training-2.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pszt import net\n",
    "import numpy as np\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Sieć pierwsza bag-of-words"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Sieć druga - TFIDF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = np.load('x_tfidf.npy')\n",
    "y_true = np.load('y_tfidf.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create a random order to shuffle the dataset.\n",
    "s = np.arange(x.shape[0])\n",
    "np.random.shuffle(s)\n",
    "\n",
    "# Split dataset into k folds.\n",
    "x_folds = np.array_split(x[s], 5)\n",
    "y_folds = np.array_split(y_true[s], 5)\n",
    "\n",
    "x_train = np.concatenate(np.delete(x_folds, 1))\n",
    "y_train = np.concatenate(np.delete(y_folds, 1))\n",
    "\n",
    "x_test = x_folds[1]\n",
    "y_test = y_folds[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1. Layer - input_dim: 8257, output_dim: 20, activation: sigmoid\n",
      "2. Layer - input_dim: 20, output_dim: 3, activation: softmax\n"
     ]
    }
   ],
   "source": [
    "mlp = net.MLP()\n",
    "\n",
    "mlp.add_layer(input_dim=x_train.shape[1], output_dim=20, activation=\"sigmoid\")\n",
    "\n",
    "mlp.add_layer(input_dim=20, output_dim=3, activation=\"softmax\")\n",
    "mlp.print_layers()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "mlp.init_layers()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 1, loss_train: 2813.33, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.22, accuracy_test: 55.2%\n",
      "Epoch: 2, loss_train: 2812.24, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.44, accuracy_test: 55.2%\n",
      "Epoch: 3, loss_train: 2812.15, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.44, accuracy_test: 55.2%\n",
      "Epoch: 4, loss_train: 2812.11, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.44, accuracy_test: 55.2%\n",
      "Epoch: 5, loss_train: 2812.06, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.43, accuracy_test: 55.2%\n",
      "Epoch: 6, loss_train: 2812.02, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.43, accuracy_test: 55.2%\n",
      "Epoch: 7, loss_train: 2811.97, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.42, accuracy_test: 55.2%\n",
      "Epoch: 8, loss_train: 2811.93, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.41, accuracy_test: 55.2%\n",
      "Epoch: 9, loss_train: 2811.89, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.41, accuracy_test: 55.2%\n",
      "Epoch: 10, loss_train: 2811.84, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.40, accuracy_test: 55.2%\n",
      "Epoch: 11, loss_train: 2811.80, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.39, accuracy_test: 55.2%\n",
      "Epoch: 12, loss_train: 2811.75, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.39, accuracy_test: 55.2%\n",
      "Epoch: 13, loss_train: 2811.71, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.38, accuracy_test: 55.2%\n",
      "Epoch: 14, loss_train: 2811.66, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.37, accuracy_test: 55.2%\n",
      "Epoch: 15, loss_train: 2811.62, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.37, accuracy_test: 55.2%\n",
      "Epoch: 16, loss_train: 2811.57, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.36, accuracy_test: 55.2%\n",
      "Epoch: 17, loss_train: 2811.53, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.35, accuracy_test: 55.2%\n",
      "Epoch: 18, loss_train: 2811.48, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.34, accuracy_test: 55.2%\n",
      "Epoch: 19, loss_train: 2811.43, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.34, accuracy_test: 55.2%\n",
      "Epoch: 20, loss_train: 2811.39, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.33, accuracy_test: 55.2%\n",
      "Epoch: 21, loss_train: 2811.34, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.32, accuracy_test: 55.2%\n",
      "Epoch: 22, loss_train: 2811.29, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.32, accuracy_test: 55.2%\n",
      "Epoch: 23, loss_train: 2811.24, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.31, accuracy_test: 55.2%\n",
      "Epoch: 24, loss_train: 2811.20, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.30, accuracy_test: 55.2%\n",
      "Epoch: 25, loss_train: 2811.15, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.29, accuracy_test: 55.2%\n",
      "Epoch: 26, loss_train: 2811.10, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.29, accuracy_test: 55.2%\n",
      "Epoch: 27, loss_train: 2811.05, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.28, accuracy_test: 55.2%\n",
      "Epoch: 28, loss_train: 2811.00, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.27, accuracy_test: 55.2%\n",
      "Epoch: 29, loss_train: 2810.95, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.26, accuracy_test: 55.2%\n",
      "Epoch: 30, loss_train: 2810.90, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.26, accuracy_test: 55.2%\n",
      "Epoch: 31, loss_train: 2810.85, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.25, accuracy_test: 55.2%\n",
      "Epoch: 32, loss_train: 2810.79, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.24, accuracy_test: 55.2%\n",
      "Epoch: 33, loss_train: 2810.74, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.23, accuracy_test: 55.2%\n",
      "Epoch: 34, loss_train: 2810.69, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.22, accuracy_test: 55.2%\n",
      "Epoch: 35, loss_train: 2810.63, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.21, accuracy_test: 55.2%\n",
      "Epoch: 36, loss_train: 2810.58, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.21, accuracy_test: 55.2%\n",
      "Epoch: 37, loss_train: 2810.52, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.20, accuracy_test: 55.2%\n",
      "Epoch: 38, loss_train: 2810.47, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.19, accuracy_test: 55.2%\n",
      "Epoch: 39, loss_train: 2810.41, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.18, accuracy_test: 55.2%\n",
      "Epoch: 40, loss_train: 2810.35, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.17, accuracy_test: 55.2%\n",
      "Epoch: 41, loss_train: 2810.30, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.16, accuracy_test: 55.2%\n",
      "Epoch: 42, loss_train: 2810.24, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.15, accuracy_test: 55.2%\n",
      "Epoch: 43, loss_train: 2810.18, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.14, accuracy_test: 55.2%\n",
      "Epoch: 44, loss_train: 2810.12, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.13, accuracy_test: 55.2%\n",
      "Epoch: 45, loss_train: 2810.05, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.12, accuracy_test: 55.2%\n",
      "Epoch: 46, loss_train: 2809.99, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.11, accuracy_test: 55.2%\n",
      "Epoch: 47, loss_train: 2809.93, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.10, accuracy_test: 55.2%\n",
      "Epoch: 48, loss_train: 2809.86, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.09, accuracy_test: 55.2%\n",
      "Epoch: 49, loss_train: 2809.80, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.08, accuracy_test: 55.2%\n",
      "Epoch: 50, loss_train: 2809.73, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.07, accuracy_test: 55.2%\n",
      "Epoch: 51, loss_train: 2809.66, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.06, accuracy_test: 55.2%\n",
      "Epoch: 52, loss_train: 2809.59, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.04, accuracy_test: 55.2%\n",
      "Epoch: 53, loss_train: 2809.52, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.03, accuracy_test: 55.2%\n",
      "Epoch: 54, loss_train: 2809.45, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.02, accuracy_test: 55.2%\n",
      "Epoch: 55, loss_train: 2809.38, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.01, accuracy_test: 55.2%\n",
      "Epoch: 56, loss_train: 2809.30, accuracy_train: 57.2%\n",
      "\t  loss_test: 726.00, accuracy_test: 55.2%\n",
      "Epoch: 57, loss_train: 2809.23, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.98, accuracy_test: 55.2%\n",
      "Epoch: 58, loss_train: 2809.15, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.97, accuracy_test: 55.2%\n",
      "Epoch: 59, loss_train: 2809.07, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.96, accuracy_test: 55.2%\n",
      "Epoch: 60, loss_train: 2808.99, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.94, accuracy_test: 55.2%\n",
      "Epoch: 61, loss_train: 2808.91, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.93, accuracy_test: 55.2%\n",
      "Epoch: 62, loss_train: 2808.83, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.91, accuracy_test: 55.2%\n",
      "Epoch: 63, loss_train: 2808.74, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.90, accuracy_test: 55.2%\n",
      "Epoch: 64, loss_train: 2808.66, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.88, accuracy_test: 55.2%\n",
      "Epoch: 65, loss_train: 2808.57, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.87, accuracy_test: 55.2%\n",
      "Epoch: 66, loss_train: 2808.48, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.85, accuracy_test: 55.2%\n",
      "Epoch: 67, loss_train: 2808.39, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.84, accuracy_test: 55.2%\n",
      "Epoch: 68, loss_train: 2808.29, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.82, accuracy_test: 55.2%\n",
      "Epoch: 69, loss_train: 2808.20, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.80, accuracy_test: 55.2%\n",
      "Epoch: 70, loss_train: 2808.10, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.79, accuracy_test: 55.2%\n",
      "Epoch: 71, loss_train: 2808.00, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.77, accuracy_test: 55.2%\n",
      "Epoch: 72, loss_train: 2807.90, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.75, accuracy_test: 55.2%\n",
      "Epoch: 73, loss_train: 2807.79, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.73, accuracy_test: 55.2%\n",
      "Epoch: 74, loss_train: 2807.68, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.71, accuracy_test: 55.2%\n",
      "Epoch: 75, loss_train: 2807.57, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.69, accuracy_test: 55.2%\n",
      "Epoch: 76, loss_train: 2807.46, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.67, accuracy_test: 55.2%\n",
      "Epoch: 77, loss_train: 2807.35, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.65, accuracy_test: 55.2%\n",
      "Epoch: 78, loss_train: 2807.23, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.63, accuracy_test: 55.2%\n",
      "Epoch: 79, loss_train: 2807.11, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.61, accuracy_test: 55.2%\n",
      "Epoch: 80, loss_train: 2806.98, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.58, accuracy_test: 55.2%\n",
      "Epoch: 81, loss_train: 2806.86, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.56, accuracy_test: 55.2%\n",
      "Epoch: 82, loss_train: 2806.72, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.54, accuracy_test: 55.2%\n",
      "Epoch: 83, loss_train: 2806.59, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.51, accuracy_test: 55.2%\n",
      "Epoch: 84, loss_train: 2806.45, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.48, accuracy_test: 55.2%\n",
      "Epoch: 85, loss_train: 2806.31, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.46, accuracy_test: 55.2%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 86, loss_train: 2806.17, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.43, accuracy_test: 55.2%\n",
      "Epoch: 87, loss_train: 2806.02, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.40, accuracy_test: 55.2%\n",
      "Epoch: 88, loss_train: 2805.86, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.37, accuracy_test: 55.2%\n",
      "Epoch: 89, loss_train: 2805.71, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.34, accuracy_test: 55.2%\n",
