+R7RsfZ88u+c2Iq2NIsIfvvDlXQe6fogOQPkjub495521j2xvsxvHyvbnGXsyWOcnK2uOEHXlOh8+Wzef+oiOjd8gs5Nc+hYex50H3sbIoBs0LpgB9nJJ1ZjIs0O7mkfcHFJ7miOl1ZtqHFEZunjBJ2AONJGbtd0cjtn0HNgEh1/PZfIZYmuzAdP1lLhJzvpv7Rdvonsxw4lHfawOnqk60MlOvvqfD+5PQTN0sIJOgV6Dkzk8KpL6Wk/CeLYGR1qytO6YAdQPwseTp09uVDgqB+ZrJhz4awaR2SWPk7QaZHPFup0DHBH1NZJZspbNM04gEbvoW896pGmpa2FhTdfREvbsS8Cm0c1c+Wtl9Q+KLOU8SyONCk3uyOTZ/SlW4i8UHYXRJbcviVEx+k1C2+4Lb7tCiR4/uF1ZLIZ8j15PjJpHN/45XWcMtOLRsycoFOk+83JNM/a1/9TtApDHYUcXpjb3jz193T96zvQPTJneUhi8W0L+dy3Lmb/awdpHTOKU86YlOjGqmZp4iGOFOnaNnPAp+h+c5byZCeM/NkOLW0tzJg3jSmzJjs5m5Vwgk6RzPgjkBv6LVGmm0zbm4M3NLMRyQk6RdTaxfHsqhIhontc9QIys0Q5QadI/u3xkDmOqXTRRM+7F1QvIDNL1Ih7SXjen19KOoSqyr91HeT+QaHYVTltaOwNtC74cS3CGrIzkg6gRFpiWTx8ewBU5sYzk47AjpOfoFNGJz0ETWeXaZGBpnno5EfJjE9Xcjaz4eUEnTLKTCAzcSWM/znQzIdvUSs0n4M++jhqSctjmZlVy4gb4mgUmdHXEC3nEId/A12vQGYctF2PRl+L5G2ZzBqBE3SKqWk2mnBf0mGYWUIqGuKQ9AtJ/5T0d0mrJU0YrsDMzBpdpWPQLwBnRcQ84FXgJ5WHZGZmUGGCjojnI6J3Pth6CtuIm5nZMBjOWRzfBP400JeSbpG0UdLG9vb2YezWzKw+DfqSUNKLwJR+vrozIv5YbHMnhZUVjw10nohYAawAmD9/fv1Unjczq5JBE3REXFHue0k3AV8ELo8YcPtTM7O6IGkR8ACQBR6OiHv7fH8rsJRCXeDDwC0RsUPSQuBeoAXoAn4UEWvL9VXRNLtioLcDF0fEkUrOZWaWdpKywHJgIbAX2CDpmYjYUdLs8Yh4sNj+KuB+YBFwCFgcEfslnQWsAU4r11+lY9C/BsYBL0jaKunBCs9nZpZmnwJej4g9EdEFPAl8qbRBRPyv5HAMxc1EI2JLROwvfr4daJM0qlxnFT1BR8TsE/m9TZs2HZL0RiV9p9xECv9bNqJGvXZfd/VMr/QE77ZvW/P08o9PHELTVkkbS45XFN+f9ToNKC3CvhdY0PckkpYC36cwnHFZP/1cA2yOiM5ywSSykjAi6nrDOUkbI2J+0nEkoVGv3dedbhGxqMb9LQeWS7oB+Cnw9d7vJH0SuA/47GDncbEkM7Oh2wdMKzmeWvxsIE8CX+49kDQVWA18LSJ2D9aZE7SZ2dBtAM6UNFOFqmXXA8+UNpBUWnj7C8Brxc8nAM8Cd0TE34bSmRN0dawYvEndatRr93U3gOLK6WUUZmDsBFZFxHZJdxdnbAAsk7Rd0lYK49C9wxvLgNnAXcVJFVslTS7Xnzx12cwsnfwEbWaWUk7QZmYp5QRdJZKuLY5D5SWlfhpSpSQtkrRL0uuS7kg6nlqR9Kik/0jalnQstSRpmqS/SNpR/Hv+3aRjqkdO0NWzDfgKsC7pQKqtZPnrlcBcYImkuclGVTO/o7CMt9F0Az+IiLnAp4GlDXTPa8YJukoiYmdE7Eo6jhoZdPlrvYqIdcDbScdRaxFxICI2F//8HoUZDWXrStjxc4K24dDf8lf/Y20QkmYA5wKvJBtJ/fGmsRUYSq1ss3omaSzwB+B7fYoE2TBwgq7AYLWyG8jxLn+1OiCpmUJyfiwinko6nnrkIQ4bDoMuf7X6IknAI8DOiLg/6XjqlRN0lUi6WtJe4ELgWUlrko6pWgZa/ppsVLUh6QngZWCOpL2Sbk46phr5DPBV4LKSZcufTzqoeuOl3mZmKeUnaDOzlHKCNjNLKSdoM7OUcoI2M0spJ2gzs5RygjYzSyknaDOzlPo/qazUaH/WF6kAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x15a36b2e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(100, 100)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x157856668>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def decision_surface(model, X, y):\n",
    "    x_vec= np.linspace(np.min(X[:,0]), np.max(X[:,0]), 100)\n",
    "    y_vec =np.linspace(np.min(X[:,1]), np.max(X[:,1]), 100)\n",
    "    x_grid, y_grid = np.meshgrid(x_vec, y_vec)\n",
    "\n",
    "    x_grid = x_grid.flatten()\n",
    "    y_grid = y_grid.flatten()\n",
    "\n",
    "    grid = np.array([x_grid, y_grid]).transpose()\n",
    "\n",
    "    out = model.predict(grid)\n",
    "    out = out.reshape(100,100,2)\n",
    "    print(np.shape(out[:,:,0]))\n",
    "\n",
    "    plt.figure()\n",
    "    plt.contourf(x_vec, y_vec, out[:,:,0],  cmap=\"coolwarm\")\n",
    "    plt.colorbar()\n",
    "    plt.scatter(X[:, 0], X[:, 1], marker='o', c=(1.-y),\n",
    "                s=50)\n",
    "    plt.show()\n",
    "\n",
    "decision_surface(model,X,y)\n",
    "decision_surface(model1,X,y)\n",
    "decision_surface(model2,X,y)\n",
    "\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x155ccc390>"
      ]
     },
     "execution_count": 174,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x13a27e080>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.datasets import make_circles\n",
    "X, y = make_circles(n_samples=400, factor=.3, noise=.1)\n",
    "labels = ['b', 'r']\n",
    "y_lab = np.take(labels, (y < 0.5))\n",
    "Y_ = one_hot(2, y)\n",
    "plt.scatter(X[:,0], X[:,1], c=y_lab)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(100, 100)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x155de5630>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "np.random.seed(123+3)\n",
    "\n",
    "# This returns a tensor to represent the input\n",
    "inputs = Input(shape=(2,))\n",
    "\n",
    "# a layer instance is callable on a tensor, and returns a tensor\n",
    "x = Dense(10)(inputs)\n",
    "x = Activation(\"relu\")(x)\n",
    "# x = Dense(10)(x)\n",
    "# x = Activation(\"softmax\")(x)\n",
    "# x = Dense(10)(x)\n",
    "# x = Activation(\"softmax\")(x)\n",
    "# x = Dense(10)(x)\n",
    "# x = Activation(\"softmax\")(x)\n",
    "x = Dense(2)(x)\n",
    "# to find out more about activations check the keras documentation\n",
    "predictions = Activation('softmax')(x)\n",
    "\n",
    "# This creates a model that includes\n",
    "# the Input layer and three Dense layers\n",
    "model = Model(inputs=inputs, outputs=predictions)\n",
    "model.compile(optimizer='sgd',\n",
    "              loss='mean_squared_error',\n",
    "              metrics=['accuracy'],\n",
    "              )\n",
    "model.fit(X, Y_, epochs=300, verbose=False)\n",
    "decision_surface(model,X,y)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Fashion neural network\n",
    "\n",
    "Now let's graduate to classifying fashion items. The structure should be very similar to the simple neural network but you might need more layers of different widths.\n",
    "\n",
    "* what network structures work?\n",
    "  * more layers or wider layers or both?\n",
    "* how good can you make your network?\n",
    "  * what did you use as baseline to compare your performance to?\n",
    "* experiment!\n",
    "* (bonus) how does your NN compare to a random forest with about 200 trees (or some other decision tree based classifier)?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "(X_train, y_train), (X_test, y_test) = fashion_mnist.load_data()\n",
    "\n",
    "X_train = X_train.astype('float32')\n",
    "X_test = X_test.astype('float32')\n",
    "X_train /= 255\n",
    "X_test /= 255"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Split off some validation data\n",
    "\n",
    "To measure our Neural Networks performance we will need some validation data. The `train_test_split` helper from scikit-learn does this for us."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_val, y_train, y_val = train_test_split(X_train, y_train,\n",
    "                                                  test_size=10000,\n",
    "                                                  random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### One more thing\n",
    "\n",
    "We need to convert the labels from integers (0, 1, 2, 3, ...) to  a one-hot encoding. The one-hot encoding for a problem with ten classes is a ten dimensional vector for each sample. For a sample in class 4 every entry is zero except for the fourth one. Let's check it out:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras import utils\n",
    "\n",
    "\n",
    "num_classes = 10\n",
    "y_train_ = utils.to_categorical(y_train, num_classes)\n",
    "y_val = utils.to_categorical(y_val, num_classes)\n",
    "y_test = utils.to_categorical(y_test, num_classes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([5, 0, 0, 1, 4], dtype=uint8)"
      ]
     },
     "execution_count": 159,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0., 0., 0., 0., 0., 1., 0., 0., 0., 0.],\n       [1., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n       [1., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n       [0., 1., 0., 0., 0., 0., 0., 0., 0., 0.],\n       [0., 0., 0., 0., 1., 0., 0., 0., 0., 0.]])"