      "Epoch: 90, loss_train: 2805.54, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.31, accuracy_test: 55.2%\n",
      "Epoch: 91, loss_train: 2805.38, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.28, accuracy_test: 55.2%\n",
      "Epoch: 92, loss_train: 2805.21, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.25, accuracy_test: 55.2%\n",
      "Epoch: 93, loss_train: 2805.03, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.21, accuracy_test: 55.2%\n",
      "Epoch: 94, loss_train: 2804.85, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.18, accuracy_test: 55.2%\n",
      "Epoch: 95, loss_train: 2804.66, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.14, accuracy_test: 55.2%\n",
      "Epoch: 96, loss_train: 2804.47, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.11, accuracy_test: 55.2%\n",
      "Epoch: 97, loss_train: 2804.27, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.07, accuracy_test: 55.2%\n",
      "Epoch: 98, loss_train: 2804.07, accuracy_train: 57.2%\n",
      "\t  loss_test: 725.03, accuracy_test: 55.2%\n",
      "Epoch: 99, loss_train: 2803.86, accuracy_train: 57.2%\n",
      "\t  loss_test: 724.99, accuracy_test: 55.2%\n",
      "Epoch: 100, loss_train: 2803.64, accuracy_train: 57.2%\n",
      "\t  loss_test: 724.94, accuracy_test: 55.2%\n",
      "Epoch: 101, loss_train: 2803.41, accuracy_train: 57.2%\n",
      "\t  loss_test: 724.90, accuracy_test: 55.2%\n",
      "Epoch: 102, loss_train: 2803.18, accuracy_train: 57.2%\n",
      "\t  loss_test: 724.85, accuracy_test: 55.2%\n",
      "Epoch: 103, loss_train: 2802.94, accuracy_train: 57.2%\n",
      "\t  loss_test: 724.81, accuracy_test: 55.2%\n",
      "Epoch: 104, loss_train: 2802.69, accuracy_train: 57.2%\n",
      "\t  loss_test: 724.76, accuracy_test: 55.2%\n",
      "Epoch: 105, loss_train: 2802.44, accuracy_train: 57.2%\n",
      "\t  loss_test: 724.71, accuracy_test: 55.2%\n",
      "Epoch: 106, loss_train: 2802.18, accuracy_train: 57.2%\n",
      "\t  loss_test: 724.65, accuracy_test: 55.2%\n",
      "Epoch: 107, loss_train: 2801.90, accuracy_train: 57.2%\n",
      "\t  loss_test: 724.60, accuracy_test: 55.2%\n",
      "Epoch: 108, loss_train: 2801.62, accuracy_train: 57.2%\n",
      "\t  loss_test: 724.54, accuracy_test: 55.2%\n",
      "Epoch: 109, loss_train: 2801.33, accuracy_train: 57.2%\n",
      "\t  loss_test: 724.48, accuracy_test: 55.2%\n",
      "Epoch: 110, loss_train: 2801.03, accuracy_train: 57.2%\n",
      "\t  loss_test: 724.42, accuracy_test: 55.2%\n",
      "Epoch: 111, loss_train: 2800.72, accuracy_train: 57.2%\n",
      "\t  loss_test: 724.36, accuracy_test: 55.2%\n",
      "Epoch: 112, loss_train: 2800.39, accuracy_train: 57.2%\n",
      "\t  loss_test: 724.29, accuracy_test: 55.2%\n",
      "Epoch: 113, loss_train: 2800.06, accuracy_train: 57.2%\n",
      "\t  loss_test: 724.23, accuracy_test: 55.2%\n",
      "Epoch: 114, loss_train: 2799.71, accuracy_train: 57.2%\n",
      "\t  loss_test: 724.16, accuracy_test: 55.2%\n",
      "Epoch: 115, loss_train: 2799.35, accuracy_train: 57.2%\n",
      "\t  loss_test: 724.08, accuracy_test: 55.2%\n",
      "Epoch: 116, loss_train: 2798.98, accuracy_train: 57.2%\n",
      "\t  loss_test: 724.01, accuracy_test: 55.2%\n",
      "Epoch: 117, loss_train: 2798.60, accuracy_train: 57.2%\n",
      "\t  loss_test: 723.93, accuracy_test: 55.2%\n",
      "Epoch: 118, loss_train: 2798.20, accuracy_train: 57.2%\n",
      "\t  loss_test: 723.85, accuracy_test: 55.2%\n",
      "Epoch: 119, loss_train: 2797.78, accuracy_train: 57.2%\n",
      "\t  loss_test: 723.76, accuracy_test: 55.2%\n",
      "Epoch: 120, loss_train: 2797.35, accuracy_train: 57.2%\n",
      "\t  loss_test: 723.67, accuracy_test: 55.2%\n",
      "Epoch: 121, loss_train: 2796.91, accuracy_train: 57.2%\n",
      "\t  loss_test: 723.58, accuracy_test: 55.2%\n",
      "Epoch: 122, loss_train: 2796.44, accuracy_train: 57.2%\n",
      "\t  loss_test: 723.48, accuracy_test: 55.2%\n",
      "Epoch: 123, loss_train: 2795.96, accuracy_train: 57.2%\n",
      "\t  loss_test: 723.38, accuracy_test: 55.2%\n",
      "Epoch: 124, loss_train: 2795.46, accuracy_train: 57.2%\n",
      "\t  loss_test: 723.28, accuracy_test: 55.2%\n",
      "Epoch: 125, loss_train: 2794.94, accuracy_train: 57.2%\n",
      "\t  loss_test: 723.17, accuracy_test: 55.2%\n",
      "Epoch: 126, loss_train: 2794.41, accuracy_train: 57.2%\n",
      "\t  loss_test: 723.06, accuracy_test: 55.2%\n",
      "Epoch: 127, loss_train: 2793.85, accuracy_train: 57.2%\n",
      "\t  loss_test: 722.94, accuracy_test: 55.2%\n",
      "Epoch: 128, loss_train: 2793.26, accuracy_train: 57.2%\n",
      "\t  loss_test: 722.82, accuracy_test: 55.2%\n",
      "Epoch: 129, loss_train: 2792.66, accuracy_train: 57.2%\n",
      "\t  loss_test: 722.69, accuracy_test: 55.2%\n",
      "Epoch: 130, loss_train: 2792.03, accuracy_train: 57.2%\n",
      "\t  loss_test: 722.56, accuracy_test: 55.2%\n",
      "Epoch: 131, loss_train: 2791.37, accuracy_train: 57.2%\n",
      "\t  loss_test: 722.42, accuracy_test: 55.2%\n",
      "Epoch: 132, loss_train: 2790.69, accuracy_train: 57.2%\n",
      "\t  loss_test: 722.28, accuracy_test: 55.2%\n",
      "Epoch: 133, loss_train: 2789.98, accuracy_train: 57.2%\n",
      "\t  loss_test: 722.13, accuracy_test: 55.2%\n",
      "Epoch: 134, loss_train: 2789.25, accuracy_train: 57.2%\n",
      "\t  loss_test: 721.97, accuracy_test: 55.2%\n",
      "Epoch: 135, loss_train: 2788.48, accuracy_train: 57.2%\n",
      "\t  loss_test: 721.81, accuracy_test: 55.2%\n",
      "Epoch: 136, loss_train: 2787.68, accuracy_train: 57.2%\n",
      "\t  loss_test: 721.64, accuracy_test: 55.2%\n",
      "Epoch: 137, loss_train: 2786.85, accuracy_train: 57.2%\n",
      "\t  loss_test: 721.47, accuracy_test: 55.2%\n",
      "Epoch: 138, loss_train: 2785.98, accuracy_train: 57.2%\n",
      "\t  loss_test: 721.28, accuracy_test: 55.2%\n",
      "Epoch: 139, loss_train: 2785.08, accuracy_train: 57.2%\n",
      "\t  loss_test: 721.09, accuracy_test: 55.2%\n",
      "Epoch: 140, loss_train: 2784.14, accuracy_train: 57.2%\n",
      "\t  loss_test: 720.89, accuracy_test: 55.2%\n",
      "Epoch: 141, loss_train: 2783.16, accuracy_train: 57.2%\n",
      "\t  loss_test: 720.68, accuracy_test: 55.2%\n",
      "Epoch: 142, loss_train: 2782.14, accuracy_train: 57.2%\n",
      "\t  loss_test: 720.46, accuracy_test: 55.2%\n",
      "Epoch: 143, loss_train: 2781.07, accuracy_train: 57.2%\n",
      "\t  loss_test: 720.24, accuracy_test: 55.2%\n",
      "Epoch: 144, loss_train: 2779.96, accuracy_train: 57.2%\n",
      "\t  loss_test: 720.00, accuracy_test: 55.2%\n",
      "Epoch: 145, loss_train: 2778.81, accuracy_train: 57.2%\n",
      "\t  loss_test: 719.75, accuracy_test: 55.2%\n",
      "Epoch: 146, loss_train: 2777.60, accuracy_train: 57.2%\n",
      "\t  loss_test: 719.49, accuracy_test: 55.2%\n",
      "Epoch: 147, loss_train: 2776.35, accuracy_train: 57.2%\n",
      "\t  loss_test: 719.22, accuracy_test: 55.2%\n",
      "Epoch: 148, loss_train: 2775.04, accuracy_train: 57.2%\n",
      "\t  loss_test: 718.94, accuracy_test: 55.2%\n",
      "Epoch: 149, loss_train: 2773.67, accuracy_train: 57.2%\n",
      "\t  loss_test: 718.65, accuracy_test: 55.2%\n",
      "Epoch: 150, loss_train: 2772.25, accuracy_train: 57.2%\n",
      "\t  loss_test: 718.34, accuracy_test: 55.2%\n",
      "Epoch: 151, loss_train: 2770.76, accuracy_train: 57.2%\n",
      "\t  loss_test: 718.02, accuracy_test: 55.2%\n",
      "Epoch: 152, loss_train: 2769.21, accuracy_train: 57.2%\n",
      "\t  loss_test: 717.69, accuracy_test: 55.2%\n",
      "Epoch: 153, loss_train: 2767.60, accuracy_train: 57.2%\n",
      "\t  loss_test: 717.34, accuracy_test: 55.2%\n",
      "Epoch: 154, loss_train: 2765.92, accuracy_train: 57.2%\n",
      "\t  loss_test: 716.98, accuracy_test: 55.2%\n",
      "Epoch: 155, loss_train: 2764.16, accuracy_train: 57.2%\n",
      "\t  loss_test: 716.60, accuracy_test: 55.2%\n",
      "Epoch: 156, loss_train: 2762.33, accuracy_train: 57.2%\n",
      "\t  loss_test: 716.20, accuracy_test: 55.2%\n",
      "Epoch: 157, loss_train: 2760.43, accuracy_train: 57.2%\n",
      "\t  loss_test: 715.79, accuracy_test: 55.2%\n",
      "Epoch: 158, loss_train: 2758.44, accuracy_train: 57.2%\n",
      "\t  loss_test: 715.36, accuracy_test: 55.2%\n",
      "Epoch: 159, loss_train: 2756.37, accuracy_train: 57.2%\n",
      "\t  loss_test: 714.91, accuracy_test: 55.2%\n",
      "Epoch: 160, loss_train: 2754.21, accuracy_train: 57.2%\n",
      "\t  loss_test: 714.44, accuracy_test: 55.2%\n",
      "Epoch: 161, loss_train: 2751.96, accuracy_train: 57.2%\n",
      "\t  loss_test: 713.95, accuracy_test: 55.2%\n",
      "Epoch: 162, loss_train: 2749.61, accuracy_train: 57.2%\n",
      "\t  loss_test: 713.44, accuracy_test: 55.2%\n",
      "Epoch: 163, loss_train: 2747.17, accuracy_train: 57.2%\n",
      "\t  loss_test: 712.91, accuracy_test: 55.2%\n",
      "Epoch: 164, loss_train: 2744.62, accuracy_train: 57.2%\n",
      "\t  loss_test: 712.36, accuracy_test: 55.2%\n",
      "Epoch: 165, loss_train: 2741.97, accuracy_train: 57.2%\n",
      "\t  loss_test: 711.78, accuracy_test: 55.2%\n",
      "Epoch: 166, loss_train: 2739.21, accuracy_train: 57.2%\n",
      "\t  loss_test: 711.18, accuracy_test: 55.2%\n",
      "Epoch: 167, loss_train: 2736.34, accuracy_train: 57.2%\n",
      "\t  loss_test: 710.56, accuracy_test: 55.2%\n",
      "Epoch: 168, loss_train: 2733.35, accuracy_train: 57.2%\n",
      "\t  loss_test: 709.91, accuracy_test: 55.2%\n",
      "Epoch: 169, loss_train: 2730.23, accuracy_train: 57.2%\n",
      "\t  loss_test: 709.23, accuracy_test: 55.2%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 170, loss_train: 2727.00, accuracy_train: 57.2%\n",
      "\t  loss_test: 708.53, accuracy_test: 55.2%\n",
      "Epoch: 171, loss_train: 2723.63, accuracy_train: 57.2%\n",
      "\t  loss_test: 707.80, accuracy_test: 55.2%\n",
      "Epoch: 172, loss_train: 2720.13, accuracy_train: 57.2%\n",
      "\t  loss_test: 707.03, accuracy_test: 55.2%\n",
      "Epoch: 173, loss_train: 2716.50, accuracy_train: 57.2%\n",
      "\t  loss_test: 706.24, accuracy_test: 55.2%\n",
      "Epoch: 174, loss_train: 2712.72, accuracy_train: 57.2%\n",
      "\t  loss_test: 705.42, accuracy_test: 55.2%\n",
      "Epoch: 175, loss_train: 2708.80, accuracy_train: 57.2%\n",
      "\t  loss_test: 704.57, accuracy_test: 55.2%\n",
      "Epoch: 176, loss_train: 2704.73, accuracy_train: 57.2%\n",
      "\t  loss_test: 703.68, accuracy_test: 55.3%\n",
      "Epoch: 177, loss_train: 2700.51, accuracy_train: 57.3%\n",
      "\t  loss_test: 702.76, accuracy_test: 55.5%\n",
      "Epoch: 178, loss_train: 2696.14, accuracy_train: 57.3%\n",
      "\t  loss_test: 701.81, accuracy_test: 55.5%\n",
      "Epoch: 179, loss_train: 2691.60, accuracy_train: 57.4%\n",
      "\t  loss_test: 700.82, accuracy_test: 55.5%\n",
      "Epoch: 180, loss_train: 2686.90, accuracy_train: 57.4%\n",
      "\t  loss_test: 699.79, accuracy_test: 55.5%\n",
      "Epoch: 181, loss_train: 2682.04, accuracy_train: 57.5%\n",
      "\t  loss_test: 698.73, accuracy_test: 55.5%\n",
      "Epoch: 182, loss_train: 2677.00, accuracy_train: 57.5%\n",
      "\t  loss_test: 697.64, accuracy_test: 55.5%\n",
      "Epoch: 183, loss_train: 2671.80, accuracy_train: 57.6%\n",
      "\t  loss_test: 696.50, accuracy_test: 55.5%\n",
      "Epoch: 184, loss_train: 2666.42, accuracy_train: 57.7%\n",
      "\t  loss_test: 695.33, accuracy_test: 55.5%\n",
      "Epoch: 185, loss_train: 2660.86, accuracy_train: 57.8%\n",
      "\t  loss_test: 694.12, accuracy_test: 55.6%\n",
      "Epoch: 186, loss_train: 2655.13, accuracy_train: 57.9%\n",
      "\t  loss_test: 692.88, accuracy_test: 55.8%\n",
      "Epoch: 187, loss_train: 2649.22, accuracy_train: 58.0%\n",
      "\t  loss_test: 691.59, accuracy_test: 55.8%\n",
      "Epoch: 188, loss_train: 2643.13, accuracy_train: 58.0%\n",
      "\t  loss_test: 690.27, accuracy_test: 55.8%\n",
      "Epoch: 189, loss_train: 2636.86, accuracy_train: 58.3%\n",
      "\t  loss_test: 688.90, accuracy_test: 55.8%\n",
      "Epoch: 190, loss_train: 2630.40, accuracy_train: 58.5%\n",