      ]
     },
     "execution_count": 160,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# modified y\n",
    "y_train_[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "metadata": {},
   "outputs": [],
   "source": [
    "# let's make y_train the same as the others\n",
    "y_train = utils.to_categorical(y_train, num_classes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Model building\n",
    "\n",
    "We now define the model architecture and train the model. To learn more about the building blocks that are available check out the [keras documentation](https://keras.io/layers/about-keras-layers/)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import Model\n",
    "from keras.layers import Input, Dense, Activation, Flatten"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {},
   "outputs": [],
   "source": [
    "# we define the input shape (i.e., how many input features) **without** the batch size\n",
    "x = Input(shape=(28, 28, ))\n",
    "\n",
    "# turn a 28x28 matrix into a 784-d vector, this removes all information\n",
    "# about the spatial relation between pixels. Using convolutions will\n",
    "# allow us to take advantage of that information (see later)\n",
    "h = Flatten()(x)\n",
    "\n",
    "h = Dense(50)(h)\n",
    "h = Activation('tanh')(h)\n",
    "h = Dense(50)(h)\n",
    "h = Activation('tanh')(h)\n",
    "h = Dense(50)(h)\n",
    "h = Activation('tanh')(h)\n",
    "h = Dense(50)(h)\n",
    "h = Activation('tanh')(h)\n",
    "h = Dense(50)(h)\n",
    "h = Activation('tanh')(h)\n",
    "\n",
    "h = Dense(2)(h)\n",
    "\n",
    "# we want to predict one of ten classes\n",
    "h = Dense(10)(h)\n",
    "y = Activation('softmax')(h)\n",
    "\n",
    "\n",
    "# Package it all up in a Model\n",
    "net = Model(x, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model structure\n",
    "\n",
    "You can print out the structure of your network and check how many parameters it has, etc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\nLayer (type)                 Output Shape              Param #   \n=================================================================\ninput_46 (InputLayer)        (None, 28, 28)            0         \n_________________________________________________________________\nflatten_8 (Flatten)          (None, 784)               0         \n_________________________________________________________________\ndense_109 (Dense)            (None, 50)                39250     \n_________________________________________________________________\nactivation_123 (Activation)  (None, 50)                0         \n_________________________________________________________________\ndense_110 (Dense)            (None, 50)                2550      \n_________________________________________________________________\nactivation_124 (Activation)  (None, 50)                0         \n_________________________________________________________________\ndense_111 (Dense)            (None, 50)                2550      \n_________________________________________________________________\nactivation_125 (Activation)  (None, 50)                0         \n_________________________________________________________________\ndense_112 (Dense)            (None, 50)                2550      \n_________________________________________________________________\nactivation_126 (Activation)  (None, 50)                0         \n_________________________________________________________________\ndense_113 (Dense)            (None, 50)                2550      \n_________________________________________________________________\nactivation_127 (Activation)  (None, 50)                0         \n_________________________________________________________________\ndense_114 (Dense)            (None, 2)                 102       \n_________________________________________________________________\ndense_115 (Dense)            (None, 10)                30        \n_________________________________________________________________\nactivation_128 (Activation)  (None, 10)                0         \n=================================================================\nTotal params: 49,582\nTrainable params: 49,582\nNon-trainable params: 0\n_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "net.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "metadata": {},
   "outputs": [],
   "source": [
    "net.compile(loss='categorical_crossentropy',\n",
    "            optimizer='sgd',\n",
    "            metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 50000 samples, validate on 10000 samples\nEpoch 1/20\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\r  128/50000 [..............................] - ETA: 6:42 - loss: 2.3053 - acc: 0.0859"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2176/50000 [>.............................] - ETA: 23s - loss: 2.2861 - acc: 0.1213 "
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4224/50000 [=>............................] - ETA: 12s - loss: 2.2729 - acc: 0.1404"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6272/50000 [==>...........................] - ETA: 8s - loss: 2.2608 - acc: 0.1496 "
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 8192/50000 [===>..........................] - ETA: 6s - loss: 2.2481 - acc: 0.1553"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 9984/50000 [====>.........................] - ETA: 5s - loss: 2.2373 - acc: 0.1558"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r11904/50000 [======>.......................] - ETA: 4s - loss: 2.2241 - acc: 0.1585"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r13952/50000 [=======>......................] - ETA: 3s - loss: 2.2118 - acc: 0.1598"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r16000/50000 [========>.....................] - ETA: 3s - loss: 2.1991 - acc: 0.1606"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18176/50000 [=========>....................] - ETA: 2s - loss: 2.1854 - acc: 0.1621"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r20224/50000 [===========>..................] - ETA: 2s - loss: 2.1707 - acc: 0.1656"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r22272/50000 [============>.................] - ETA: 1s - loss: 2.1585 - acc: 0.1696"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r24320/50000 [=============>................] - ETA: 1s - loss: 2.1452 - acc: 0.1741"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r26496/50000 [==============>...............] - ETA: 1s - loss: 2.1320 - acc: 0.1789"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r28672/50000 [================>.............] - ETA: 1s - loss: 2.1193 - acc: 0.1846"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r30848/50000 [=================>............] - ETA: 1s - loss: 2.1048 - acc: 0.1891"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r33024/50000 [==================>...........] - ETA: 0s - loss: 2.0919 - acc: 0.1946"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r35072/50000 [====================>.........] - ETA: 0s - loss: 2.0785 - acc: 0.2009"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r37248/50000 [=====================>........] - ETA: 0s - loss: 2.0656 - acc: 0.2063"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r39296/50000 [======================>.......] - ETA: 0s - loss: 2.0531 - acc: 0.2121"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r41344/50000 [=======================>......] - ETA: 0s - loss: 2.0400 - acc: 0.2188"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r43520/50000 [=========================>....] - ETA: 0s - loss: 2.0257 - acc: 0.2245"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r45696/50000 [==========================>...] - ETA: 0s - loss: 2.0116 - acc: 0.2307"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r47872/50000 [===========================>..] - ETA: 0s - loss: 1.9964 - acc: 0.2368"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 3s 50us/step - loss: 1.9829 - acc: 0.2437 - val_loss: 1.6638 - val_acc: 0.3943\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 2/20\n\r  128/50000 [..............................] - ETA: 2s - loss: 1.5671 - acc: 0.3828"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2048/50000 [>.............................] - ETA: 1s - loss: 1.6286 - acc: 0.3945"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4224/50000 [=>............................] - ETA: 1s - loss: 1.6359 - acc: 0.4020"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6400/50000 [==>...........................] - ETA: 1s - loss: 1.6203 - acc: 0.4031"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 8576/50000 [====>.........................] - ETA: 1s - loss: 1.6046 - acc: 0.4067"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r10752/50000 [=====>........................] - ETA: 0s - loss: 1.5898 - acc: 0.4116"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r12928/50000 [======>.......................] - ETA: 0s - loss: 1.5785 - acc: 0.4174"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r14976/50000 [=======>......................] - ETA: 0s - loss: 1.5658 - acc: 0.4207"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r17024/50000 [=========>....................] - ETA: 0s - loss: 1.5514 - acc: 0.4256"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r19072/50000 [==========>...................] - ETA: 0s - loss: 1.5393 - acc: 0.4277"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r21120/50000 [===========>..................] - ETA: 0s - loss: 1.5272 - acc: 0.4328"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r23296/50000 [============>.................] - ETA: 0s - loss: 1.5152 - acc: 0.4367"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r25472/50000 [==============>...............] - ETA: 0s - loss: 1.5015 - acc: 0.4413"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r27520/50000 [===============>..............] - ETA: 0s - loss: 1.4902 - acc: 0.4429"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r29568/50000 [================>.............] - ETA: 0s - loss: 1.4780 - acc: 0.4450"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r31616/50000 [=================>............] - ETA: 0s - loss: 1.4655 - acc: 0.4487"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r33792/50000 [===================>..........] - ETA: 0s - loss: 1.4548 - acc: 0.4518"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r36096/50000 [====================>.........] - ETA: 0s - loss: 1.4421 - acc: 0.4557"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r38272/50000 [=====================>........] - ETA: 0s - loss: 1.4297 - acc: 0.4595"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r40448/50000 [=======================>......] - ETA: 0s - loss: 1.4176 - acc: 0.4628"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r42496/50000 [========================>.....] - ETA: 0s - loss: 1.4074 - acc: 0.4662"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r44544/50000 [=========================>....] - ETA: 0s - loss: 1.3966 - acc: 0.4693"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r46592/50000 [==========================>...] - ETA: 0s - loss: 1.3872 - acc: 0.4722"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r48768/50000 [============================>.] - ETA: 0s - loss: 1.3758 - acc: 0.4759"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 1s 27us/step - loss: 1.3702 - acc: 0.4770 - val_loss: 1.1568 - val_acc: 0.5402\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 3/20\n\r  128/50000 [..............................] - ETA: 1s - loss: 1.1578 - acc: 0.5625"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2176/50000 [>.............................] - ETA: 1s - loss: 1.1268 - acc: 0.5680"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4224/50000 [=>............................] - ETA: 1s - loss: 1.1312 - acc: 0.5620"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6400/50000 [==>...........................] - ETA: 1s - loss: 1.1288 - acc: 0.5636"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 8576/50000 [====>.........................] - ETA: 1s - loss: 1.1250 - acc: 0.5654"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r10624/50000 [=====>........................] - ETA: 0s - loss: 1.1158 - acc: 0.5730"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r12800/50000 [======>.......................] - ETA: 0s - loss: 1.1086 - acc: 0.5782"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r14976/50000 [=======>......................] - ETA: 0s - loss: 1.1026 - acc: 0.5768"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r17152/50000 [=========>....................] - ETA: 0s - loss: 1.0968 - acc: 0.5803"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r19456/50000 [==========>...................] - ETA: 0s - loss: 1.0905 - acc: 0.5833"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r21760/50000 [============>.................] - ETA: 0s - loss: 1.0854 - acc: 0.5890"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r23936/50000 [=============>................] - ETA: 0s - loss: 1.0835 - acc: 0.5934"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r26112/50000 [==============>...............] - ETA: 0s - loss: 1.0761 - acc: 0.5980"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r28160/50000 [===============>..............] - ETA: 0s - loss: 1.0715 - acc: 0.6012"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r30208/50000 [=================>............] - ETA: 0s - loss: 1.0687 - acc: 0.6035"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r32384/50000 [==================>...........] - ETA: 0s - loss: 1.0648 - acc: 0.6059"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r34560/50000 [===================>..........] - ETA: 0s - loss: 1.0618 - acc: 0.6080"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r36608/50000 [====================>.........] - ETA: 0s - loss: 1.0608 - acc: 0.6099"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r38912/50000 [======================>.......] - ETA: 0s - loss: 1.0561 - acc: 0.6126"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r41088/50000 [=======================>......] - ETA: 0s - loss: 1.0533 - acc: 0.6141"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r43264/50000 [========================>.....] - ETA: 0s - loss: 1.0518 - acc: 0.6160"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r45568/50000 [==========================>...] - ETA: 0s - loss: 1.0488 - acc: 0.6177"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r47872/50000 [===========================>..] - ETA: 0s - loss: 1.0451 - acc: 0.6202"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r49920/50000 [============================>.] - ETA: 0s - loss: 1.0419 - acc: 0.6217"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 1s 26us/step - loss: 1.0420 - acc: 0.6217 - val_loss: 0.9767 - val_acc: 0.6518\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 4/20\n\r  128/50000 [..............................] - ETA: 1s - loss: 0.9202 - acc: 0.6406"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2048/50000 [>.............................] - ETA: 1s - loss: 0.9451 - acc: 0.6626"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4096/50000 [=>............................] - ETA: 1s - loss: 0.9592 - acc: 0.6636"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6144/50000 [==>...........................] - ETA: 1s - loss: 0.9528 - acc: 0.6709"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 8192/50000 [===>..........................] - ETA: 1s - loss: 0.9527 - acc: 0.6652"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r10240/50000 [=====>........................] - ETA: 1s - loss: 0.9606 - acc: 0.6654"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r12288/50000 [======>.......................] - ETA: 0s - loss: 0.9618 - acc: 0.6654"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r14464/50000 [=======>......................] - ETA: 0s - loss: 0.9635 - acc: 0.6643"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r16640/50000 [========>.....................] - ETA: 0s - loss: 0.9598 - acc: 0.6643"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18688/50000 [==========>...................] - ETA: 0s - loss: 0.9571 - acc: 0.6654"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r20864/50000 [===========>..................] - ETA: 0s - loss: 0.9516 - acc: 0.6665"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r23040/50000 [============>.................] - ETA: 0s - loss: 0.9480 - acc: 0.6672"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r25216/50000 [==============>...............] - ETA: 0s - loss: 0.9473 - acc: 0.6689"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r27264/50000 [===============>..............] - ETA: 0s - loss: 0.9441 - acc: 0.6692"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r29440/50000 [================>.............] - ETA: 0s - loss: 0.9425 - acc: 0.6687"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r31616/50000 [=================>............] - ETA: 0s - loss: 0.9393 - acc: 0.6685"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r33792/50000 [===================>..........] - ETA: 0s - loss: 0.9376 - acc: 0.6686"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r35968/50000 [====================>.........] - ETA: 0s - loss: 0.9351 - acc: 0.6695"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r38016/50000 [=====================>........] - ETA: 0s - loss: 0.9343 - acc: 0.6698"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r40192/50000 [=======================>......] - ETA: 0s - loss: 0.9323 - acc: 0.6707"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r42368/50000 [========================>.....] - ETA: 0s - loss: 0.9302 - acc: 0.6712"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r44544/50000 [=========================>....] - ETA: 0s - loss: 0.9292 - acc: 0.6720"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r46720/50000 [===========================>..] - ETA: 0s - loss: 0.9284 - acc: 0.6718"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r48768/50000 [============================>.] - ETA: 0s - loss: 0.9256 - acc: 0.6730"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 1s 27us/step - loss: 0.9246 - acc: 0.6732 - val_loss: 0.8999 - val_acc: 0.6818\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 5/20\n\r  128/50000 [..............................] - ETA: 2s - loss: 0.8904 - acc: 0.7109"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2048/50000 [>.............................] - ETA: 1s - loss: 0.9006 - acc: 0.6865"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4096/50000 [=>............................] - ETA: 1s - loss: 0.8972 - acc: 0.6807"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6272/50000 [==>...........................] - ETA: 1s - loss: 0.8822 - acc: 0.6842"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 8448/50000 [====>.........................] - ETA: 1s - loss: 0.8749 - acc: 0.6857"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r10624/50000 [=====>........................] - ETA: 0s - loss: 0.8713 - acc: 0.6890"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r12800/50000 [======>.......................] - ETA: 0s - loss: 0.8777 - acc: 0.6863"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r14848/50000 [=======>......................] - ETA: 0s - loss: 0.8783 - acc: 0.6868"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r17024/50000 [=========>....................] - ETA: 0s - loss: 0.8782 - acc: 0.6874"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r19200/50000 [==========>...................] - ETA: 0s - loss: 0.8808 - acc: 0.6878"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r21376/50000 [===========>..................] - ETA: 0s - loss: 0.8808 - acc: 0.6861"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r23424/50000 [=============>................] - ETA: 0s - loss: 0.8778 - acc: 0.6866"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r25472/50000 [==============>...............] - ETA: 0s - loss: 0.8777 - acc: 0.6869"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r27520/50000 [===============>..............] - ETA: 0s - loss: 0.8749 - acc: 0.6886"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r29696/50000 [================>.............] - ETA: 0s - loss: 0.8750 - acc: 