      "\t  loss_test: 687.50, accuracy_test: 56.1%\n",
      "Epoch: 191, loss_train: 2623.77, accuracy_train: 58.7%\n",
      "\t  loss_test: 686.06, accuracy_test: 56.4%\n",
      "Epoch: 192, loss_train: 2616.95, accuracy_train: 59.0%\n",
      "\t  loss_test: 684.58, accuracy_test: 56.6%\n",
      "Epoch: 193, loss_train: 2609.96, accuracy_train: 59.2%\n",
      "\t  loss_test: 683.06, accuracy_test: 57.0%\n",
      "Epoch: 194, loss_train: 2602.78, accuracy_train: 59.4%\n",
      "\t  loss_test: 681.50, accuracy_test: 57.4%\n",
      "Epoch: 195, loss_train: 2595.43, accuracy_train: 59.6%\n",
      "\t  loss_test: 679.91, accuracy_test: 58.1%\n",
      "Epoch: 196, loss_train: 2587.91, accuracy_train: 59.8%\n",
      "\t  loss_test: 678.28, accuracy_test: 58.2%\n",
      "Epoch: 197, loss_train: 2580.22, accuracy_train: 60.3%\n",
      "\t  loss_test: 676.62, accuracy_test: 58.6%\n",
      "Epoch: 198, loss_train: 2572.37, accuracy_train: 60.8%\n",
      "\t  loss_test: 674.92, accuracy_test: 58.9%\n",
      "Epoch: 199, loss_train: 2564.35, accuracy_train: 61.1%\n",
      "\t  loss_test: 673.19, accuracy_test: 59.1%\n",
      "Epoch: 200, loss_train: 2556.18, accuracy_train: 61.4%\n",
      "\t  loss_test: 671.42, accuracy_test: 59.7%\n",
      "Epoch: 201, loss_train: 2547.85, accuracy_train: 61.8%\n",
      "\t  loss_test: 669.63, accuracy_test: 59.8%\n",
      "Epoch: 202, loss_train: 2539.38, accuracy_train: 62.4%\n",
      "\t  loss_test: 667.81, accuracy_test: 60.4%\n",
      "Epoch: 203, loss_train: 2530.78, accuracy_train: 62.8%\n",
      "\t  loss_test: 665.95, accuracy_test: 60.7%\n",
      "Epoch: 204, loss_train: 2522.04, accuracy_train: 63.1%\n",
      "\t  loss_test: 664.08, accuracy_test: 61.2%\n",
      "Epoch: 205, loss_train: 2513.18, accuracy_train: 63.4%\n",
      "\t  loss_test: 662.18, accuracy_test: 61.4%\n",
      "Epoch: 206, loss_train: 2504.20, accuracy_train: 64.0%\n",
      "\t  loss_test: 660.25, accuracy_test: 61.8%\n",
      "Epoch: 207, loss_train: 2495.11, accuracy_train: 64.8%\n",
      "\t  loss_test: 658.31, accuracy_test: 62.0%\n",
      "Epoch: 208, loss_train: 2485.93, accuracy_train: 65.2%\n",
      "\t  loss_test: 656.35, accuracy_test: 62.5%\n",
      "Epoch: 209, loss_train: 2476.65, accuracy_train: 66.0%\n",
      "\t  loss_test: 654.37, accuracy_test: 62.7%\n",
      "Epoch: 210, loss_train: 2467.29, accuracy_train: 66.4%\n",
      "\t  loss_test: 652.38, accuracy_test: 63.5%\n",
      "Epoch: 211, loss_train: 2457.87, accuracy_train: 67.0%\n",
      "\t  loss_test: 650.38, accuracy_test: 63.2%\n",
      "Epoch: 212, loss_train: 2448.38, accuracy_train: 67.3%\n",
      "\t  loss_test: 648.37, accuracy_test: 63.5%\n",
      "Epoch: 213, loss_train: 2438.83, accuracy_train: 67.6%\n",
      "\t  loss_test: 646.35, accuracy_test: 63.5%\n",
      "Epoch: 214, loss_train: 2429.24, accuracy_train: 68.1%\n",
      "\t  loss_test: 644.32, accuracy_test: 64.1%\n",
      "Epoch: 215, loss_train: 2419.62, accuracy_train: 68.7%\n",
      "\t  loss_test: 642.30, accuracy_test: 64.4%\n",
      "Epoch: 216, loss_train: 2409.97, accuracy_train: 69.4%\n",
      "\t  loss_test: 640.27, accuracy_test: 65.0%\n",
      "Epoch: 217, loss_train: 2400.31, accuracy_train: 69.8%\n",
      "\t  loss_test: 638.25, accuracy_test: 65.0%\n",
      "Epoch: 218, loss_train: 2390.64, accuracy_train: 70.1%\n",
      "\t  loss_test: 636.23, accuracy_test: 65.6%\n",
      "Epoch: 219, loss_train: 2380.98, accuracy_train: 70.6%\n",
      "\t  loss_test: 634.22, accuracy_test: 66.0%\n",
      "Epoch: 220, loss_train: 2371.32, accuracy_train: 70.8%\n",
      "\t  loss_test: 632.21, accuracy_test: 66.4%\n",
      "Epoch: 221, loss_train: 2361.69, accuracy_train: 70.9%\n",
      "\t  loss_test: 630.22, accuracy_test: 67.1%\n",
      "Epoch: 222, loss_train: 2352.08, accuracy_train: 71.4%\n",
      "\t  loss_test: 628.24, accuracy_test: 67.3%\n",
      "Epoch: 223, loss_train: 2342.50, accuracy_train: 71.8%\n",
      "\t  loss_test: 626.27, accuracy_test: 67.7%\n",
      "Epoch: 224, loss_train: 2332.97, accuracy_train: 72.2%\n",
      "\t  loss_test: 624.32, accuracy_test: 67.9%\n",
      "Epoch: 225, loss_train: 2323.48, accuracy_train: 72.3%\n",
      "\t  loss_test: 622.39, accuracy_test: 68.3%\n",
      "Epoch: 226, loss_train: 2314.05, accuracy_train: 72.6%\n",
      "\t  loss_test: 620.47, accuracy_test: 68.6%\n",
      "Epoch: 227, loss_train: 2304.68, accuracy_train: 73.0%\n",
      "\t  loss_test: 618.58, accuracy_test: 68.9%\n",
      "Epoch: 228, loss_train: 2295.38, accuracy_train: 73.3%\n",
      "\t  loss_test: 616.70, accuracy_test: 69.1%\n",
      "Epoch: 229, loss_train: 2286.15, accuracy_train: 73.7%\n",
      "\t  loss_test: 614.85, accuracy_test: 69.6%\n",
      "Epoch: 230, loss_train: 2276.99, accuracy_train: 74.0%\n",
      "\t  loss_test: 613.03, accuracy_test: 70.3%\n",
      "Epoch: 231, loss_train: 2267.91, accuracy_train: 74.3%\n",
      "\t  loss_test: 611.23, accuracy_test: 70.4%\n",
      "Epoch: 232, loss_train: 2258.91, accuracy_train: 74.6%\n",
      "\t  loss_test: 609.45, accuracy_test: 70.6%\n",
      "Epoch: 233, loss_train: 2250.00, accuracy_train: 74.6%\n",
      "\t  loss_test: 607.70, accuracy_test: 71.0%\n",
      "Epoch: 234, loss_train: 2241.17, accuracy_train: 74.9%\n",
      "\t  loss_test: 605.98, accuracy_test: 70.7%\n",
      "Epoch: 235, loss_train: 2232.44, accuracy_train: 75.2%\n",
      "\t  loss_test: 604.28, accuracy_test: 70.7%\n",
      "Epoch: 236, loss_train: 2223.80, accuracy_train: 75.3%\n",
      "\t  loss_test: 602.62, accuracy_test: 71.0%\n",
      "Epoch: 237, loss_train: 2215.25, accuracy_train: 75.6%\n",
      "\t  loss_test: 600.98, accuracy_test: 71.5%\n",
      "Epoch: 238, loss_train: 2206.80, accuracy_train: 75.7%\n",
      "\t  loss_test: 599.37, accuracy_test: 71.9%\n",
      "Epoch: 239, loss_train: 2198.44, accuracy_train: 75.9%\n",
      "\t  loss_test: 597.79, accuracy_test: 71.9%\n",
      "Epoch: 240, loss_train: 2190.18, accuracy_train: 76.0%\n",
      "\t  loss_test: 596.23, accuracy_test: 72.0%\n",
      "Epoch: 241, loss_train: 2182.01, accuracy_train: 76.1%\n",
      "\t  loss_test: 594.71, accuracy_test: 72.0%\n",
      "Epoch: 242, loss_train: 2173.94, accuracy_train: 76.3%\n",
      "\t  loss_test: 593.21, accuracy_test: 71.9%\n",
      "Epoch: 243, loss_train: 2165.96, accuracy_train: 76.4%\n",
      "\t  loss_test: 591.74, accuracy_test: 72.4%\n",
      "Epoch: 244, loss_train: 2158.08, accuracy_train: 76.7%\n",
      "\t  loss_test: 590.30, accuracy_test: 72.4%\n",
      "Epoch: 245, loss_train: 2150.29, accuracy_train: 76.8%\n",
      "\t  loss_test: 588.89, accuracy_test: 72.4%\n",
      "Epoch: 246, loss_train: 2142.59, accuracy_train: 76.9%\n",
      "\t  loss_test: 587.51, accuracy_test: 72.5%\n",
      "Epoch: 247, loss_train: 2134.99, accuracy_train: 77.0%\n",
      "\t  loss_test: 586.15, accuracy_test: 72.5%\n",
      "Epoch: 248, loss_train: 2127.47, accuracy_train: 77.1%\n",
      "\t  loss_test: 584.82, accuracy_test: 72.4%\n",
      "Epoch: 249, loss_train: 2120.04, accuracy_train: 77.3%\n",
      "\t  loss_test: 583.52, accuracy_test: 72.4%\n",
      "Epoch: 250, loss_train: 2112.71, accuracy_train: 77.4%\n",
      "\t  loss_test: 582.24, accuracy_test: 72.1%\n",
      "Epoch: 251, loss_train: 2105.45, accuracy_train: 77.5%\n",
      "\t  loss_test: 580.99, accuracy_test: 72.1%\n",
      "Epoch: 252, loss_train: 2098.28, accuracy_train: 77.6%\n",
      "\t  loss_test: 579.76, accuracy_test: 72.3%\n",
      "Epoch: 253, loss_train: 2091.19, accuracy_train: 77.6%\n",
      "\t  loss_test: 578.56, accuracy_test: 72.4%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 254, loss_train: 2084.18, accuracy_train: 77.7%\n",
      "\t  loss_test: 577.38, accuracy_test: 72.4%\n",
      "Epoch: 255, loss_train: 2077.25, accuracy_train: 77.7%\n",
      "\t  loss_test: 576.23, accuracy_test: 72.4%\n",
      "Epoch: 256, loss_train: 2070.40, accuracy_train: 77.8%\n",
      "\t  loss_test: 575.10, accuracy_test: 72.3%\n",
      "Epoch: 257, loss_train: 2063.62, accuracy_train: 77.9%\n",
      "\t  loss_test: 573.99, accuracy_test: 72.4%\n",
      "Epoch: 258, loss_train: 2056.91, accuracy_train: 77.9%\n",
      "\t  loss_test: 572.91, accuracy_test: 72.4%\n",
      "Epoch: 259, loss_train: 2050.27, accuracy_train: 78.0%\n",
      "\t  loss_test: 571.84, accuracy_test: 72.4%\n",
      "Epoch: 260, loss_train: 2043.70, accuracy_train: 78.1%\n",
      "\t  loss_test: 570.80, accuracy_test: 72.4%\n",
      "Epoch: 261, loss_train: 2037.20, accuracy_train: 78.3%\n",
      "\t  loss_test: 569.78, accuracy_test: 72.4%\n",
      "Epoch: 262, loss_train: 2030.76, accuracy_train: 78.4%\n",
      "\t  loss_test: 568.77, accuracy_test: 72.8%\n",
      "Epoch: 263, loss_train: 2024.38, accuracy_train: 78.4%\n",
      "\t  loss_test: 567.79, accuracy_test: 72.8%\n",
      "Epoch: 264, loss_train: 2018.06, accuracy_train: 78.4%\n",
      "\t  loss_test: 566.83, accuracy_test: 72.8%\n",
      "Epoch: 265, loss_train: 2011.81, accuracy_train: 78.4%\n",
      "\t  loss_test: 565.88, accuracy_test: 72.9%\n",
      "Epoch: 266, loss_train: 2005.61, accuracy_train: 78.5%\n",
      "\t  loss_test: 564.95, accuracy_test: 73.1%\n",
      "Epoch: 267, loss_train: 1999.46, accuracy_train: 78.5%\n",
      "\t  loss_test: 564.04, accuracy_test: 73.1%\n",
      "Epoch: 268, loss_train: 1993.37, accuracy_train: 78.6%\n",
      "\t  loss_test: 563.15, accuracy_test: 73.2%\n",
      "Epoch: 269, loss_train: 1987.33, accuracy_train: 78.6%\n",
      "\t  loss_test: 562.27, accuracy_test: 73.2%\n",
      "Epoch: 270, loss_train: 1981.34, accuracy_train: 78.6%\n",
      "\t  loss_test: 561.41, accuracy_test: 73.3%\n",
      "Epoch: 271, loss_train: 1975.40, accuracy_train: 78.7%\n",
      "\t  loss_test: 560.57, accuracy_test: 73.3%\n",
      "Epoch: 272, loss_train: 1969.51, accuracy_train: 78.7%\n",
      "\t  loss_test: 559.74, accuracy_test: 73.5%\n",
      "Epoch: 273, loss_train: 1963.66, accuracy_train: 78.8%\n",
      "\t  loss_test: 558.92, accuracy_test: 73.5%\n",
      "Epoch: 274, loss_train: 1957.86, accuracy_train: 79.0%\n",
      "\t  loss_test: 558.12, accuracy_test: 73.5%\n",
      "Epoch: 275, loss_train: 1952.10, accuracy_train: 79.1%\n",
      "\t  loss_test: 557.33, accuracy_test: 73.5%\n",
      "Epoch: 276, loss_train: 1946.39, accuracy_train: 79.2%\n",
      "\t  loss_test: 556.56, accuracy_test: 73.3%\n",
      "Epoch: 277, loss_train: 1940.71, accuracy_train: 79.2%\n",
      "\t  loss_test: 555.80, accuracy_test: 73.3%\n",
      "Epoch: 278, loss_train: 1935.07, accuracy_train: 79.2%\n",
      "\t  loss_test: 555.06, accuracy_test: 73.3%\n",
      "Epoch: 279, loss_train: 1929.47, accuracy_train: 79.3%\n",
      "\t  loss_test: 554.32, accuracy_test: 73.2%\n",
      "Epoch: 280, loss_train: 1923.91, accuracy_train: 79.3%\n",
      "\t  loss_test: 553.60, accuracy_test: 73.2%\n",
      "Epoch: 281, loss_train: 1918.38, accuracy_train: 79.3%\n",
      "\t  loss_test: 552.89, accuracy_test: 73.2%\n",
      "Epoch: 282, loss_train: 1912.89, accuracy_train: 79.3%\n",
      "\t  loss_test: 552.19, accuracy_test: 73.2%\n",
      "Epoch: 283, loss_train: 1907.43, accuracy_train: 79.4%\n",
      "\t  loss_test: 551.50, accuracy_test: 73.2%\n",
      "Epoch: 284, loss_train: 1902.01, accuracy_train: 79.5%\n",
      "\t  loss_test: 550.83, accuracy_test: 73.3%\n",
      "Epoch: 285, loss_train: 1896.61, accuracy_train: 79.5%\n",
      "\t  loss_test: 550.16, accuracy_test: 73.5%\n",
      "Epoch: 286, loss_train: 1891.25, accuracy_train: 79.5%\n",
      "\t  loss_test: 549.50, accuracy_test: 73.6%\n",
      "Epoch: 287, loss_train: 1885.92, accuracy_train: 79.6%\n",
      "\t  loss_test: 548.86, accuracy_test: 73.7%\n",
      "Epoch: 288, loss_train: 1880.61, accuracy_train: 79.6%\n",
      "\t  loss_test: 548.22, accuracy_test: 73.7%\n",
      "Epoch: 289, loss_train: 1875.34, accuracy_train: 79.6%\n",
      "\t  loss_test: 547.60, accuracy_test: 73.6%\n",
      "Epoch: 290, loss_train: 1870.09, accuracy_train: 79.7%\n",
      "\t  loss_test: 546.98, accuracy_test: 73.6%\n",
      "Epoch: 291, loss_train: 1864.87, accuracy_train: 79.7%\n",
      "\t  loss_test: 546.37, accuracy_test: 73.6%\n",
      "Epoch: 292, loss_train: 1859.68, accuracy_train: 79.7%\n",
      "\t  loss_test: 545.78, accuracy_test: 73.6%\n",
      "Epoch: 293, loss_train: 1854.51, accuracy_train: 79.8%\n",
      "\t  loss_test: 545.19, accuracy_test: 73.7%\n",
      "Epoch: 294, loss_train: 1849.37, accuracy_train: 79.8%\n",
      "\t  loss_test: 544.60, accuracy_test: 73.7%\n",
      "Epoch: 295, loss_train: 1844.26, accuracy_train: 79.8%\n",
      "\t  loss_test: 544.03, accuracy_test: 73.7%\n",
      "Epoch: 296, loss_train: 1839.16, accuracy_train: 79.9%\n",
      "\t  loss_test: 543.47, accuracy_test: 73.7%\n",
      "Epoch: 297, loss_train: 1834.10, accuracy_train: 80.0%\n",
      "\t  loss_test: 542.91, accuracy_test: 73.7%\n",
      "Epoch: 298, loss_train: 1829.05, accuracy_train: 80.0%\n",
      "\t  loss_test: 542.36, accuracy_test: 73.7%\n",
      "Epoch: 299, loss_train: 1824.03, accuracy_train: 80.0%\n",
      "\t  loss_test: 541.82, accuracy_test: 73.7%\n",
      "Epoch: 300, loss_train: 1819.03, accuracy_train: 80.1%\n",
      "\t  loss_test: 541.29, accuracy_test: 73.7%\n",
      "Epoch: 301, loss_train: 1814.05, accuracy_train: 80.1%\n",
      "\t  loss_test: 540.76, accuracy_test: 73.7%\n",
      "Epoch: 302, loss_train: 1809.09, accuracy_train: 80.1%\n",
      "\t  loss_test: 540.24, accuracy_test: 73.7%\n",
      "Epoch: 303, loss_train: 1804.16, accuracy_train: 80.2%\n",
      "\t  loss_test: 539.73, accuracy_test: 73.7%\n",
      "Epoch: 304, loss_train: 1799.25, accuracy_train: 80.2%\n",
      "\t  loss_test: 539.22, accuracy_test: 73.7%\n",
      "Epoch: 305, loss_train: 1794.35, accuracy_train: 80.2%\n",
      "\t  loss_test: 538.72, accuracy_test: 73.9%\n",
      "Epoch: 306, loss_train: 1789.48, accuracy_train: 80.2%\n",
      "\t  loss_test: 538.23, accuracy_test: 73.9%\n",
      "Epoch: 307, loss_train: 1784.63, accuracy_train: 80.3%\n",
      "\t  loss_test: 537.75, accuracy_test: 73.9%\n",
      "Epoch: 308, loss_train: 1779.79, accuracy_train: 80.3%\n",
      "\t  loss_test: 537.27, accuracy_test: 73.9%\n",
      "Epoch: 309, loss_train: 1774.98, accuracy_train: 80.4%\n",
      "\t  loss_test: 536.79, accuracy_test: 73.9%\n",
      "Epoch: 310, loss_train: 1770.19, accuracy_train: 80.5%\n",
      "\t  loss_test: 536.33, accuracy_test: 73.7%\n",
      "Epoch: 311, loss_train: 1765.41, accuracy_train: 80.5%\n",
      "\t  loss_test: 535.86, accuracy_test: 73.9%\n",
      "Epoch: 312, loss_train: 1760.66, accuracy_train: 80.6%\n",
      "\t  loss_test: 535.41, accuracy_test: 73.9%\n",
      "Epoch: 313, loss_train: 1755.92, accuracy_train: 80.7%\n",
      "\t  loss_test: 534.96, accuracy_test: 73.9%\n",
      "Epoch: 314, loss_train: 1751.20, accuracy_train: 80.7%\n",
      "\t  loss_test: 534.51, accuracy_test: 73.9%\n",
      "Epoch: 315, loss_train: 1746.50, accuracy_train: 80.8%\n",
      "\t  loss_test: 534.08, accuracy_test: 73.9%\n",
      "Epoch: 316, loss_train: 1741.82, accuracy_train: 80.8%\n",
      "\t  loss_test: 533.64, accuracy_test: 73.9%\n",
      "Epoch: 317, loss_train: 1737.15, accuracy_train: 80.8%\n",
      "\t  loss_test: 533.22, accuracy_test: 74.0%\n",
      "Epoch: 318, loss_train: 1732.51, accuracy_train: 80.9%\n",
      "\t  loss_test: 532.79, accuracy_test: 74.0%\n",
      "Epoch: 319, loss_train: 1727.88, accuracy_train: 81.0%\n",
      "\t  loss_test: 532.38, accuracy_test: 74.0%\n",
      "Epoch: 320, loss_train: 1723.27, accuracy_train: 81.0%\n",
      "\t  loss_test: 531.96, accuracy_test: 74.1%\n",
      "Epoch: 321, loss_train: 1718.67, accuracy_train: 81.1%\n",
      "\t  loss_test: 531.56, accuracy_test: 74.2%\n",
      "Epoch: 322, loss_train: 1714.10, accuracy_train: 81.1%\n",
      "\t  loss_test: 531.15, accuracy_test: 74.2%\n",
      "Epoch: 323, loss_train: 1709.54, accuracy_train: 81.2%\n",
      "\t  loss_test: 530.76, accuracy_test: 74.2%\n",
      "Epoch: 324, loss_train: 1705.00, accuracy_train: 81.2%\n",
      "\t  loss_test: 530.36, accuracy_test: 74.2%\n",
      "Epoch: 325, loss_train: 1700.48, accuracy_train: 81.2%\n",
      "\t  loss_test: 529.97, accuracy_test: 74.2%\n",
      "Epoch: 326, loss_train: 1695.97, accuracy_train: 81.2%\n",
      "\t  loss_test: 529.59, accuracy_test: 74.2%\n",
      "Epoch: 327, loss_train: 1691.48, accuracy_train: 81.2%\n",
      "\t  loss_test: 529.21, accuracy_test: 74.2%\n",
      "Epoch: 328, loss_train: 1687.01, accuracy_train: 81.2%\n",
      "\t  loss_test: 528.84, accuracy_test: 74.2%\n",
      "Epoch: 329, loss_train: 1682.55, accuracy_train: 81.3%\n",
      "\t  loss_test: 528.47, accuracy_test: 74.2%\n",
      "Epoch: 330, loss_train: 1678.11, accuracy_train: 81.4%\n",
      "\t  loss_test: 528.10, accuracy_test: 74.2%\n",
      "Epoch: 331, loss_train: 1673.69, accuracy_train: 81.4%\n",
      "\t  loss_test: 527.74, accuracy_test: 74.4%\n",
      "Epoch: 332, loss_train: 1669.29, accuracy_train: 81.4%\n",
      "\t  loss_test: 527.38, accuracy_test: 74.4%\n",
      "Epoch: 333, loss_train: 1664.90, accuracy_train: 81.4%\n",
      "\t  loss_test: 527.03, accuracy_test: 74.4%\n",
      "Epoch: 334, loss_train: 1660.53, accuracy_train: 81.4%\n",
      "\t  loss_test: 526.68, accuracy_test: 74.4%\n",
      "Epoch: 335, loss_train: 1656.17, accuracy_train: 81.4%\n",
      "\t  loss_test: 526.34, accuracy_test: 74.5%\n",
      "Epoch: 336, loss_train: 1651.84, accuracy_train: 81.4%\n",
      "\t  loss_test: 525.99, accuracy_test: 74.5%\n",
      "Epoch: 337, loss_train: 1647.51, accuracy_train: 81.4%\n",
      "\t  loss_test: 525.66, accuracy_test: 74.5%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 338, loss_train: 1643.21, accuracy_train: 81.5%\n",
      "\t  loss_test: 525.32, accuracy_test: 74.5%\n",
      "Epoch: 339, loss_train: 1638.92, accuracy_train: 81.5%\n",
      "\t  loss_test: 524.99, accuracy_test: 74.5%\n",
      "Epoch: 340, loss_train: 1634.65, accuracy_train: 81.6%\n",
      "\t  loss_test: 524.67, accuracy_test: 74.5%\n",
      "Epoch: 341, loss_train: 1630.40, accuracy_train: 81.6%\n",
      "\t  loss_test: 524.35, accuracy_test: 74.5%\n",
      "Epoch: 342, loss_train: 1626.16, accuracy_train: 81.6%\n",
      "\t  loss_test: 524.03, accuracy_test: 74.5%\n",
      "Epoch: 343, loss_train: 1621.94, accuracy_train: 81.7%\n",
      "\t  loss_test: 523.71, accuracy_test: 74.5%\n",
      "Epoch: 344, loss_train: 1617.73, accuracy_train: 81.7%\n",
      "\t  loss_test: 523.40, accuracy_test: 74.5%\n",
      "Epoch: 345, loss_train: 1613.54, accuracy_train: 81.7%\n",
      "\t  loss_test: 523.10, accuracy_test: 74.5%\n",
      "Epoch: 346, loss_train: 1609.37, accuracy_train: 81.8%\n",
      "\t  loss_test: 522.79, accuracy_test: 74.5%\n",
      "Epoch: 347, loss_train: 1605.21, accuracy_train: 81.9%\n",
      "\t  loss_test: 522.49, accuracy_test: 74.6%\n",
      "Epoch: 348, loss_train: 1601.08, accuracy_train: 81.9%\n",
      "\t  loss_test: 522.19, accuracy_test: 74.6%\n",
      "Epoch: 349, loss_train: 1596.95, accuracy_train: 82.0%\n",
      "\t  loss_test: 521.90, accuracy_test: 74.6%\n",
      "Epoch: 350, loss_train: 1592.85, accuracy_train: 82.0%\n",
      "\t  loss_test: 521.61, accuracy_test: 74.6%\n",
      "Epoch: 351, loss_train: 1588.75, accuracy_train: 82.0%\n",
      "\t  loss_test: 521.32, accuracy_test: 74.6%\n",
      "Epoch: 352, loss_train: 1584.68, accuracy_train: 82.1%\n",
      "\t  loss_test: 521.04, accuracy_test: 74.6%\n",
      "Epoch: 353, loss_train: 1580.62, accuracy_train: 82.1%\n",
      "\t  loss_test: 520.76, accuracy_test: 74.6%\n",
      "Epoch: 354, loss_train: 1576.58, accuracy_train: 82.1%\n",
      "\t  loss_test: 520.48, accuracy_test: 74.6%\n",
      "Epoch: 355, loss_train: 1572.56, accuracy_train: 82.2%\n",
      "\t  loss_test: 520.21, accuracy_test: 74.6%\n",
      "Epoch: 356, loss_train: 1568.55, accuracy_train: 82.2%\n",
      "\t  loss_test: 519.94, accuracy_test: 74.6%\n",
      "Epoch: 357, loss_train: 1564.55, accuracy_train: 82.2%\n",
      "\t  loss_test: 519.67, accuracy_test: 74.6%\n",
      "Epoch: 358, loss_train: 1560.58, accuracy_train: 82.3%\n",
      "\t  loss_test: 519.40, accuracy_test: 74.6%\n",
      "Epoch: 359, loss_train: 1556.61, accuracy_train: 82.3%\n",
      "\t  loss_test: 519.14, accuracy_test: 74.6%\n",
      "Epoch: 360, loss_train: 1552.67, accuracy_train: 82.4%\n",
      "\t  loss_test: 518.88, accuracy_test: 74.6%\n",
      "Epoch: 361, loss_train: 1548.74, accuracy_train: 82.4%\n",
      "\t  loss_test: 518.63, accuracy_test: 74.6%\n",
      "Epoch: 362, loss_train: 1544.83, accuracy_train: 82.5%\n",
      "\t  loss_test: 518.37, accuracy_test: 74.6%\n",
      "Epoch: 363, loss_train: 1540.93, accuracy_train: 82.5%\n",
      "\t  loss_test: 518.12, accuracy_test: 74.5%\n",
      "Epoch: 364, loss_train: 1537.05, accuracy_train: 82.6%\n",
      "\t  loss_test: 517.87, accuracy_test: 74.5%\n",
      "Epoch: 365, loss_train: 1533.18, accuracy_train: 82.7%\n",
      "\t  loss_test: 517.63, accuracy_test: 74.5%\n",
      "Epoch: 366, loss_train: 1529.33, accuracy_train: 82.7%\n",
      "\t  loss_test: 517.39, accuracy_test: 74.5%\n",
      "Epoch: 367, loss_train: 1525.50, accuracy_train: 82.8%\n",
      "\t  loss_test: 517.15, accuracy_test: 74.5%\n",
      "Epoch: 368, loss_train: 1521.68, accuracy_train: 82.8%\n",
      "\t  loss_test: 516.91, accuracy_test: 74.5%\n",
      "Epoch: 369, loss_train: 1517.87, accuracy_train: 82.8%\n",
      "\t  loss_test: 516.68, accuracy_test: 74.5%\n",
      "Epoch: 370, loss_train: 1514.09, accuracy_train: 82.8%\n",
      "\t  loss_test: 516.44, accuracy_test: 74.5%\n",
      "Epoch: 371, loss_train: 1510.31, accuracy_train: 82.9%\n",
      "\t  loss_test: 516.22, accuracy_test: 74.5%\n",
      "Epoch: 372, loss_train: 1506.56, accuracy_train: 83.0%\n",
      "\t  loss_test: 515.99, accuracy_test: 74.5%\n",
      "Epoch: 373, loss_train: 1502.81, accuracy_train: 83.0%\n",
      "\t  loss_test: 515.77, accuracy_test: 74.5%\n",
      "Epoch: 374, loss_train: 1499.09, accuracy_train: 83.1%\n",
      "\t  loss_test: 515.54, accuracy_test: 74.5%\n",
      "Epoch: 375, loss_train: 1495.38, accuracy_train: 83.1%\n",
      "\t  loss_test: 515.33, accuracy_test: 74.5%\n",
      "Epoch: 376, loss_train: 1491.68, accuracy_train: 83.2%\n",
      "\t  loss_test: 515.11, accuracy_test: 74.5%\n",
      "Epoch: 377, loss_train: 1488.00, accuracy_train: 83.3%\n",
      "\t  loss_test: 514.90, accuracy_test: 74.6%\n",
      "Epoch: 378, loss_train: 1484.33, accuracy_train: 83.4%\n",
      "\t  loss_test: 514.68, accuracy_test: 74.6%\n",
      "Epoch: 379, loss_train: 1480.68, accuracy_train: 83.4%\n",
      "\t  loss_test: 514.48, accuracy_test: 74.6%\n",
      "Epoch: 380, loss_train: 1477.04, accuracy_train: 83.4%\n",
      "\t  loss_test: 514.27, accuracy_test: 74.6%\n",
      "Epoch: 381, loss_train: 1473.42, accuracy_train: 83.5%\n",
      "\t  loss_test: 514.06, accuracy_test: 74.6%\n",
      "Epoch: 382, loss_train: 1469.81, accuracy_train: 83.6%\n",
      "\t  loss_test: 513.86, accuracy_test: 74.6%\n",
      "Epoch: 383, loss_train: 1466.22, accuracy_train: 83.6%\n",
      "\t  loss_test: 513.66, accuracy_test: 74.6%\n",
      "Epoch: 384, loss_train: 1462.64, accuracy_train: 83.6%\n",
      "\t  loss_test: 513.46, accuracy_test: 74.6%\n",
      "Epoch: 385, loss_train: 1459.08, accuracy_train: 83.6%\n",
      "\t  loss_test: 513.27, accuracy_test: 74.6%\n",
      "Epoch: 386, loss_train: 1455.53, accuracy_train: 83.7%\n",
      "\t  loss_test: 513.08, accuracy_test: 74.6%\n",
      "Epoch: 387, loss_train: 1451.99, accuracy_train: 83.7%\n",
      "\t  loss_test: 512.88, accuracy_test: 74.6%\n",
      "Epoch: 388, loss_train: 1448.47, accuracy_train: 83.7%\n",
      "\t  loss_test: 512.69, accuracy_test: 74.6%\n",
      "Epoch: 389, loss_train: 1444.96, accuracy_train: 83.8%\n",
      "\t  loss_test: 512.51, accuracy_test: 74.6%\n",
      "Epoch: 390, loss_train: 1441.47, accuracy_train: 83.8%\n",
      "\t  loss_test: 512.32, accuracy_test: 74.6%\n",
      "Epoch: 391, loss_train: 1437.99, accuracy_train: 83.8%\n",