0.6884"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r31872/50000 [==================>...........] - ETA: 0s - loss: 0.8749 - acc: 0.6884"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r34048/50000 [===================>..........] - ETA: 0s - loss: 0.8733 - acc: 0.6902"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r36224/50000 [====================>.........] - ETA: 0s - loss: 0.8711 - acc: 0.6912"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r38272/50000 [=====================>........] - ETA: 0s - loss: 0.8700 - acc: 0.6914"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r40448/50000 [=======================>......] - ETA: 0s - loss: 0.8696 - acc: 0.6923"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r42752/50000 [========================>.....] - ETA: 0s - loss: 0.8672 - acc: 0.6935"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r44928/50000 [=========================>....] - ETA: 0s - loss: 0.8647 - acc: 0.6944"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r47104/50000 [===========================>..] - ETA: 0s - loss: 0.8651 - acc: 0.6940"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r49280/50000 [============================>.] - ETA: 0s - loss: 0.8644 - acc: 0.6943"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 1s 27us/step - loss: 0.8635 - acc: 0.6947 - val_loss: 0.8533 - val_acc: 0.7000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 6/20\n\r  128/50000 [..............................] - ETA: 2s - loss: 0.9415 - acc: 0.7109"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2048/50000 [>.............................] - ETA: 1s - loss: 0.8501 - acc: 0.7070"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4224/50000 [=>............................] - ETA: 1s - loss: 0.8532 - acc: 0.7086"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6400/50000 [==>...........................] - ETA: 1s - loss: 0.8303 - acc: 0.7066"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 8576/50000 [====>.........................] - ETA: 1s - loss: 0.8375 - acc: 0.7037"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r10624/50000 [=====>........................] - ETA: 0s - loss: 0.8319 - acc: 0.7043"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r12800/50000 [======>.......................] - ETA: 0s - loss: 0.8253 - acc: 0.7080"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r14848/50000 [=======>......................] - ETA: 0s - loss: 0.8285 - acc: 0.7068"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r17024/50000 [=========>....................] - ETA: 0s - loss: 0.8304 - acc: 0.7070"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r19200/50000 [==========>...................] - ETA: 0s - loss: 0.8295 - acc: 0.7081"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r21376/50000 [===========>..................] - ETA: 0s - loss: 0.8290 - acc: 0.7091"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r23552/50000 [=============>................] - ETA: 0s - loss: 0.8311 - acc: 0.7084"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r25600/50000 [==============>...............] - ETA: 0s - loss: 0.8301 - acc: 0.7101"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r27648/50000 [===============>..............] - ETA: 0s - loss: 0.8283 - acc: 0.7104"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r29824/50000 [================>.............] - ETA: 0s - loss: 0.8268 - acc: 0.7101"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r32000/50000 [==================>...........] - ETA: 0s - loss: 0.8258 - acc: 0.7108"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r34176/50000 [===================>..........] - ETA: 0s - loss: 0.8240 - acc: 0.7120"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r36352/50000 [====================>.........] - ETA: 0s - loss: 0.8234 - acc: 0.7123"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r38400/50000 [======================>.......] - ETA: 0s - loss: 0.8230 - acc: 0.7123"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r40576/50000 [=======================>......] - ETA: 0s - loss: 0.8214 - acc: 0.7136"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r42752/50000 [========================>.....] - ETA: 0s - loss: 0.8203 - acc: 0.7139"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r44928/50000 [=========================>....] - ETA: 0s - loss: 0.8220 - acc: 0.7138"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r46976/50000 [===========================>..] - ETA: 0s - loss: 0.8233 - acc: 0.7135"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r49152/50000 [============================>.] - ETA: 0s - loss: 0.8220 - acc: 0.7137"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 1s 27us/step - loss: 0.8215 - acc: 0.7139 - val_loss: 0.8207 - val_acc: 0.7170\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 7/20\n\r  128/50000 [..............................] - ETA: 1s - loss: 0.7900 - acc: 0.7109"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2176/50000 [>.............................] - ETA: 1s - loss: 0.8445 - acc: 0.7050"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4224/50000 [=>............................] - ETA: 1s - loss: 0.8249 - acc: 0.7131"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6272/50000 [==>...........................] - ETA: 1s - loss: 0.8156 - acc: 0.7176"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 8320/50000 [===>..........................] - ETA: 1s - loss: 0.8144 - acc: 0.7215"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r10368/50000 [=====>........................] - ETA: 0s - loss: 0.8082 - acc: 0.7220"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r12544/50000 [======>.......................] - ETA: 0s - loss: 0.8095 - acc: 0.7230"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r14592/50000 [=======>......................] - ETA: 0s - loss: 0.8142 - acc: 0.7223"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r16512/50000 [========>.....................] - ETA: 0s - loss: 0.8102 - acc: 0.7220"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18432/50000 [==========>...................] - ETA: 0s - loss: 0.8038 - acc: 0.7230"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r20480/50000 [===========>..................] - ETA: 0s - loss: 0.8033 - acc: 0.7236"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r22656/50000 [============>.................] - ETA: 0s - loss: 0.8021 - acc: 0.7230"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r24832/50000 [=============>................] - ETA: 0s - loss: 0.8001 - acc: 0.7247"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r27008/50000 [===============>..............] - ETA: 0s - loss: 0.8037 - acc: 0.7240"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r29184/50000 [================>.............] - ETA: 0s - loss: 0.7988 - acc: 0.7253"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r31232/50000 [=================>............] - ETA: 0s - loss: 0.7965 - acc: 0.7262"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r33408/50000 [===================>..........] - ETA: 0s - loss: 0.7955 - acc: 0.7266"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r35456/50000 [====================>.........] - ETA: 0s - loss: 0.7957 - acc: 0.7259"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r37504/50000 [=====================>........] - ETA: 0s - loss: 0.7948 - acc: 0.7268"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r39680/50000 [======================>.......] - ETA: 0s - loss: 0.7948 - acc: 0.7267"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r41856/50000 [========================>.....] - ETA: 0s - loss: 0.7936 - acc: 0.7274"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r43904/50000 [=========================>....] - ETA: 0s - loss: 0.7936 - acc: 0.7270"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r46080/50000 [==========================>...] - ETA: 0s - loss: 0.7933 - acc: 0.7279"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r48256/50000 [===========================>..] - ETA: 0s - loss: 0.7912 - acc: 0.7289"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 1s 27us/step - loss: 0.7897 - acc: 0.7297 - val_loss: 0.7820 - val_acc: 0.7360\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 8/20\n\r  128/50000 [..............................] - ETA: 1s - loss: 0.5808 - acc: 0.8203"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2176/50000 [>.............................] - ETA: 1s - loss: 0.7702 - acc: 0.7431"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4224/50000 [=>............................] - ETA: 1s - loss: 0.7854 - acc: 0.7391"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6400/50000 [==>...........................] - ETA: 1s - loss: 0.7703 - acc: 0.7414"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 8576/50000 [====>.........................] - ETA: 1s - loss: 0.7625 - acc: 0.7382"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r10752/50000 [=====>........................] - ETA: 0s - loss: 0.7671 - acc: 0.7393"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r12800/50000 [======>.......................] - ETA: 0s - loss: 0.7690 - acc: 0.7393"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r14976/50000 [=======>......................] - ETA: 0s - loss: 0.7644 - acc: 0.7402"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r17024/50000 [=========>....................] - ETA: 0s - loss: 0.7658 - acc: 0.7417"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r19200/50000 [==========>...................] - ETA: 0s - loss: 0.7653 - acc: 0.7412"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r21376/50000 [===========>..................] - ETA: 0s - loss: 0.7609 - acc: 0.7432"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r23424/50000 [=============>................] - ETA: 0s - loss: 0.7606 - acc: 0.7439"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r25600/50000 [==============>...............] - ETA: 0s - loss: 0.7618 - acc: 0.7445"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r27776/50000 [===============>..............] - ETA: 0s - loss: 0.7610 - acc: 