      "\t  loss_test: 512.14, accuracy_test: 74.6%\n",
      "Epoch: 392, loss_train: 1434.52, accuracy_train: 83.8%\n",
      "\t  loss_test: 511.96, accuracy_test: 74.6%\n",
      "Epoch: 393, loss_train: 1431.07, accuracy_train: 83.9%\n",
      "\t  loss_test: 511.78, accuracy_test: 74.6%\n",
      "Epoch: 394, loss_train: 1427.63, accuracy_train: 83.9%\n",
      "\t  loss_test: 511.60, accuracy_test: 74.6%\n",
      "Epoch: 395, loss_train: 1424.20, accuracy_train: 84.0%\n",
      "\t  loss_test: 511.43, accuracy_test: 74.6%\n",
      "Epoch: 396, loss_train: 1420.79, accuracy_train: 84.0%\n",
      "\t  loss_test: 511.26, accuracy_test: 74.6%\n",
      "Epoch: 397, loss_train: 1417.39, accuracy_train: 84.0%\n",
      "\t  loss_test: 511.08, accuracy_test: 74.6%\n",
      "Epoch: 398, loss_train: 1414.00, accuracy_train: 84.0%\n",
      "\t  loss_test: 510.92, accuracy_test: 74.6%\n",
      "Epoch: 399, loss_train: 1410.63, accuracy_train: 84.0%\n",
      "\t  loss_test: 510.75, accuracy_test: 74.6%\n",
      "Epoch: 400, loss_train: 1407.27, accuracy_train: 84.1%\n",
      "\t  loss_test: 510.58, accuracy_test: 74.6%\n",
      "Epoch: 401, loss_train: 1403.92, accuracy_train: 84.1%\n",
      "\t  loss_test: 510.42, accuracy_test: 74.6%\n",
      "Epoch: 402, loss_train: 1400.58, accuracy_train: 84.2%\n",
      "\t  loss_test: 510.26, accuracy_test: 74.8%\n",
      "Epoch: 403, loss_train: 1397.26, accuracy_train: 84.2%\n",
      "\t  loss_test: 510.10, accuracy_test: 74.6%\n",
      "Epoch: 404, loss_train: 1393.95, accuracy_train: 84.3%\n",
      "\t  loss_test: 509.94, accuracy_test: 74.6%\n",
      "Epoch: 405, loss_train: 1390.65, accuracy_train: 84.4%\n",
      "\t  loss_test: 509.78, accuracy_test: 74.6%\n",
      "Epoch: 406, loss_train: 1387.37, accuracy_train: 84.5%\n",
      "\t  loss_test: 509.63, accuracy_test: 74.6%\n",
      "Epoch: 407, loss_train: 1384.09, accuracy_train: 84.5%\n",
      "\t  loss_test: 509.47, accuracy_test: 74.6%\n",
      "Epoch: 408, loss_train: 1380.83, accuracy_train: 84.5%\n",
      "\t  loss_test: 509.32, accuracy_test: 74.6%\n",
      "Epoch: 409, loss_train: 1377.58, accuracy_train: 84.5%\n",
      "\t  loss_test: 509.17, accuracy_test: 74.8%\n",
      "Epoch: 410, loss_train: 1374.34, accuracy_train: 84.5%\n",
      "\t  loss_test: 509.02, accuracy_test: 74.8%\n",
      "Epoch: 411, loss_train: 1371.11, accuracy_train: 84.6%\n",
      "\t  loss_test: 508.88, accuracy_test: 74.8%\n",
      "Epoch: 412, loss_train: 1367.90, accuracy_train: 84.6%\n",
      "\t  loss_test: 508.73, accuracy_test: 74.9%\n",
      "Epoch: 413, loss_train: 1364.70, accuracy_train: 84.6%\n",
      "\t  loss_test: 508.59, accuracy_test: 74.9%\n",
      "Epoch: 414, loss_train: 1361.50, accuracy_train: 84.6%\n",
      "\t  loss_test: 508.44, accuracy_test: 74.9%\n",
      "Epoch: 415, loss_train: 1358.32, accuracy_train: 84.7%\n",
      "\t  loss_test: 508.30, accuracy_test: 74.9%\n",
      "Epoch: 416, loss_train: 1355.15, accuracy_train: 84.7%\n",
      "\t  loss_test: 508.16, accuracy_test: 74.9%\n",
      "Epoch: 417, loss_train: 1351.99, accuracy_train: 84.7%\n",
      "\t  loss_test: 508.02, accuracy_test: 74.9%\n",
      "Epoch: 418, loss_train: 1348.84, accuracy_train: 84.8%\n",
      "\t  loss_test: 507.89, accuracy_test: 74.9%\n",
      "Epoch: 419, loss_train: 1345.70, accuracy_train: 84.8%\n",
      "\t  loss_test: 507.75, accuracy_test: 74.9%\n",
      "Epoch: 420, loss_train: 1342.58, accuracy_train: 84.8%\n",
      "\t  loss_test: 507.62, accuracy_test: 74.9%\n",
      "Epoch: 421, loss_train: 1339.46, accuracy_train: 84.9%\n",
      "\t  loss_test: 507.49, accuracy_test: 74.8%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 422, loss_train: 1336.35, accuracy_train: 84.9%\n",
      "\t  loss_test: 507.35, accuracy_test: 74.8%\n",
      "Epoch: 423, loss_train: 1333.26, accuracy_train: 84.9%\n",
      "\t  loss_test: 507.22, accuracy_test: 74.8%\n",
      "Epoch: 424, loss_train: 1330.17, accuracy_train: 84.9%\n",
      "\t  loss_test: 507.10, accuracy_test: 74.8%\n",
      "Epoch: 425, loss_train: 1327.09, accuracy_train: 84.9%\n",
      "\t  loss_test: 506.97, accuracy_test: 74.8%\n",
      "Epoch: 426, loss_train: 1324.03, accuracy_train: 85.0%\n",
      "\t  loss_test: 506.84, accuracy_test: 74.9%\n",
      "Epoch: 427, loss_train: 1320.97, accuracy_train: 85.0%\n",
      "\t  loss_test: 506.72, accuracy_test: 74.8%\n",
      "Epoch: 428, loss_train: 1317.92, accuracy_train: 85.0%\n",
      "\t  loss_test: 506.59, accuracy_test: 74.8%\n",
      "Epoch: 429, loss_train: 1314.88, accuracy_train: 85.1%\n",
      "\t  loss_test: 506.47, accuracy_test: 74.8%\n",
      "Epoch: 430, loss_train: 1311.85, accuracy_train: 85.1%\n",
      "\t  loss_test: 506.35, accuracy_test: 74.8%\n",
      "Epoch: 431, loss_train: 1308.84, accuracy_train: 85.1%\n",
      "\t  loss_test: 506.23, accuracy_test: 74.6%\n",
      "Epoch: 432, loss_train: 1305.82, accuracy_train: 85.1%\n",
      "\t  loss_test: 506.11, accuracy_test: 74.6%\n",
      "Epoch: 433, loss_train: 1302.82, accuracy_train: 85.1%\n",
      "\t  loss_test: 505.99, accuracy_test: 74.6%\n",
      "Epoch: 434, loss_train: 1299.83, accuracy_train: 85.1%\n",
      "\t  loss_test: 505.87, accuracy_test: 74.6%\n",
      "Epoch: 435, loss_train: 1296.85, accuracy_train: 85.2%\n",
      "\t  loss_test: 505.76, accuracy_test: 74.6%\n",
      "Epoch: 436, loss_train: 1293.87, accuracy_train: 85.2%\n",
      "\t  loss_test: 505.64, accuracy_test: 74.6%\n",
      "Epoch: 437, loss_train: 1290.91, accuracy_train: 85.3%\n",
      "\t  loss_test: 505.53, accuracy_test: 74.6%\n",
      "Epoch: 438, loss_train: 1287.95, accuracy_train: 85.3%\n",
      "\t  loss_test: 505.42, accuracy_test: 74.6%\n",
      "Epoch: 439, loss_train: 1285.00, accuracy_train: 85.3%\n",
      "\t  loss_test: 505.30, accuracy_test: 74.6%\n",
      "Epoch: 440, loss_train: 1282.06, accuracy_train: 85.3%\n",
      "\t  loss_test: 505.19, accuracy_test: 74.6%\n",
      "Epoch: 441, loss_train: 1279.12, accuracy_train: 85.3%\n",
      "\t  loss_test: 505.08, accuracy_test: 74.6%\n",
      "Epoch: 442, loss_train: 1276.20, accuracy_train: 85.3%\n",
      "\t  loss_test: 504.97, accuracy_test: 74.6%\n",
      "Epoch: 443, loss_train: 1273.28, accuracy_train: 85.3%\n",
      "\t  loss_test: 504.86, accuracy_test: 74.6%\n",
      "Epoch: 444, loss_train: 1270.37, accuracy_train: 85.3%\n",
      "\t  loss_test: 504.76, accuracy_test: 74.6%\n",
      "Epoch: 445, loss_train: 1267.47, accuracy_train: 85.3%\n",
      "\t  loss_test: 504.65, accuracy_test: 74.6%\n",
      "Epoch: 446, loss_train: 1264.57, accuracy_train: 85.3%\n",
      "\t  loss_test: 504.54, accuracy_test: 74.6%\n",
      "Epoch: 447, loss_train: 1261.68, accuracy_train: 85.3%\n",
      "\t  loss_test: 504.44, accuracy_test: 74.6%\n",
      "Epoch: 448, loss_train: 1258.80, accuracy_train: 85.4%\n",
      "\t  loss_test: 504.33, accuracy_test: 74.8%\n",
      "Epoch: 449, loss_train: 1255.93, accuracy_train: 85.4%\n",
      "\t  loss_test: 504.23, accuracy_test: 74.9%\n",
      "Epoch: 450, loss_train: 1253.06, accuracy_train: 85.4%\n",
      "\t  loss_test: 504.13, accuracy_test: 74.9%\n",
      "Epoch: 451, loss_train: 1250.20, accuracy_train: 85.4%\n",
      "\t  loss_test: 504.03, accuracy_test: 75.0%\n",
      "Epoch: 452, loss_train: 1247.35, accuracy_train: 85.4%\n",
      "\t  loss_test: 503.92, accuracy_test: 75.0%\n",
      "Epoch: 453, loss_train: 1244.50, accuracy_train: 85.4%\n",
      "\t  loss_test: 503.82, accuracy_test: 75.0%\n",
      "Epoch: 454, loss_train: 1241.66, accuracy_train: 85.4%\n",
      "\t  loss_test: 503.72, accuracy_test: 75.0%\n",
      "Epoch: 455, loss_train: 1238.83, accuracy_train: 85.4%\n",
      "\t  loss_test: 503.62, accuracy_test: 75.0%\n",
      "Epoch: 456, loss_train: 1236.00, accuracy_train: 85.4%\n",
      "\t  loss_test: 503.53, accuracy_test: 75.0%\n",
      "Epoch: 457, loss_train: 1233.18, accuracy_train: 85.5%\n",
      "\t  loss_test: 503.43, accuracy_test: 75.2%\n",
      "Epoch: 458, loss_train: 1230.37, accuracy_train: 85.5%\n",
      "\t  loss_test: 503.33, accuracy_test: 75.2%\n",
      "Epoch: 459, loss_train: 1227.56, accuracy_train: 85.5%\n",
      "\t  loss_test: 503.23, accuracy_test: 75.2%\n",
      "Epoch: 460, loss_train: 1224.75, accuracy_train: 85.5%\n",
      "\t  loss_test: 503.14, accuracy_test: 75.2%\n",
      "Epoch: 461, loss_train: 1221.95, accuracy_train: 85.6%\n",
      "\t  loss_test: 503.04, accuracy_test: 75.2%\n",
      "Epoch: 462, loss_train: 1219.16, accuracy_train: 85.6%\n",
      "\t  loss_test: 502.94, accuracy_test: 75.2%\n",
      "Epoch: 463, loss_train: 1216.38, accuracy_train: 85.6%\n",
      "\t  loss_test: 502.85, accuracy_test: 75.2%\n",
      "Epoch: 464, loss_train: 1213.59, accuracy_train: 85.6%\n",
      "\t  loss_test: 502.76, accuracy_test: 75.2%\n",
      "Epoch: 465, loss_train: 1210.82, accuracy_train: 86.2%\n",
      "\t  loss_test: 502.66, accuracy_test: 75.6%\n",
      "Epoch: 466, loss_train: 1208.05, accuracy_train: 86.2%\n",
      "\t  loss_test: 502.57, accuracy_test: 75.6%\n",
      "Epoch: 467, loss_train: 1205.28, accuracy_train: 86.2%\n",
      "\t  loss_test: 502.47, accuracy_test: 75.6%\n",
      "Epoch: 468, loss_train: 1202.52, accuracy_train: 86.2%\n",
      "\t  loss_test: 502.38, accuracy_test: 75.6%\n",
      "Epoch: 469, loss_train: 1199.76, accuracy_train: 86.2%\n",
      "\t  loss_test: 502.29, accuracy_test: 75.6%\n",
      "Epoch: 470, loss_train: 1197.01, accuracy_train: 86.2%\n",
      "\t  loss_test: 502.20, accuracy_test: 75.6%\n",
      "Epoch: 471, loss_train: 1194.27, accuracy_train: 86.3%\n",
      "\t  loss_test: 502.11, accuracy_test: 75.4%\n",
      "Epoch: 472, loss_train: 1191.52, accuracy_train: 86.3%\n",
      "\t  loss_test: 502.02, accuracy_test: 75.4%\n",
      "Epoch: 473, loss_train: 1188.78, accuracy_train: 86.3%\n",
      "\t  loss_test: 501.92, accuracy_test: 75.6%\n",
      "Epoch: 474, loss_train: 1186.05, accuracy_train: 86.3%\n",
      "\t  loss_test: 501.83, accuracy_test: 75.7%\n",
      "Epoch: 475, loss_train: 1183.32, accuracy_train: 86.3%\n",
      "\t  loss_test: 501.74, accuracy_test: 75.7%\n",
      "Epoch: 476, loss_train: 1180.60, accuracy_train: 86.3%\n",
      "\t  loss_test: 501.65, accuracy_test: 75.6%\n",
      "Epoch: 477, loss_train: 1177.87, accuracy_train: 86.3%\n",
      "\t  loss_test: 501.56, accuracy_test: 75.6%\n",
      "Epoch: 478, loss_train: 1175.16, accuracy_train: 86.3%\n",
      "\t  loss_test: 501.47, accuracy_test: 75.6%\n",
      "Epoch: 479, loss_train: 1172.44, accuracy_train: 86.3%\n",
      "\t  loss_test: 501.39, accuracy_test: 75.7%\n",
      "Epoch: 480, loss_train: 1169.73, accuracy_train: 86.3%\n",
      "\t  loss_test: 501.30, accuracy_test: 75.7%\n",
      "Epoch: 481, loss_train: 1167.02, accuracy_train: 86.3%\n",
      "\t  loss_test: 501.21, accuracy_test: 75.7%\n",
      "Epoch: 482, loss_train: 1164.32, accuracy_train: 86.3%\n",
      "\t  loss_test: 501.12, accuracy_test: 75.6%\n",
      "Epoch: 483, loss_train: 1161.62, accuracy_train: 86.4%\n",
      "\t  loss_test: 501.03, accuracy_test: 75.6%\n",
      "Epoch: 484, loss_train: 1158.93, accuracy_train: 86.4%\n",
      "\t  loss_test: 500.94, accuracy_test: 75.6%\n",
      "Epoch: 485, loss_train: 1156.23, accuracy_train: 86.4%\n",
      "\t  loss_test: 500.86, accuracy_test: 75.6%\n",
      "Epoch: 486, loss_train: 1153.54, accuracy_train: 86.5%\n",
      "\t  loss_test: 500.77, accuracy_test: 75.6%\n",
      "Epoch: 487, loss_train: 1150.85, accuracy_train: 86.5%\n",
      "\t  loss_test: 500.68, accuracy_test: 75.6%\n",
      "Epoch: 488, loss_train: 1148.17, accuracy_train: 86.5%\n",
      "\t  loss_test: 500.59, accuracy_test: 75.6%\n",
      "Epoch: 489, loss_train: 1145.49, accuracy_train: 86.5%\n",
      "\t  loss_test: 500.50, accuracy_test: 75.6%\n",
      "Epoch: 490, loss_train: 1142.81, accuracy_train: 86.5%\n",
      "\t  loss_test: 500.42, accuracy_test: 75.6%\n",
      "Epoch: 491, loss_train: 1140.14, accuracy_train: 86.5%\n",
      "\t  loss_test: 500.33, accuracy_test: 75.6%\n",
      "Epoch: 492, loss_train: 1137.46, accuracy_train: 86.5%\n",
      "\t  loss_test: 500.24, accuracy_test: 75.6%\n",
      "Epoch: 493, loss_train: 1134.79, accuracy_train: 86.5%\n",
      "\t  loss_test: 500.16, accuracy_test: 75.6%\n",
      "Epoch: 494, loss_train: 1132.12, accuracy_train: 86.5%\n",
      "\t  loss_test: 500.07, accuracy_test: 75.4%\n",
      "Epoch: 495, loss_train: 1129.46, accuracy_train: 86.5%\n",
      "\t  loss_test: 499.98, accuracy_test: 75.4%\n",
      "Epoch: 496, loss_train: 1126.79, accuracy_train: 86.5%\n",
      "\t  loss_test: 499.89, accuracy_test: 75.4%\n",
      "Epoch: 497, loss_train: 1124.13, accuracy_train: 86.5%\n",
      "\t  loss_test: 499.81, accuracy_test: 75.4%\n",
      "Epoch: 498, loss_train: 1121.47, accuracy_train: 86.6%\n",
      "\t  loss_test: 499.72, accuracy_test: 75.4%\n",
      "Epoch: 499, loss_train: 1118.82, accuracy_train: 86.6%\n",
      "\t  loss_test: 499.63, accuracy_test: 75.4%\n",
      "Epoch: 500, loss_train: 1116.16, accuracy_train: 86.6%\n",
      "\t  loss_test: 499.55, accuracy_test: 75.4%\n",
      "Epoch: 501, loss_train: 1113.51, accuracy_train: 86.6%\n",
      "\t  loss_test: 499.46, accuracy_test: 75.4%\n",
      "Epoch: 502, loss_train: 1110.86, accuracy_train: 86.6%\n",
      "\t  loss_test: 499.37, accuracy_test: 75.4%\n",
      "Epoch: 503, loss_train: 1108.21, accuracy_train: 86.6%\n",
      "\t  loss_test: 499.28, accuracy_test: 75.4%\n",
      "Epoch: 504, loss_train: 1105.56, accuracy_train: 86.7%\n",
      "\t  loss_test: 499.20, accuracy_test: 75.4%\n",
      "Epoch: 505, loss_train: 1102.92, accuracy_train: 86.7%\n",
      "\t  loss_test: 499.11, accuracy_test: 75.4%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 506, loss_train: 1100.27, accuracy_train: 86.8%\n",
      "\t  loss_test: 499.02, accuracy_test: 75.4%\n",
      "Epoch: 507, loss_train: 1097.63, accuracy_train: 86.8%\n",
      "\t  loss_test: 498.93, accuracy_test: 75.4%\n",
      "Epoch: 508, loss_train: 1094.99, accuracy_train: 86.8%\n",
      "\t  loss_test: 498.85, accuracy_test: 75.4%\n",
      "Epoch: 509, loss_train: 1092.35, accuracy_train: 86.8%\n",
      "\t  loss_test: 498.76, accuracy_test: 75.4%\n",
      "Epoch: 510, loss_train: 1089.71, accuracy_train: 86.9%\n",
      "\t  loss_test: 498.67, accuracy_test: 75.4%\n",
      "Epoch: 511, loss_train: 1087.08, accuracy_train: 86.9%\n",
      "\t  loss_test: 498.58, accuracy_test: 75.4%\n",
      "Epoch: 512, loss_train: 1084.44, accuracy_train: 86.9%\n",
      "\t  loss_test: 498.50, accuracy_test: 75.4%\n",
      "Epoch: 513, loss_train: 1081.81, accuracy_train: 86.9%\n",
      "\t  loss_test: 498.41, accuracy_test: 75.4%\n",
      "Epoch: 514, loss_train: 1079.18, accuracy_train: 86.9%\n",
      "\t  loss_test: 498.32, accuracy_test: 75.4%\n",
      "Epoch: 515, loss_train: 1076.55, accuracy_train: 86.9%\n",
      "\t  loss_test: 498.23, accuracy_test: 75.4%\n",
      "Epoch: 516, loss_train: 1073.92, accuracy_train: 86.9%\n",
      "\t  loss_test: 498.14, accuracy_test: 75.4%\n",
      "Epoch: 517, loss_train: 1071.29, accuracy_train: 86.9%\n",
      "\t  loss_test: 498.05, accuracy_test: 75.4%\n",
      "Epoch: 518, loss_train: 1068.66, accuracy_train: 86.9%\n",
      "\t  loss_test: 497.97, accuracy_test: 75.4%\n",
      "Epoch: 519, loss_train: 1066.03, accuracy_train: 86.9%\n",
      "\t  loss_test: 497.88, accuracy_test: 75.4%\n",
      "Epoch: 520, loss_train: 1063.41, accuracy_train: 86.9%\n",
      "\t  loss_test: 497.79, accuracy_test: 75.3%\n",
      "Epoch: 521, loss_train: 1060.79, accuracy_train: 86.9%\n",
      "\t  loss_test: 497.70, accuracy_test: 75.3%\n",
      "Epoch: 522, loss_train: 1058.16, accuracy_train: 87.0%\n",
      "\t  loss_test: 497.61, accuracy_test: 75.3%\n",
      "Epoch: 523, loss_train: 1055.54, accuracy_train: 87.0%\n",
      "\t  loss_test: 497.52, accuracy_test: 75.2%\n",
      "Epoch: 524, loss_train: 1052.92, accuracy_train: 87.0%\n",
      "\t  loss_test: 497.43, accuracy_test: 75.2%\n",
      "Epoch: 525, loss_train: 1050.30, accuracy_train: 87.0%\n",
      "\t  loss_test: 497.34, accuracy_test: 75.2%\n",
      "Epoch: 526, loss_train: 1047.68, accuracy_train: 87.1%\n",
      "\t  loss_test: 497.25, accuracy_test: 75.2%\n",
      "Epoch: 527, loss_train: 1045.06, accuracy_train: 87.2%\n",
      "\t  loss_test: 497.16, accuracy_test: 75.2%\n",
      "Epoch: 528, loss_train: 1042.45, accuracy_train: 87.2%\n",
      "\t  loss_test: 497.07, accuracy_test: 75.2%\n",
      "Epoch: 529, loss_train: 1039.83, accuracy_train: 87.2%\n",
      "\t  loss_test: 496.98, accuracy_test: 75.2%\n",
      "Epoch: 530, loss_train: 1037.22, accuracy_train: 87.2%\n",
      "\t  loss_test: 496.89, accuracy_test: 75.2%\n",
      "Epoch: 531, loss_train: 1034.60, accuracy_train: 87.2%\n",
      "\t  loss_test: 496.80, accuracy_test: 75.2%\n",
      "Epoch: 532, loss_train: 1031.99, accuracy_train: 87.2%\n",
      "\t  loss_test: 496.71, accuracy_test: 75.2%\n",
      "Epoch: 533, loss_train: 1029.38, accuracy_train: 87.2%\n",
      "\t  loss_test: 496.62, accuracy_test: 75.2%\n",
      "Epoch: 534, loss_train: 1026.77, accuracy_train: 87.2%\n",
      "\t  loss_test: 496.52, accuracy_test: 75.2%\n",
      "Epoch: 535, loss_train: 1024.16, accuracy_train: 87.2%\n",
      "\t  loss_test: 496.43, accuracy_test: 75.2%\n",
      "Epoch: 536, loss_train: 1021.55, accuracy_train: 87.2%\n",
      "\t  loss_test: 496.34, accuracy_test: 75.2%\n",
      "Epoch: 537, loss_train: 1018.94, accuracy_train: 87.2%\n",
      "\t  loss_test: 496.25, accuracy_test: 75.3%\n",
      "Epoch: 538, loss_train: 1016.34, accuracy_train: 87.3%\n",
      "\t  loss_test: 496.16, accuracy_test: 75.3%\n",
      "Epoch: 539, loss_train: 1013.73, accuracy_train: 87.3%\n",
      "\t  loss_test: 496.07, accuracy_test: 75.4%\n",
      "Epoch: 540, loss_train: 1011.13, accuracy_train: 87.3%\n",
      "\t  loss_test: 495.97, accuracy_test: 75.4%\n",
      "Epoch: 541, loss_train: 1008.52, accuracy_train: 87.3%\n",
      "\t  loss_test: 495.88, accuracy_test: 75.4%\n",
      "Epoch: 542, loss_train: 1005.92, accuracy_train: 87.3%\n",
      "\t  loss_test: 495.79, accuracy_test: 75.6%\n",
      "Epoch: 543, loss_train: 1003.32, accuracy_train: 87.6%\n",
      "\t  loss_test: 495.70, accuracy_test: 75.6%\n",
      "Epoch: 544, loss_train: 1000.72, accuracy_train: 87.6%\n",
      "\t  loss_test: 495.60, accuracy_test: 75.6%\n",
      "Epoch: 545, loss_train: 998.12, accuracy_train: 87.8%\n",
      "\t  loss_test: 495.51, accuracy_test: 75.6%\n",
      "Epoch: 546, loss_train: 995.52, accuracy_train: 87.9%\n",
      "\t  loss_test: 495.42, accuracy_test: 75.6%\n",
      "Epoch: 547, loss_train: 992.93, accuracy_train: 87.9%\n",
      "\t  loss_test: 495.32, accuracy_test: 75.6%\n",
      "Epoch: 548, loss_train: 990.33, accuracy_train: 87.9%\n",
      "\t  loss_test: 495.23, accuracy_test: 75.6%\n",
      "Epoch: 549, loss_train: 987.74, accuracy_train: 88.0%\n",
      "\t  loss_test: 495.14, accuracy_test: 75.6%\n",
      "Epoch: 550, loss_train: 985.14, accuracy_train: 88.1%\n",
      "\t  loss_test: 495.04, accuracy_test: 75.6%\n",
      "Epoch: 551, loss_train: 982.55, accuracy_train: 88.1%\n",
      "\t  loss_test: 494.95, accuracy_test: 75.6%\n",
      "Epoch: 552, loss_train: 979.96, accuracy_train: 88.1%\n",
      "\t  loss_test: 494.86, accuracy_test: 75.6%\n",
      "Epoch: 553, loss_train: 977.38, accuracy_train: 88.2%\n",
      "\t  loss_test: 494.76, accuracy_test: 75.6%\n",
      "Epoch: 554, loss_train: 974.79, accuracy_train: 88.2%\n",
      "\t  loss_test: 494.67, accuracy_test: 75.6%\n",
      "Epoch: 555, loss_train: 972.20, accuracy_train: 88.2%\n",
      "\t  loss_test: 494.58, accuracy_test: 75.6%\n",
      "Epoch: 556, loss_train: 969.62, accuracy_train: 88.4%\n",
      "\t  loss_test: 494.48, accuracy_test: 75.6%\n",
      "Epoch: 557, loss_train: 967.04, accuracy_train: 88.4%\n",
      "\t  loss_test: 494.39, accuracy_test: 75.7%\n",
      "Epoch: 558, loss_train: 964.46, accuracy_train: 88.5%\n",
      "\t  loss_test: 494.30, accuracy_test: 75.7%\n",
      "Epoch: 559, loss_train: 961.88, accuracy_train: 88.5%\n",
      "\t  loss_test: 494.20, accuracy_test: 75.7%\n",
      "Epoch: 560, loss_train: 959.30, accuracy_train: 88.5%\n",
      "\t  loss_test: 494.11, accuracy_test: 75.7%\n",
      "Epoch: 561, loss_train: 956.73, accuracy_train: 88.6%\n",
      "\t  loss_test: 494.01, accuracy_test: 75.7%\n",
      "Epoch: 562, loss_train: 954.16, accuracy_train: 88.6%\n",
      "\t  loss_test: 493.92, accuracy_test: 75.8%\n",
      "Epoch: 563, loss_train: 951.59, accuracy_train: 88.7%\n",
      "\t  loss_test: 493.83, accuracy_test: 75.8%\n",
      "Epoch: 564, loss_train: 949.02, accuracy_train: 88.7%\n",
      "\t  loss_test: 493.73, accuracy_test: 75.8%\n",
      "Epoch: 565, loss_train: 946.45, accuracy_train: 88.8%\n",
      "\t  loss_test: 493.64, accuracy_test: 75.8%\n",
      "Epoch: 566, loss_train: 943.89, accuracy_train: 88.8%\n",
      "\t  loss_test: 493.55, accuracy_test: 75.8%\n",
      "Epoch: 567, loss_train: 941.33, accuracy_train: 88.8%\n",
      "\t  loss_test: 493.45, accuracy_test: 75.8%\n",
      "Epoch: 568, loss_train: 938.77, accuracy_train: 88.8%\n",
      "\t  loss_test: 493.36, accuracy_test: 75.8%\n",
      "Epoch: 569, loss_train: 936.21, accuracy_train: 88.9%\n",
      "\t  loss_test: 493.26, accuracy_test: 75.8%\n",
      "Epoch: 570, loss_train: 933.65, accuracy_train: 88.9%\n",
      "\t  loss_test: 493.17, accuracy_test: 75.8%\n",
      "Epoch: 571, loss_train: 931.10, accuracy_train: 88.9%\n",
      "\t  loss_test: 493.08, accuracy_test: 75.8%\n",
      "Epoch: 572, loss_train: 928.55, accuracy_train: 89.0%\n",
      "\t  loss_test: 492.99, accuracy_test: 75.8%\n",
      "Epoch: 573, loss_train: 926.00, accuracy_train: 89.1%\n",
      "\t  loss_test: 492.89, accuracy_test: 75.8%\n",
      "Epoch: 574, loss_train: 923.46, accuracy_train: 89.2%\n",
      "\t  loss_test: 492.80, accuracy_test: 75.8%\n",
      "Epoch: 575, loss_train: 920.92, accuracy_train: 89.3%\n",
      "\t  loss_test: 492.71, accuracy_test: 75.8%\n",
      "Epoch: 576, loss_train: 918.38, accuracy_train: 89.4%\n",
      "\t  loss_test: 492.61, accuracy_test: 75.8%\n",
      "Epoch: 577, loss_train: 915.84, accuracy_train: 89.4%\n",
      "\t  loss_test: 492.52, accuracy_test: 75.8%\n",
      "Epoch: 578, loss_train: 913.31, accuracy_train: 89.4%\n",
      "\t  loss_test: 492.43, accuracy_test: 75.8%\n",
      "Epoch: 579, loss_train: 910.78, accuracy_train: 89.4%\n",
      "\t  loss_test: 492.34, accuracy_test: 75.8%\n",
      "Epoch: 580, loss_train: 908.26, accuracy_train: 89.4%\n",
      "\t  loss_test: 492.24, accuracy_test: 75.8%\n",
      "Epoch: 581, loss_train: 905.73, accuracy_train: 89.5%\n",
      "\t  loss_test: 492.15, accuracy_test: 76.0%\n",
      "Epoch: 582, loss_train: 903.21, accuracy_train: 89.5%\n",
      "\t  loss_test: 492.06, accuracy_test: 76.0%\n",
      "Epoch: 583, loss_train: 900.70, accuracy_train: 89.7%\n",
      "\t  loss_test: 491.97, accuracy_test: 76.2%\n",
      "Epoch: 584, loss_train: 898.18, accuracy_train: 89.8%\n",
      "\t  loss_test: 491.88, accuracy_test: 76.2%\n",
      "Epoch: 585, loss_train: 895.67, accuracy_train: 89.8%\n",
      "\t  loss_test: 491.79, accuracy_test: 76.2%\n",
      "Epoch: 586, loss_train: 893.17, accuracy_train: 89.8%\n",
      "\t  loss_test: 491.70, accuracy_test: 76.2%\n",
      "Epoch: 587, loss_train: 890.66, accuracy_train: 89.8%\n",
      "\t  loss_test: 491.61, accuracy_test: 76.3%\n",
      "Epoch: 588, loss_train: 888.17, accuracy_train: 89.9%\n",
      "\t  loss_test: 491.52, accuracy_test: 76.3%\n",
      "Epoch: 589, loss_train: 885.67, accuracy_train: 89.9%\n",
      "\t  loss_test: 491.43, accuracy_test: 76.3%\n",
      "Epoch: 590, loss_train: 883.18, accuracy_train: 90.0%\n",
      "\t  loss_test: 491.34, accuracy_test: 76.3%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 591, loss_train: 880.69, accuracy_train: 90.1%\n",
      "\t  loss_test: 491.25, accuracy_test: 76.3%\n",
      "Epoch: 592, loss_train: 878.21, accuracy_train: 90.2%\n",
      "\t  loss_test: 491.16, accuracy_test: 76.3%\n",