0.7441"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r29952/50000 [================>.............] - ETA: 0s - loss: 0.7615 - acc: 0.7433"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r32128/50000 [==================>...........] - ETA: 0s - loss: 0.7612 - acc: 0.7435"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r34304/50000 [===================>..........] - ETA: 0s - loss: 0.7613 - acc: 0.7437"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r36352/50000 [====================>.........] - ETA: 0s - loss: 0.7616 - acc: 0.7437"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r38528/50000 [======================>.......] - ETA: 0s - loss: 0.7634 - acc: 0.7434"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r40704/50000 [=======================>......] - ETA: 0s - loss: 0.7635 - acc: 0.7440"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r42752/50000 [========================>.....] - ETA: 0s - loss: 0.7627 - acc: 0.7448"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r44800/50000 [=========================>....] - ETA: 0s - loss: 0.7613 - acc: 0.7456"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r46848/50000 [===========================>..] - ETA: 0s - loss: 0.7624 - acc: 0.7452"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r48896/50000 [============================>.] - ETA: 0s - loss: 0.7625 - acc: 0.7453"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 1s 27us/step - loss: 0.7618 - acc: 0.7455 - val_loss: 0.7600 - val_acc: 0.7310\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 9/20\n\r  128/50000 [..............................] - ETA: 2s - loss: 0.6830 - acc: 0.7734"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2048/50000 [>.............................] - ETA: 1s - loss: 0.7350 - acc: 0.7402"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4224/50000 [=>............................] - ETA: 1s - loss: 0.7509 - acc: 0.7446"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6400/50000 [==>...........................] - ETA: 1s - loss: 0.7455 - acc: 0.7484"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 8448/50000 [====>.........................] - ETA: 1s - loss: 0.7473 - acc: 0.7502"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r10624/50000 [=====>........................] - ETA: 0s - loss: 0.7412 - acc: 0.7518"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r12672/50000 [======>.......................] - ETA: 0s - loss: 0.7468 - acc: 0.7517"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r14848/50000 [=======>......................] - ETA: 0s - loss: 0.7435 - acc: 0.7531"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r17024/50000 [=========>....................] - ETA: 0s - loss: 0.7466 - acc: 0.7535"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r19072/50000 [==========>...................] - ETA: 0s - loss: 0.7443 - acc: 0.7542"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r21376/50000 [===========>..................] - ETA: 0s - loss: 0.7429 - acc: 0.7550"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r23680/50000 [=============>................] - ETA: 0s - loss: 0.7421 - acc: 0.7549"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r25984/50000 [==============>...............] - ETA: 0s - loss: 0.7414 - acc: 0.7542"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r28288/50000 [===============>..............] - ETA: 0s - loss: 0.7387 - acc: 0.7547"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r30592/50000 [=================>............] - ETA: 0s - loss: 0.7403 - acc: 0.7535"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r32896/50000 [==================>...........] - ETA: 0s - loss: 0.7381 - acc: 0.7548"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r35200/50000 [====================>.........] - ETA: 0s - loss: 0.7363 - acc: 0.7555"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r37504/50000 [=====================>........] - ETA: 0s - loss: 0.7372 - acc: 0.7552"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r39808/50000 [======================>.......] - ETA: 0s - loss: 0.7380 - acc: 0.7551"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r42112/50000 [========================>.....] - ETA: 0s - loss: 0.7363 - acc: 0.7551"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r44416/50000 [=========================>....] - ETA: 0s - loss: 0.7355 - acc: 0.7559"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r46592/50000 [==========================>...] - ETA: 0s - loss: 0.7349 - acc: 0.7563"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r48768/50000 [============================>.] - ETA: 0s - loss: 0.7349 - acc: 0.7564"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 1s 25us/step - loss: 0.7353 - acc: 0.7562 - val_loss: 0.7516 - val_acc: 0.7579\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 10/20\n\r  128/50000 [..............................] - ETA: 1s - loss: 0.9168 - acc: 0.7969"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2176/50000 [>.............................] - ETA: 1s - loss: 0.7525 - acc: 0.7486"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4352/50000 [=>............................] - ETA: 1s - loss: 0.7288 - acc: 0.7597"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6656/50000 [==>...........................] - ETA: 1s - loss: 0.7225 - acc: 0.7638"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 8960/50000 [====>.........................] - ETA: 0s - loss: 0.7135 - acc: 0.7660"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r11264/50000 [=====>........................] - ETA: 0s - loss: 0.7266 - acc: 0.7614"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r13568/50000 [=======>......................] - ETA: 0s - loss: 0.7221 - acc: 0.7636"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r15872/50000 [========>.....................] - ETA: 0s - loss: 0.7254 - acc: 0.7617"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18176/50000 [=========>....................] - ETA: 0s - loss: 0.7252 - acc: 0.7624"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r20480/50000 [===========>..................] - ETA: 0s - loss: 0.7236 - acc: 0.7632"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r22784/50000 [============>.................] - ETA: 0s - loss: 0.7228 - acc: 0.7643"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r25088/50000 [==============>...............] - ETA: 0s - loss: 0.7204 - acc: 0.7641"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r27392/50000 [===============>..............] - ETA: 0s - loss: 0.7183 - acc: 0.7658"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r29696/50000 [================>.............] - ETA: 0s - loss: 0.7184 - acc: 0.7658"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r32000/50000 [==================>...........] - ETA: 0s - loss: 0.7185 - acc: 0.7658"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r34304/50000 [===================>..........] - ETA: 0s - loss: 0.7164 - acc: 0.7666"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r36608/50000 [====================>.........] - ETA: 0s - loss: 0.7177 - acc: 0.7665"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r38912/50000 [======================>.......] - ETA: 0s - loss: 0.7171 - acc: 0.7670"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r41216/50000 [=======================>......] - ETA: 0s - loss: 0.7137 - acc: 0.7676"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r43520/50000 [=========================>....] - ETA: 0s - loss: 0.7143 - acc: 0.7679"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r45824/50000 [==========================>...] - ETA: 0s - loss: 0.7132 - acc: 0.7683"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r47872/50000 [===========================>..] - ETA: 0s - loss: 0.7121 - acc: 0.7683"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 1s 25us/step - loss: 0.7112 - acc: 0.7686 - val_loss: 0.7143 - val_acc: 0.7639\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 11/20\n\r  128/50000 [..............................] - ETA: 1s - loss: 0.6145 - acc: 0.7500"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2304/50000 [>.............................] - ETA: 1s - loss: 0.6896 - acc: 0.7747"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4608/50000 [=>............................] - ETA: 1s - loss: 0.7067 - acc: 0.7689"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6656/50000 [==>...........................] - ETA: 1s - loss: 0.6992 - acc: 0.7704"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 8960/50000 [====>.........................] - ETA: 0s - loss: 0.6963 - acc: 0.7725"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r11264/50000 [=====>........................] - ETA: 0s - loss: 0.6885 - acc: 0.7724"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r13568/50000 [=======>......................] - ETA: 0s - loss: 0.6845 - acc: 0.7752"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r15744/50000 [========>.....................] - ETA: 0s - loss: 0.6881 - acc: 0.7727"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18048/50000 [=========>....................] - ETA: 0s - loss: 0.6909 - acc: 0.7723"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r20352/50000 [===========>..................] - ETA: 0s - loss: 0.6920 - acc: 0.7720"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r22656/50000 [============>.................] - ETA: 0s - loss: 0.6962 - acc: 0.7729"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r24960/50000 [=============>................] - ETA: 0s - loss: 0.6956 - acc: 0.7730"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r27264/50000 [===============>..............] - ETA: 0s - loss: 0.6950 - acc: 0.7726"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r29568/50000 [================>.............] - ETA: 0s - loss: 0.6931 - acc: 0.7732"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r31872/50000 [==================>...........] - ETA: 0s - loss: 0.6956 - acc: 0.7726"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r34176/50000 [===================>..........] - ETA: 0s - loss: 0.6960 - acc: 0.7720"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r36480/50000 [====================>.........] - ETA: 0s - loss: 0.6946 - acc: 0.7727"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r38784/50000 [======================>.......] - ETA: 0s - loss: 0.6909 - acc: 0.7741"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r41088/50000 [=======================>......] - ETA: 0s - loss: 0.6906 - acc: 0.7737"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r43392/50000 [=========================>....] - ETA: 0s - loss: 0.6905 - acc: 0.7738"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r45696/50000 [==========================>...] - ETA: 0s - loss: 0.6897 - acc: 0.7740"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r47872/50000 [===========================>..] - ETA: 0s - loss: 0.6895 - acc: 0.7738"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 1s 25us/step - loss: 0.6887 - acc: 0.7741 - val_loss: 0.7266 - val_acc: 0.7624\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 12/20\n\r  128/50000 [..............................] - ETA: 1s - loss: 0.7443 - acc: 0.8281"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2176/50000 [>.............................] - ETA: 1s - loss: 0.6688 - acc: 0.7808"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4480/50000 [=>............................] - ETA: 1s - loss: 0.6723 - acc: 0.7850"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6784/50000 [===>..........................] - ETA: 1s - loss: 0.6718 - acc: 0.7811"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 9088/50000 [====>.........................] - ETA: 0s - loss: 0.6567 - acc: 0.7841"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r11392/50000 [=====>........................] - ETA: 0s - loss: 0.6531 - acc: 0.7865"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r13696/50000 [=======>......................] - ETA: 0s - loss: 0.6494 - acc: 0.7870"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r16000/50000 [========>.....................] - ETA: 0s - loss: 0.6587 - acc: 0.7836"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18304/50000 [=========>....................] - ETA: 0s - loss: 0.6549 - acc: 0.7834"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r20608/50000 [===========>..................] - ETA: 0s - loss: 0.6566 - acc: 0.7833"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r22784/50000 [============>.................] - ETA: 0s - loss: 0.6598 - acc: 0.7832"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r25088/50000 [==============>...............] - ETA: 0s - loss: 0.6613 - acc: 0.7827"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r27392/50000 [===============>..............] - ETA: 0s - loss: 0.6633 - acc: 0.7818"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r29696/50000 [================>.............] - ETA: 0s - loss: 0.6616 - acc: 0.7824"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r32128/50000 [==================>...........] - ETA: 0s - loss: 0.6633 - acc: 0.7813"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r34432/50000 [===================>..........] - ETA: 0s - loss: 0.6628 - acc: 0.7816"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r36736/50000 [=====================>........] - ETA: 0s - loss: 0.6621 - acc: 0.7816"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r39040/50000 [======================>.......] - ETA: 0s - loss: 0.6625 - acc: 0.7815"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r41344/50000 [=======================>......] - ETA: 0s - loss: 0.6635 - acc: 0.7819"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r43648/50000 [=========================>....] - ETA: 0s - loss: 0.6628 - acc: 0.7819"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r45952/50000 [==========================>...] - ETA: 0s - loss: 0.6617 - acc: 0.7824"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r48128/50000 [===========================>..] - ETA: 0s - loss: 0.6640 - acc: 0.7825"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 1s 25us/step - loss: 0.6634 - acc: 0.7824 - val_loss: 0.6713 - val_acc: 0.7783\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 13/20\n\r  128/50000 [..............................] - ETA: 1s - loss: 0.7240 - acc: 0.7188"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2304/50000 [>.............................] - ETA: 1s - loss: 0.6606 - acc: 0.7865"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4608/50000 [=>............................] - ETA: 1s - loss: 0.6552 - acc: 0.7891"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6912/50000 [===>..........................] - ETA: 0s - loss: 0.6578 - acc: 0.7849"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 9216/50000 [====>.........................] - ETA: 0s - loss: 0.6508 - acc: 0.7865"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r11520/50000 [=====>........................] - ETA: 0s - loss: 0.6511 - acc: 0.7865"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r13824/50000 [=======>......................] - ETA: 0s - loss: 0.6532 - acc: 0.7861"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r16128/50000 [========>.....................] - ETA: 0s - loss: 0.6510 - acc: 0.7875"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18432/50000 [==========>...................] - ETA: 0s - loss: 0.6511 - acc: 0.7892"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r20736/50000 [===========>..................] - ETA: 0s - loss: 0.6500 - acc: 0.7890"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r23040/50000 [============>.................] - ETA: 0s - loss: 0.6468 - acc: 0.7896"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r25344/50000 [==============>...............] - ETA: 0s - loss: 0.6478 - acc: 0.7883"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r27648/50000 [===============>..............] - ETA: 0s - loss: 0.6482 - acc: 0.7888"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r29952/50000 [================>.............] - ETA: 0s - loss: 0.6500 - acc: 0.7877"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r32256/50000 [==================>...........] - ETA: 0s - loss: 0.6502 - acc: 0.7881"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r34560/50000 [===================>..........] - ETA: 0s - loss: 0.6517 - acc: 0.7868"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r36864/50000 [=====================>........] - ETA: 0s - loss: 0.6547 - acc: 0.7858"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r39168/50000 [======================>.......] - ETA: 0s - loss: 0.6517 - acc: 0.7867"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r41472/50000 [=======================>......] - ETA: 0s - loss: 0.6508 - acc: 0.7869"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r43648/50000 [=========================>....] - ETA: 0s - loss: 0.6488 - acc: 0.7868"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r45952/50000 [==========================>...] - ETA: 0s - loss: 0.6494 - acc: 0.7865"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r48256/50000 [===========================>..] - ETA: 0s - loss: 0.6484 - acc: 0.7865"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 1s 25us/step - loss: 0.6464 - acc: 0.7868 - val_loss: 0.6596 - val_acc: 0.7899\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 14/20\n\r  128/50000 [..............................] - ETA: 1s - loss: 0.6266 - acc: 0.8047"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2304/50000 [>.............................] - ETA: 1s - loss: 0.6333 - acc: 0.7943"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4480/50000 [=>............................] - ETA: 1s - loss: 0.6210 - acc: 0.8002"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6784/50000 [===>..........................] - ETA: 1s - loss: 0.6330 - acc: 0.7941"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 9088/50000 [====>.........................] - ETA: 0s - loss: 0.6268 - acc: 0.7968"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r11392/50000 [=====>........................] - ETA: 0s - loss: 0.6261 - acc: 0.7944"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r13696/50000 [=======>......................] - ETA: 0s - loss: 0.6227 - acc: 0.7941"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r16000/50000 [========>.....................] - ETA: 0s - loss: 0.6260 - acc: 0.7933"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18304/50000 [=========>....................] - ETA: 0s - loss: 0.6283 - acc: 0.7935"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r20608/50000 [===========>..................] - ETA: 0s - loss: 0.6337 - acc: 0.7912"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r22912/50000 [============>.................] - ETA: 0s - loss: 0.6306 - acc: 0.7914"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r25216/50000 [==============>...............] - ETA: 0s - loss: 0.6313 - acc: 0.7907"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r27520/50000 [===============>..............] - ETA: 0s - loss: 0.6312 - acc: 0.7903"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r29824/50000 [================>.............] - ETA: 0s - loss: 0.6317 - acc: 0.7910"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r32128/50000 [==================>...........] - ETA: 0s - loss: 0.6334 - acc: 0.7908"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r34432/50000 [===================>..........] - ETA: 0s - loss: 0.6314 - acc: 0.7913"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r36736/50000 [=====================>........] - ETA: 0s - loss: 0.6310 - acc: 0.7909"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r39040/50000 [======================>.......] - ETA: 0s - loss: 0.6310 - acc: 0.7908"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r41344/50000 [=======================>......] - ETA: 0s - loss: 0.6300 - acc: 0.7911"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r43648/50000 [=========================>....] - ETA: 0s - loss: 0.6311 - acc: 0.7909"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r45952/50000 [==========================>...] - ETA: 0s - loss: 0.6301 - acc: 0.7911"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r48256/50000 [===========================>..] - ETA: 