      "Epoch: 593, loss_train: 875.73, accuracy_train: 90.3%\n",
      "\t  loss_test: 491.07, accuracy_test: 76.3%\n",
      "Epoch: 594, loss_train: 873.26, accuracy_train: 90.4%\n",
      "\t  loss_test: 490.99, accuracy_test: 76.3%\n",
      "Epoch: 595, loss_train: 870.79, accuracy_train: 90.4%\n",
      "\t  loss_test: 490.90, accuracy_test: 76.3%\n",
      "Epoch: 596, loss_train: 868.32, accuracy_train: 90.4%\n",
      "\t  loss_test: 490.81, accuracy_test: 76.3%\n",
      "Epoch: 597, loss_train: 865.86, accuracy_train: 90.5%\n",
      "\t  loss_test: 490.72, accuracy_test: 76.3%\n",
      "Epoch: 598, loss_train: 863.40, accuracy_train: 90.6%\n",
      "\t  loss_test: 490.64, accuracy_test: 76.5%\n",
      "Epoch: 599, loss_train: 860.95, accuracy_train: 90.7%\n",
      "\t  loss_test: 490.55, accuracy_test: 76.5%\n",
      "Epoch: 600, loss_train: 858.50, accuracy_train: 90.7%\n",
      "\t  loss_test: 490.47, accuracy_test: 76.5%\n",
      "Epoch: 601, loss_train: 856.06, accuracy_train: 90.7%\n",
      "\t  loss_test: 490.38, accuracy_test: 76.5%\n",
      "Epoch: 602, loss_train: 853.62, accuracy_train: 90.8%\n",
      "\t  loss_test: 490.30, accuracy_test: 76.5%\n",
      "Epoch: 603, loss_train: 851.19, accuracy_train: 90.8%\n",
      "\t  loss_test: 490.21, accuracy_test: 76.5%\n",
      "Epoch: 604, loss_train: 848.76, accuracy_train: 90.8%\n",
      "\t  loss_test: 490.13, accuracy_test: 76.5%\n",
      "Epoch: 605, loss_train: 846.33, accuracy_train: 90.8%\n",
      "\t  loss_test: 490.05, accuracy_test: 76.5%\n",
      "Epoch: 606, loss_train: 843.92, accuracy_train: 90.8%\n",
      "\t  loss_test: 489.96, accuracy_test: 76.5%\n",
      "Epoch: 607, loss_train: 841.50, accuracy_train: 90.9%\n",
      "\t  loss_test: 489.88, accuracy_test: 76.5%\n",
      "Epoch: 608, loss_train: 839.09, accuracy_train: 90.9%\n",
      "\t  loss_test: 489.80, accuracy_test: 76.5%\n",
      "Epoch: 609, loss_train: 836.69, accuracy_train: 91.1%\n",
      "\t  loss_test: 489.72, accuracy_test: 76.5%\n",
      "Epoch: 610, loss_train: 834.29, accuracy_train: 91.1%\n",
      "\t  loss_test: 489.64, accuracy_test: 76.5%\n",
      "Epoch: 611, loss_train: 831.90, accuracy_train: 91.3%\n",
      "\t  loss_test: 489.56, accuracy_test: 76.5%\n",
      "Epoch: 612, loss_train: 829.51, accuracy_train: 91.3%\n",
      "\t  loss_test: 489.48, accuracy_test: 76.5%\n",
      "Epoch: 613, loss_train: 827.13, accuracy_train: 91.5%\n",
      "\t  loss_test: 489.40, accuracy_test: 76.7%\n",
      "Epoch: 614, loss_train: 824.76, accuracy_train: 91.5%\n",
      "\t  loss_test: 489.32, accuracy_test: 76.7%\n",
      "Epoch: 615, loss_train: 822.39, accuracy_train: 91.5%\n",
      "\t  loss_test: 489.24, accuracy_test: 76.7%\n",
      "Epoch: 616, loss_train: 820.02, accuracy_train: 91.6%\n",
      "\t  loss_test: 489.16, accuracy_test: 76.7%\n",
      "Epoch: 617, loss_train: 817.67, accuracy_train: 91.6%\n",
      "\t  loss_test: 489.09, accuracy_test: 76.7%\n",
      "Epoch: 618, loss_train: 815.31, accuracy_train: 91.7%\n",
      "\t  loss_test: 489.01, accuracy_test: 76.7%\n",
      "Epoch: 619, loss_train: 812.97, accuracy_train: 91.7%\n",
      "\t  loss_test: 488.94, accuracy_test: 76.7%\n",
      "Epoch: 620, loss_train: 810.63, accuracy_train: 91.7%\n",
      "\t  loss_test: 488.86, accuracy_test: 76.7%\n",
      "Epoch: 621, loss_train: 808.29, accuracy_train: 91.8%\n",
      "\t  loss_test: 488.79, accuracy_test: 76.6%\n",
      "Epoch: 622, loss_train: 805.97, accuracy_train: 91.8%\n",
      "\t  loss_test: 488.71, accuracy_test: 76.6%\n",
      "Epoch: 623, loss_train: 803.64, accuracy_train: 91.8%\n",
      "\t  loss_test: 488.64, accuracy_test: 76.6%\n",
      "Epoch: 624, loss_train: 801.33, accuracy_train: 91.8%\n",
      "\t  loss_test: 488.57, accuracy_test: 76.6%\n",
      "Epoch: 625, loss_train: 799.02, accuracy_train: 91.9%\n",
      "\t  loss_test: 488.50, accuracy_test: 76.6%\n",
      "Epoch: 626, loss_train: 796.72, accuracy_train: 92.0%\n",
      "\t  loss_test: 488.43, accuracy_test: 76.5%\n",
      "Epoch: 627, loss_train: 794.42, accuracy_train: 92.1%\n",
      "\t  loss_test: 488.36, accuracy_test: 76.5%\n",
      "Epoch: 628, loss_train: 792.13, accuracy_train: 92.3%\n",
      "\t  loss_test: 488.29, accuracy_test: 76.5%\n",
      "Epoch: 629, loss_train: 789.85, accuracy_train: 92.3%\n",
      "\t  loss_test: 488.22, accuracy_test: 76.5%\n",
      "Epoch: 630, loss_train: 787.57, accuracy_train: 92.4%\n",
      "\t  loss_test: 488.15, accuracy_test: 76.3%\n",
      "Epoch: 631, loss_train: 785.30, accuracy_train: 92.4%\n",
      "\t  loss_test: 488.08, accuracy_test: 76.3%\n",
      "Epoch: 632, loss_train: 783.03, accuracy_train: 92.4%\n",
      "\t  loss_test: 488.01, accuracy_test: 76.3%\n",
      "Epoch: 633, loss_train: 780.78, accuracy_train: 92.4%\n",
      "\t  loss_test: 487.95, accuracy_test: 76.3%\n",
      "Epoch: 634, loss_train: 778.53, accuracy_train: 92.4%\n",
      "\t  loss_test: 487.88, accuracy_test: 76.3%\n",
      "Epoch: 635, loss_train: 776.28, accuracy_train: 92.5%\n",
      "\t  loss_test: 487.82, accuracy_test: 76.2%\n",
      "Epoch: 636, loss_train: 774.05, accuracy_train: 92.7%\n",
      "\t  loss_test: 487.75, accuracy_test: 76.2%\n",
      "Epoch: 637, loss_train: 771.82, accuracy_train: 92.7%\n",
      "\t  loss_test: 487.69, accuracy_test: 76.2%\n",
      "Epoch: 638, loss_train: 769.60, accuracy_train: 92.7%\n",
      "\t  loss_test: 487.63, accuracy_test: 76.2%\n",
      "Epoch: 639, loss_train: 767.38, accuracy_train: 92.8%\n",
      "\t  loss_test: 487.57, accuracy_test: 76.2%\n",
      "Epoch: 640, loss_train: 765.17, accuracy_train: 92.9%\n",
      "\t  loss_test: 487.51, accuracy_test: 76.2%\n",
      "Epoch: 641, loss_train: 762.97, accuracy_train: 93.0%\n",
      "\t  loss_test: 487.45, accuracy_test: 76.2%\n",
      "Epoch: 642, loss_train: 760.78, accuracy_train: 93.0%\n",
      "\t  loss_test: 487.39, accuracy_test: 76.2%\n",
      "Epoch: 643, loss_train: 758.59, accuracy_train: 93.0%\n",
      "\t  loss_test: 487.33, accuracy_test: 76.3%\n",
      "Epoch: 644, loss_train: 756.41, accuracy_train: 93.0%\n",
      "\t  loss_test: 487.27, accuracy_test: 76.3%\n",
      "Epoch: 645, loss_train: 754.24, accuracy_train: 93.0%\n",
      "\t  loss_test: 487.21, accuracy_test: 76.5%\n",
      "Epoch: 646, loss_train: 752.07, accuracy_train: 93.0%\n",
      "\t  loss_test: 487.16, accuracy_test: 76.5%\n",
      "Epoch: 647, loss_train: 749.91, accuracy_train: 93.0%\n",
      "\t  loss_test: 487.10, accuracy_test: 76.5%\n",
      "Epoch: 648, loss_train: 747.76, accuracy_train: 93.0%\n",
      "\t  loss_test: 487.05, accuracy_test: 76.5%\n",
      "Epoch: 649, loss_train: 745.62, accuracy_train: 93.1%\n",
      "\t  loss_test: 486.99, accuracy_test: 76.6%\n",
      "Epoch: 650, loss_train: 743.48, accuracy_train: 93.1%\n",
      "\t  loss_test: 486.94, accuracy_test: 76.7%\n",
      "Epoch: 651, loss_train: 741.35, accuracy_train: 93.2%\n",
      "\t  loss_test: 486.89, accuracy_test: 76.7%\n",
      "Epoch: 652, loss_train: 739.23, accuracy_train: 93.2%\n",
      "\t  loss_test: 486.84, accuracy_test: 76.7%\n",
      "Epoch: 653, loss_train: 737.12, accuracy_train: 93.3%\n",
      "\t  loss_test: 486.79, accuracy_test: 76.7%\n",
      "Epoch: 654, loss_train: 735.01, accuracy_train: 93.3%\n",
      "\t  loss_test: 486.74, accuracy_test: 76.7%\n",
      "Epoch: 655, loss_train: 732.92, accuracy_train: 93.4%\n",
      "\t  loss_test: 486.69, accuracy_test: 76.7%\n",
      "Epoch: 656, loss_train: 730.82, accuracy_train: 93.4%\n",
      "\t  loss_test: 486.64, accuracy_test: 76.7%\n",
      "Epoch: 657, loss_train: 728.74, accuracy_train: 93.4%\n",
      "\t  loss_test: 486.59, accuracy_test: 76.7%\n",
      "Epoch: 658, loss_train: 726.67, accuracy_train: 93.5%\n",
      "\t  loss_test: 486.55, accuracy_test: 76.7%\n",
      "Epoch: 659, loss_train: 724.60, accuracy_train: 93.5%\n",
      "\t  loss_test: 486.50, accuracy_test: 76.6%\n",
      "Epoch: 660, loss_train: 722.54, accuracy_train: 93.5%\n",
      "\t  loss_test: 486.46, accuracy_test: 76.6%\n",
      "Epoch: 661, loss_train: 720.49, accuracy_train: 93.5%\n",
      "\t  loss_test: 486.41, accuracy_test: 76.6%\n",
      "Epoch: 662, loss_train: 718.44, accuracy_train: 93.5%\n",
      "\t  loss_test: 486.37, accuracy_test: 76.6%\n",
      "Epoch: 663, loss_train: 716.40, accuracy_train: 93.6%\n",
      "\t  loss_test: 486.33, accuracy_test: 76.6%\n",
      "Epoch: 664, loss_train: 714.38, accuracy_train: 93.6%\n",
      "\t  loss_test: 486.29, accuracy_test: 76.6%\n",
      "Epoch: 665, loss_train: 712.35, accuracy_train: 93.7%\n",
      "\t  loss_test: 486.25, accuracy_test: 76.6%\n",
      "Epoch: 666, loss_train: 710.34, accuracy_train: 93.7%\n",
      "\t  loss_test: 486.21, accuracy_test: 76.6%\n",
      "Epoch: 667, loss_train: 708.34, accuracy_train: 93.7%\n",
      "\t  loss_test: 486.17, accuracy_test: 76.6%\n",
      "Epoch: 668, loss_train: 706.34, accuracy_train: 93.7%\n",
      "\t  loss_test: 486.13, accuracy_test: 76.6%\n",
      "Epoch: 669, loss_train: 704.35, accuracy_train: 93.8%\n",
      "\t  loss_test: 486.09, accuracy_test: 76.6%\n",
      "Epoch: 670, loss_train: 702.37, accuracy_train: 93.8%\n",
      "\t  loss_test: 486.06, accuracy_test: 76.6%\n",
      "Epoch: 671, loss_train: 700.39, accuracy_train: 93.9%\n",
      "\t  loss_test: 486.02, accuracy_test: 76.6%\n",
      "Epoch: 672, loss_train: 698.43, accuracy_train: 93.9%\n",
      "\t  loss_test: 485.99, accuracy_test: 76.5%\n",
      "Epoch: 673, loss_train: 696.47, accuracy_train: 93.9%\n",
      "\t  loss_test: 485.95, accuracy_test: 76.5%\n",
      "Epoch: 674, loss_train: 694.52, accuracy_train: 93.9%\n",
      "\t  loss_test: 485.92, accuracy_test: 76.5%\n",
      "Epoch: 675, loss_train: 692.58, accuracy_train: 94.0%\n",
      "\t  loss_test: 485.89, accuracy_test: 76.5%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 676, loss_train: 690.64, accuracy_train: 94.0%\n",
      "\t  loss_test: 485.86, accuracy_test: 76.5%\n",
      "Epoch: 677, loss_train: 688.71, accuracy_train: 94.1%\n",
      "\t  loss_test: 485.83, accuracy_test: 76.5%\n",
      "Epoch: 678, loss_train: 686.80, accuracy_train: 94.2%\n",
      "\t  loss_test: 485.80, accuracy_test: 76.5%\n",
      "Epoch: 679, loss_train: 684.89, accuracy_train: 94.2%\n",
      "\t  loss_test: 485.77, accuracy_test: 76.5%\n",
      "Epoch: 680, loss_train: 682.98, accuracy_train: 94.2%\n",
      "\t  loss_test: 485.74, accuracy_test: 76.5%\n",
      "Epoch: 681, loss_train: 681.09, accuracy_train: 94.2%\n",
      "\t  loss_test: 485.72, accuracy_test: 76.6%\n",
      "Epoch: 682, loss_train: 679.20, accuracy_train: 94.2%\n",
      "\t  loss_test: 485.69, accuracy_test: 76.6%\n",
      "Epoch: 683, loss_train: 677.32, accuracy_train: 94.3%\n",
      "\t  loss_test: 485.67, accuracy_test: 76.5%\n",
      "Epoch: 684, loss_train: 675.45, accuracy_train: 94.3%\n",
      "\t  loss_test: 485.64, accuracy_test: 76.6%\n",
      "Epoch: 685, loss_train: 673.59, accuracy_train: 94.4%\n",
      "\t  loss_test: 485.62, accuracy_test: 76.6%\n",
      "Epoch: 686, loss_train: 671.74, accuracy_train: 94.5%\n",
      "\t  loss_test: 485.60, accuracy_test: 76.6%\n",
      "Epoch: 687, loss_train: 669.89, accuracy_train: 94.5%\n",
      "\t  loss_test: 485.58, accuracy_test: 76.6%\n",
      "Epoch: 688, loss_train: 668.05, accuracy_train: 94.5%\n",
      "\t  loss_test: 485.56, accuracy_test: 76.6%\n",
      "Epoch: 689, loss_train: 666.22, accuracy_train: 94.5%\n",
      "\t  loss_test: 485.54, accuracy_test: 76.7%\n",
      "Epoch: 690, loss_train: 664.40, accuracy_train: 94.5%\n",
      "\t  loss_test: 485.52, accuracy_test: 76.7%\n",
      "Epoch: 691, loss_train: 662.58, accuracy_train: 94.5%\n",
      "\t  loss_test: 485.50, accuracy_test: 76.7%\n",
      "Epoch: 692, loss_train: 660.78, accuracy_train: 94.6%\n",
      "\t  loss_test: 485.48, accuracy_test: 76.7%\n",
      "Epoch: 693, loss_train: 658.98, accuracy_train: 94.6%\n",
      "\t  loss_test: 485.47, accuracy_test: 76.7%\n",
      "Epoch: 694, loss_train: 657.19, accuracy_train: 94.6%\n",