0s - loss: 0.6277 - acc: 0.7918"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 1s 25us/step - loss: 0.6262 - acc: 0.7923 - val_loss: 0.6427 - val_acc: 0.7859\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 15/20\n\r  128/50000 [..............................] - ETA: 1s - loss: 0.5212 - acc: 0.8516"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2304/50000 [>.............................] - ETA: 1s - loss: 0.6136 - acc: 0.7990"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4608/50000 [=>............................] - ETA: 1s - loss: 0.6217 - acc: 0.7930"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6912/50000 [===>..........................] - ETA: 0s - loss: 0.6108 - acc: 0.7985"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 9216/50000 [====>.........................] - ETA: 0s - loss: 0.6070 - acc: 0.7987"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r11520/50000 [=====>........................] - ETA: 0s - loss: 0.6095 - acc: 0.7984"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r13824/50000 [=======>......................] - ETA: 0s - loss: 0.6102 - acc: 0.7998"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r16128/50000 [========>.....................] - ETA: 0s - loss: 0.6098 - acc: 0.8011"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18432/50000 [==========>...................] - ETA: 0s - loss: 0.6119 - acc: 0.7999"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r20736/50000 [===========>..................] - ETA: 0s - loss: 0.6127 - acc: 0.7995"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r23040/50000 [============>.................] - ETA: 0s - loss: 0.6101 - acc: 0.7993"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r25344/50000 [==============>...............] - ETA: 0s - loss: 0.6114 - acc: 0.7977"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r27648/50000 [===============>..............] - ETA: 0s - loss: 0.6079 - acc: 0.7988"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r29952/50000 [================>.............] - ETA: 0s - loss: 0.6086 - acc: 0.7990"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r32256/50000 [==================>...........] - ETA: 0s - loss: 0.6088 - acc: 0.7995"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r34560/50000 [===================>..........] - ETA: 0s - loss: 0.6076 - acc: 0.7999"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r36864/50000 [=====================>........] - ETA: 0s - loss: 0.6096 - acc: 0.7993"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r39168/50000 [======================>.......] - ETA: 0s - loss: 0.6081 - acc: 0.7995"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r41472/50000 [=======================>......] - ETA: 0s - loss: 0.6070 - acc: 0.7989"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r43776/50000 [=========================>....] - ETA: 0s - loss: 0.6073 - acc: 0.7986"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r46080/50000 [==========================>...] - ETA: 0s - loss: 0.6104 - acc: 0.7976"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r48256/50000 [===========================>..] - ETA: 0s - loss: 0.6091 - acc: 0.7977"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 1s 25us/step - loss: 0.6100 - acc: 0.7969 - val_loss: 0.6364 - val_acc: 0.7900\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 16/20\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\r  128/50000 [..............................] - ETA: 1s - loss: 0.7641 - acc: 0.7891"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2176/50000 [>.............................] - ETA: 1s - loss: 0.6309 - acc: 0.8042"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4352/50000 [=>............................] - ETA: 1s - loss: 0.5981 - acc: 0.8047"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6656/50000 [==>...........................] - ETA: 1s - loss: 0.6001 - acc: 0.8002"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 8960/50000 [====>.........................] - ETA: 0s - loss: 0.6029 - acc: 0.7990"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r11264/50000 [=====>........................] - ETA: 0s - loss: 0.6081 - acc: 0.7984"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r13568/50000 [=======>......................] - ETA: 0s - loss: 0.6151 - acc: 0.7974"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r15872/50000 [========>.....................] - ETA: 0s - loss: 0.6081 - acc: 0.7984"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18176/50000 [=========>....................] - ETA: 0s - loss: 0.6070 - acc: 0.7992"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r20480/50000 [===========>..................] - ETA: 0s - loss: 0.6072 - acc: 0.7989"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r22784/50000 [============>.................] - ETA: 0s - loss: 0.6007 - acc: 0.8017"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r25088/50000 [==============>...............] - ETA: 0s - loss: 0.6003 - acc: 0.8025"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r27392/50000 [===============>..............] - ETA: 0s - loss: 0.5994 - acc: 0.8029"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r29696/50000 [================>.............] - ETA: 0s - loss: 0.5978 - acc: 0.8037"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r32000/50000 [==================>...........] - ETA: 0s - loss: 0.5947 - acc: 0.8046"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r34304/50000 [===================>..........] - ETA: 0s - loss: 0.5934 - acc: 0.8049"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r36608/50000 [====================>.........] - ETA: 0s - loss: 0.5964 - acc: 0.8042"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r38912/50000 [======================>.......] - ETA: 0s - loss: 0.5983 - acc: 0.8035"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r41216/50000 [=======================>......] - ETA: 0s - loss: 0.5979 - acc: 0.8037"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r43520/50000 [=========================>....] - ETA: 0s - loss: 0.5962 - acc: 0.8037"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r45824/50000 [==========================>...] - ETA: 0s - loss: 0.5933 - acc: 0.8039"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r48128/50000 [===========================>..] - ETA: 0s - loss: 0.5947 - acc: 0.8034"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 1s 25us/step - loss: 0.5942 - acc: 0.8031 - val_loss: 0.6122 - val_acc: 0.7958\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 17/20\n\r  128/50000 [..............................] - ETA: 1s - loss: 0.4844 - acc: 0.8203"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2304/50000 [>.............................] - ETA: 1s - loss: 0.5587 - acc: 0.8086"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4608/50000 [=>............................] - ETA: 1s - loss: 0.5874 - acc: 0.8006"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6912/50000 [===>..........................] - ETA: 0s - loss: 0.5893 - acc: 0.8024"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 9216/50000 [====>.........................] - ETA: 0s - loss: 0.5838 - acc: 0.8032"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r11392/50000 [=====>........................] - ETA: 0s - loss: 0.5862 - acc: 0.8006"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r13696/50000 [=======>......................] - ETA: 0s - loss: 0.5889 - acc: 0.7991"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r16000/50000 [========>.....................] - ETA: 0s - loss: 0.5877 - acc: 0.8001"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18304/50000 [=========>....................] - ETA: 0s - loss: 0.5829 - acc: 0.8018"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r20608/50000 [===========>..................] - ETA: 0s - loss: 0.5812 - acc: 0.8028"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r22912/50000 [============>.................] - ETA: 0s - loss: 0.5833 - acc: 0.8030"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r25216/50000 [==============>...............] - ETA: 0s - loss: 0.5806 - acc: 0.8038"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r27520/50000 [===============>..............] - ETA: 0s - loss: 0.5829 - acc: 0.8037"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r29824/50000 [================>.............] - ETA: 0s - loss: 0.5820 - acc: 0.8034"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r32128/50000 [==================>...........] - ETA: 0s - loss: 0.5810 - acc: 0.8040"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r34432/50000 [===================>..........] - ETA: 0s - loss: 0.5814 - acc: 0.8040"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r36736/50000 [=====================>........] - ETA: 0s - loss: 0.5824 - acc: 0.8046"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r39040/50000 [======================>.......] - ETA: 0s - loss: 0.5835 - acc: 0.8038"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r41344/50000 [=======================>......] - ETA: 0s - loss: 0.5828 - acc: 0.8035"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r43648/50000 [=========================>....] - ETA: 0s - loss: 0.5799 - acc: 0.8046"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r45952/50000 [==========================>...] - ETA: 0s - loss: 0.5788 - acc: 0.8051"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r48128/50000 [===========================>..] - ETA: 0s - loss: 0.5801 - acc: 0.8055"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 1s 25us/step - loss: 0.5790 - acc: 0.8058 - val_loss: 0.6147 - val_acc: 0.7994\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 18/20\n\r  128/50000 [..............................] - ETA: 1s - loss: 0.4826 - acc: 0.8203"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2304/50000 [>.............................] - ETA: 1s - loss: 0.5500 - acc: 0.8164"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4608/50000 [=>............................] - ETA: 1s - loss: 0.5663 - acc: 0.8116"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6784/50000 [===>..........................] - ETA: 1s - loss: 0.5697 - acc: 0.8085"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 8960/50000 [====>.........................] - ETA: 0s - loss: 0.5741 - acc: 0.8092"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r11264/50000 [=====>........................] - ETA: 0s - loss: 0.5725 - acc: 0.8125"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r13568/50000 [=======>......................] - ETA: 0s - loss: 0.5750 - acc: 0.8115"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r15872/50000 [========>.....................] - ETA: 0s - loss: 0.5823 - acc: 0.8108"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18176/50000 [=========>....................] - ETA: 0s - loss: 0.5764 - acc: 0.8106"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r20480/50000 [===========>..................] - ETA: 0s - loss: 0.5755 - acc: 0.8110"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r22784/50000 [============>.................] - ETA: 0s - loss: 0.5755 - acc: 0.8098"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r25088/50000 [==============>...............] - ETA: 0s - loss: 0.5744 - acc: 0.8087"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r27392/50000 [===============>..............] - ETA: 0s - loss: 0.5779 - acc: 0.8079"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r29696/50000 [================>.............] - ETA: 0s - loss: 0.5735 - acc: 0.8090"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r32000/50000 [==================>...........] - ETA: 0s - loss: 0.5737 - acc: 0.8096"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r34304/50000 [===================>..........] - ETA: 0s - loss: 0.5700 - acc: 0.8106"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r36608/50000 [====================>.........] - ETA: 0s - loss: 0.5683 - acc: 0.8113"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r38912/50000 [======================>.......] - ETA: 0s - loss: 0.5690 - acc: 0.8109"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r41216/50000 [=======================>......] - ETA: 0s - loss: 0.5677 - acc: 0.8109"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r43520/50000 [=========================>....] - ETA: 0s - loss: 0.5655 - acc: 0.8111"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r45824/50000 [==========================>...] - ETA: 0s - loss: 0.5652 - acc: 0.8113"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r48000/50000 [===========================>..] - ETA: 0s - loss: 0.5656 - acc: 0.8108"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 1s 25us/step - loss: 0.5648 - acc: 0.8108 - val_loss: 0.5929 - val_acc: 0.8075\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 19/20\n\r  128/50000 [..............................] - ETA: 1s - loss: 0.6954 - acc: 0.7812"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2304/50000 [>.............................] - ETA: 1s - loss: 0.5231 - acc: 0.8290"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4608/50000 [=>............................] - ETA: 1s - loss: 0.5529 - acc: 0.8186"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6912/50000 [===>..........................] - ETA: 0s - loss: 0.5520 - acc: 0.8180"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 9216/50000 [====>.........................] - ETA: 0s - loss: 0.5588 - acc: 0.8103"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r11520/50000 [=====>........................] - ETA: 0s - loss: 0.5535 - acc: 0.8095"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r13824/50000 [=======>......................] - ETA: 0s - loss: 0.5547 - acc: 0.8095"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r16128/50000 [========>.....................] - ETA: 0s - loss: 0.5607 - acc: 0.8087"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18432/50000 [==========>...................] - ETA: 0s - loss: 0.5591 - acc: 0.8098"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r20736/50000 [===========>..................] - ETA: 0s - loss: 0.5572 - acc: 0.8101"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r23040/50000 [============>.................] - ETA: 0s - loss: 0.5519 - acc: 0.8131"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r25344/50000 [==============>...............] - ETA: 0s - loss: 0.5505 - acc: 0.8140"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r27648/50000 [===============>..............] - ETA: 0s - loss: 0.5492 - acc: 0.8146"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r29952/50000 [================>.............] - ETA: 0s - loss: 0.5481 - acc: 0.8151"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r32256/50000 [==================>...........] - ETA: 0s - loss: 0.5517 - acc: 0.8152"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r34560/50000 [===================>..........] - ETA: 0s - loss: 0.5552 - acc: 0.8146"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r36864/50000 [=====================>........] - ETA: 0s - loss: 0.5545 - acc: 0.8150"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r39168/50000 [======================>.......] - ETA: 0s - loss: 0.5563 - acc: 0.8150"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r41472/50000 [=======================>......] - ETA: 0s - loss: 0.5562 - acc: 0.8146"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r43776/50000 [=========================>....] - ETA: 0s - loss: 0.5555 - acc: 0.8149"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r46080/50000 [==========================>...] - ETA: 0s - loss: 0.5560 - acc: 0.8151"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r48384/50000 [============================>.] - ETA: 0s - loss: 0.5552 - acc: 0.8156"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 1s 25us/step - loss: 0.5551 - acc: 0.8157 - val_loss: 0.5878 - val_acc: 0.8035\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 20/20\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\r  128/50000 [..............................] - ETA: 1s - loss: 0.4184 - acc: 0.8594"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 2304/50000 [>.............................] - ETA: 1s - loss: 0.5495 - acc: 0.8108"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 4608/50000 [=>............................] - ETA: 1s - loss: 0.5405 - acc: 0.8132"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 6912/50000 [===>..........................] - ETA: 0s - loss: 0.5343 - acc: 0.8167"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r 9216/50000 [====>.........................] - ETA: 0s - loss: 0.5415 - acc: 0.8143"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r11520/50000 [=====>........................] - ETA: 0s - loss: 0.5431 - acc: 0.8125"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r13824/50000 [=======>......................] - ETA: 0s - loss: 0.5448 - acc: 0.8128"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r16128/50000 [========>.....................] - ETA: 0s - loss: 0.5461 - acc: 0.8138"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18432/50000 [==========>...................] - ETA: 0s - loss: 0.5400 - acc: 0.8153"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r20736/50000 [===========>..................] - ETA: 0s - loss: 0.5411 - acc: 0.8153"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r23040/50000 [============>.................] - ETA: 0s - loss: 0.5427 - acc: 0.8149"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r25344/50000 [==============>...............] - ETA: 0s - loss: 0.5448 - acc: 0.8151"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r27648/50000 [===============>..............] - ETA: 0s - loss: 0.5436 - acc: 0.8152"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r29952/50000 [================>.............] - ETA: 0s - loss: 0.5427 - acc: 0.8157"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r32256/50000 [==================>...........] - ETA: 0s - loss: 0.5435 - acc: 0.8156"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r34560/50000 [===================>..........] - ETA: 0s - loss: 0.5419 - acc: 0.8164"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r36864/50000 [=====================>........] - ETA: 0s - loss: 0.5400 - acc: 0.8170"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r39168/50000 [======================>.......] - ETA: 0s - loss: 0.5385 - acc: 0.8173"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r41472/50000 [=======================>......] - ETA: 0s - loss: 0.5399 - acc: 0.8169"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r43776/50000 [=========================>....] - ETA: 0s - loss: 0.5403 - acc: 0.8173"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r46080/50000 [==========================>...] - ETA: 0s - loss: 0.5396 - acc: 0.8182"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r48256/50000 [===========================>..] - ETA: 0s - loss: 0.5410 - acc: 0.8181"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r50000/50000 [==============================] - 1s 25us/step - loss: 0.5418 - acc: 0.8179 - val_loss: 0.5950 - val_acc: 0.7985\n"
     ]
    }
   ],
   "source": [
    "batch_size = 128\n",
    "history = net.fit(X_train, y_train,\n",
    "                  batch_size=batch_size,\n",
    "                  epochs=20,\n",
    "                  verbose=1,\n",
    "                  validation_data=(X_val, y_val))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "From my experimentation, a network of many layer is needed. Adding an activation layer is also improving the performances. I chose tanh. sigmoid doesn't seem to work well. Deep network seems to be better at extracting all the features of the data set. \n"
     ]
    }
   ],
   "source": [
    "print(\"From my experimentation, a network of many layer is needed. Adding an activation layer is also improving the performances. I chose tanh. sigmoid doesn't seem to work well. Deep network seems to be better at extracting all the features of the data set. 82% accuracy is the maximum that I had.\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  }
 ],
 "metadata": {
  "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": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}