      "\t  loss_test: 485.45, accuracy_test: 76.7%\n",
      "Epoch: 695, loss_train: 655.40, accuracy_train: 94.6%\n",
      "\t  loss_test: 485.44, accuracy_test: 76.7%\n",
      "Epoch: 696, loss_train: 653.63, accuracy_train: 94.6%\n",
      "\t  loss_test: 485.42, accuracy_test: 76.7%\n",
      "Epoch: 697, loss_train: 651.86, accuracy_train: 94.6%\n",
      "\t  loss_test: 485.41, accuracy_test: 76.7%\n",
      "Epoch: 698, loss_train: 650.10, accuracy_train: 94.6%\n",
      "\t  loss_test: 485.40, accuracy_test: 76.7%\n",
      "Epoch: 699, loss_train: 648.35, accuracy_train: 94.7%\n",
      "\t  loss_test: 485.39, accuracy_test: 76.7%\n",
      "Epoch: 700, loss_train: 646.61, accuracy_train: 94.7%\n",
      "\t  loss_test: 485.38, accuracy_test: 76.7%\n",
      "Epoch: 701, loss_train: 644.87, accuracy_train: 94.7%\n",
      "\t  loss_test: 485.37, accuracy_test: 76.7%\n",
      "Epoch: 702, loss_train: 643.15, accuracy_train: 94.7%\n",
      "\t  loss_test: 485.36, accuracy_test: 76.7%\n",
      "Epoch: 703, loss_train: 641.43, accuracy_train: 94.7%\n",
      "\t  loss_test: 485.35, accuracy_test: 76.7%\n",
      "Epoch: 704, loss_train: 639.71, accuracy_train: 94.8%\n",
      "\t  loss_test: 485.35, accuracy_test: 76.7%\n",
      "Epoch: 705, loss_train: 638.01, accuracy_train: 94.9%\n",
      "\t  loss_test: 485.34, accuracy_test: 76.7%\n",
      "Epoch: 706, loss_train: 636.32, accuracy_train: 94.9%\n",
      "\t  loss_test: 485.34, accuracy_test: 76.9%\n",
      "Epoch: 707, loss_train: 634.63, accuracy_train: 94.9%\n",
      "\t  loss_test: 485.33, accuracy_test: 76.9%\n",
      "Epoch: 708, loss_train: 632.95, accuracy_train: 94.9%\n",
      "\t  loss_test: 485.33, accuracy_test: 76.9%\n",
      "Epoch: 709, loss_train: 631.27, accuracy_train: 94.9%\n",
      "\t  loss_test: 485.33, accuracy_test: 76.9%\n",
      "Epoch: 710, loss_train: 629.61, accuracy_train: 94.9%\n",
      "\t  loss_test: 485.33, accuracy_test: 76.7%\n",
      "Epoch: 711, loss_train: 627.95, accuracy_train: 94.9%\n",
      "\t  loss_test: 485.33, accuracy_test: 76.7%\n",
      "Epoch: 712, loss_train: 626.30, accuracy_train: 94.9%\n",
      "\t  loss_test: 485.33, accuracy_test: 76.7%\n",
      "Epoch: 713, loss_train: 624.66, accuracy_train: 95.0%\n",
      "\t  loss_test: 485.33, accuracy_test: 76.7%\n",
      "Epoch: 714, loss_train: 623.03, accuracy_train: 95.0%\n",
      "\t  loss_test: 485.33, accuracy_test: 76.7%\n",
      "Epoch: 715, loss_train: 621.40, accuracy_train: 95.0%\n",
      "\t  loss_test: 485.33, accuracy_test: 76.7%\n",
      "Epoch: 716, loss_train: 619.79, accuracy_train: 95.1%\n",
      "\t  loss_test: 485.34, accuracy_test: 76.7%\n",
      "Epoch: 717, loss_train: 618.18, accuracy_train: 95.1%\n",
      "\t  loss_test: 485.34, accuracy_test: 76.7%\n",
      "Epoch: 718, loss_train: 616.57, accuracy_train: 95.1%\n",
      "\t  loss_test: 485.35, accuracy_test: 76.7%\n",
      "Epoch: 719, loss_train: 614.98, accuracy_train: 95.1%\n",
      "\t  loss_test: 485.35, accuracy_test: 76.7%\n",
      "Epoch: 720, loss_train: 613.39, accuracy_train: 95.2%\n",
      "\t  loss_test: 485.36, accuracy_test: 76.7%\n",
      "Epoch: 721, loss_train: 611.81, accuracy_train: 95.2%\n",
      "\t  loss_test: 485.37, accuracy_test: 76.7%\n",
      "Epoch: 722, loss_train: 610.24, accuracy_train: 95.2%\n",
      "\t  loss_test: 485.38, accuracy_test: 76.6%\n",
      "Epoch: 723, loss_train: 608.67, accuracy_train: 95.3%\n",
      "\t  loss_test: 485.39, accuracy_test: 76.6%\n",
      "Epoch: 724, loss_train: 607.12, accuracy_train: 95.3%\n",
      "\t  loss_test: 485.40, accuracy_test: 76.6%\n",
      "Epoch: 725, loss_train: 605.57, accuracy_train: 95.3%\n",
      "\t  loss_test: 485.41, accuracy_test: 76.6%\n",
      "Epoch: 726, loss_train: 604.03, accuracy_train: 95.3%\n",
      "\t  loss_test: 485.42, accuracy_test: 76.6%\n",
      "Epoch: 727, loss_train: 602.49, accuracy_train: 95.3%\n",
      "\t  loss_test: 485.43, accuracy_test: 76.6%\n",
      "Epoch: 728, loss_train: 600.96, accuracy_train: 95.3%\n",
      "\t  loss_test: 485.45, accuracy_test: 76.5%\n",
      "Epoch: 729, loss_train: 599.44, accuracy_train: 95.3%\n",
      "\t  loss_test: 485.46, accuracy_test: 76.5%\n",
      "Epoch: 730, loss_train: 597.93, accuracy_train: 95.3%\n",
      "\t  loss_test: 485.48, accuracy_test: 76.5%\n",
      "Epoch: 731, loss_train: 596.43, accuracy_train: 95.3%\n",
      "\t  loss_test: 485.49, accuracy_test: 76.5%\n",
      "Epoch: 732, loss_train: 594.93, accuracy_train: 95.3%\n",
      "\t  loss_test: 485.51, accuracy_test: 76.5%\n",
      "Epoch: 733, loss_train: 593.44, accuracy_train: 95.3%\n",
      "\t  loss_test: 485.53, accuracy_test: 76.5%\n",
      "Epoch: 734, loss_train: 591.96, accuracy_train: 95.4%\n",
      "\t  loss_test: 485.55, accuracy_test: 76.5%\n",
      "Epoch: 735, loss_train: 590.48, accuracy_train: 95.4%\n",
      "\t  loss_test: 485.57, accuracy_test: 76.5%\n",
      "Epoch: 736, loss_train: 589.01, accuracy_train: 95.5%\n",
      "\t  loss_test: 485.59, accuracy_test: 76.5%\n",
      "Epoch: 737, loss_train: 587.55, accuracy_train: 95.5%\n",
      "\t  loss_test: 485.61, accuracy_test: 76.5%\n",
      "Epoch: 738, loss_train: 586.10, accuracy_train: 95.5%\n",
      "\t  loss_test: 485.63, accuracy_test: 76.3%\n",
      "Epoch: 739, loss_train: 584.65, accuracy_train: 95.5%\n",
      "\t  loss_test: 485.65, accuracy_test: 76.3%\n",
      "Epoch: 740, loss_train: 583.21, accuracy_train: 95.5%\n",
      "\t  loss_test: 485.67, accuracy_test: 76.3%\n",
      "Epoch: 741, loss_train: 581.78, accuracy_train: 95.5%\n",
      "\t  loss_test: 485.70, accuracy_test: 76.3%\n",
      "Epoch: 742, loss_train: 580.35, accuracy_train: 95.5%\n",
      "\t  loss_test: 485.72, accuracy_test: 76.3%\n",
      "Epoch: 743, loss_train: 578.94, accuracy_train: 95.6%\n",
      "\t  loss_test: 485.75, accuracy_test: 76.5%\n",
      "Epoch: 744, loss_train: 577.53, accuracy_train: 95.6%\n",
      "\t  loss_test: 485.77, accuracy_test: 76.5%\n",
      "Epoch: 745, loss_train: 576.12, accuracy_train: 95.6%\n",
      "\t  loss_test: 485.80, accuracy_test: 76.5%\n",
      "Epoch: 746, loss_train: 574.72, accuracy_train: 95.6%\n",
      "\t  loss_test: 485.83, accuracy_test: 76.5%\n",
      "Epoch: 747, loss_train: 573.33, accuracy_train: 95.7%\n",
      "\t  loss_test: 485.86, accuracy_test: 76.5%\n",
      "Epoch: 748, loss_train: 571.95, accuracy_train: 95.7%\n",
      "\t  loss_test: 485.89, accuracy_test: 76.5%\n",
      "Epoch: 749, loss_train: 570.58, accuracy_train: 95.7%\n",
      "\t  loss_test: 485.92, accuracy_test: 76.5%\n",
      "Epoch: 750, loss_train: 569.21, accuracy_train: 95.7%\n",
      "\t  loss_test: 485.95, accuracy_test: 76.5%\n",
      "Epoch: 751, loss_train: 567.84, accuracy_train: 95.7%\n",
      "\t  loss_test: 485.98, accuracy_test: 76.5%\n",
      "Epoch: 752, loss_train: 566.49, accuracy_train: 95.7%\n",
      "\t  loss_test: 486.01, accuracy_test: 76.5%\n",
      "Epoch: 753, loss_train: 565.14, accuracy_train: 95.7%\n",
      "\t  loss_test: 486.04, accuracy_test: 76.5%\n",
      "Epoch: 754, loss_train: 563.80, accuracy_train: 95.7%\n",
      "\t  loss_test: 486.08, accuracy_test: 76.5%\n",
      "Epoch: 755, loss_train: 562.46, accuracy_train: 95.7%\n",
      "\t  loss_test: 486.11, accuracy_test: 76.5%\n",
      "Epoch: 756, loss_train: 561.14, accuracy_train: 95.7%\n",
      "\t  loss_test: 486.15, accuracy_test: 76.3%\n",
      "Epoch: 757, loss_train: 559.82, accuracy_train: 95.7%\n",
      "\t  loss_test: 486.18, accuracy_test: 76.3%\n",
      "Epoch: 758, loss_train: 558.50, accuracy_train: 95.7%\n",
      "\t  loss_test: 486.22, accuracy_test: 76.3%\n",
      "Epoch: 759, loss_train: 557.19, accuracy_train: 95.7%\n",
      "\t  loss_test: 486.26, accuracy_test: 76.3%\n",
      "Epoch: 760, loss_train: 555.89, accuracy_train: 95.7%\n",
      "\t  loss_test: 486.30, accuracy_test: 76.3%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 761, loss_train: 554.60, accuracy_train: 95.7%\n",
      "\t  loss_test: 486.33, accuracy_test: 76.3%\n",
      "Epoch: 762, loss_train: 553.31, accuracy_train: 95.7%\n",
      "\t  loss_test: 486.37, accuracy_test: 76.3%\n",
      "Epoch: 763, loss_train: 552.03, accuracy_train: 95.7%\n",
      "\t  loss_test: 486.41, accuracy_test: 76.3%\n",
      "Epoch: 764, loss_train: 550.75, accuracy_train: 95.7%\n",
      "\t  loss_test: 486.45, accuracy_test: 76.6%\n",
      "Epoch: 765, loss_train: 549.48, accuracy_train: 95.7%\n",
      "\t  loss_test: 486.50, accuracy_test: 76.6%\n",
      "Epoch: 766, loss_train: 548.22, accuracy_train: 95.7%\n",
      "\t  loss_test: 486.54, accuracy_test: 76.6%\n",
      "Epoch: 767, loss_train: 546.96, accuracy_train: 95.7%\n",
      "\t  loss_test: 486.58, accuracy_test: 76.6%\n",
      "Epoch: 768, loss_train: 545.72, accuracy_train: 95.8%\n",
      "\t  loss_test: 486.62, accuracy_test: 76.6%\n",
      "Epoch: 769, loss_train: 544.47, accuracy_train: 95.8%\n",
      "\t  loss_test: 486.67, accuracy_test: 76.6%\n",
      "Epoch: 770, loss_train: 543.24, accuracy_train: 95.8%\n",
      "\t  loss_test: 486.71, accuracy_test: 76.6%\n",
      "Epoch: 771, loss_train: 542.01, accuracy_train: 95.8%\n",
      "\t  loss_test: 486.76, accuracy_test: 76.6%\n",
      "Epoch: 772, loss_train: 540.78, accuracy_train: 95.8%\n",
      "\t  loss_test: 486.80, accuracy_test: 76.6%\n",
      "Epoch: 773, loss_train: 539.56, accuracy_train: 95.8%\n",
      "\t  loss_test: 486.85, accuracy_test: 76.5%\n",
      "Epoch: 774, loss_train: 538.35, accuracy_train: 95.8%\n",
      "\t  loss_test: 486.90, accuracy_test: 76.5%\n",
      "Epoch: 775, loss_train: 537.15, accuracy_train: 95.8%\n",
      "\t  loss_test: 486.95, accuracy_test: 76.5%\n",
      "Epoch: 776, loss_train: 535.95, accuracy_train: 95.8%\n",
      "\t  loss_test: 487.00, accuracy_test: 76.5%\n",
      "Epoch: 777, loss_train: 534.76, accuracy_train: 95.8%\n",
      "\t  loss_test: 487.04, accuracy_test: 76.6%\n",
      "Epoch: 778, loss_train: 533.57, accuracy_train: 95.8%\n",
      "\t  loss_test: 487.09, accuracy_test: 76.6%\n",
      "Epoch: 779, loss_train: 532.39, accuracy_train: 95.9%\n",
      "\t  loss_test: 487.15, accuracy_test: 76.6%\n",
      "Epoch: 780, loss_train: 531.21, accuracy_train: 95.9%\n",
      "\t  loss_test: 487.20, accuracy_test: 76.6%\n",
      "Epoch: 781, loss_train: 530.04, accuracy_train: 95.9%\n",
      "\t  loss_test: 487.25, accuracy_test: 76.6%\n",
      "Epoch: 782, loss_train: 528.88, accuracy_train: 95.9%\n",
      "\t  loss_test: 487.30, accuracy_test: 76.7%\n",
      "Epoch: 783, loss_train: 527.73, accuracy_train: 95.9%\n",
      "\t  loss_test: 487.35, accuracy_test: 76.7%\n",
      "Epoch: 784, loss_train: 526.57, accuracy_train: 95.9%\n",
      "\t  loss_test: 487.41, accuracy_test: 76.7%\n",
      "Epoch: 785, loss_train: 525.43, accuracy_train: 95.9%\n",
      "\t  loss_test: 487.46, accuracy_test: 76.7%\n",
      "Epoch: 786, loss_train: 524.29, accuracy_train: 95.9%\n",
      "\t  loss_test: 487.52, accuracy_test: 76.6%\n",
      "Epoch: 787, loss_train: 523.16, accuracy_train: 95.9%\n",
      "\t  loss_test: 487.57, accuracy_test: 76.6%\n",
      "Epoch: 788, loss_train: 522.03, accuracy_train: 95.9%\n",
      "\t  loss_test: 487.63, accuracy_test: 76.6%\n",
      "Epoch: 789, loss_train: 520.91, accuracy_train: 95.9%\n",
      "\t  loss_test: 487.69, accuracy_test: 76.6%\n",
      "Epoch: 790, loss_train: 519.79, accuracy_train: 96.0%\n",
      "\t  loss_test: 487.74, accuracy_test: 76.6%\n",
      "Epoch: 791, loss_train: 518.68, accuracy_train: 96.0%\n",
      "\t  loss_test: 487.80, accuracy_test: 76.6%\n",
      "Epoch: 792, loss_train: 517.58, accuracy_train: 96.0%\n",
      "\t  loss_test: 487.86, accuracy_test: 76.6%\n",
      "Epoch: 793, loss_train: 516.48, accuracy_train: 96.0%\n",
      "\t  loss_test: 487.92, accuracy_test: 76.6%\n",
      "Epoch: 794, loss_train: 515.39, accuracy_train: 96.0%\n",
      "\t  loss_test: 487.98, accuracy_test: 76.6%\n",
      "Epoch: 795, loss_train: 514.30, accuracy_train: 96.1%\n",
      "\t  loss_test: 488.04, accuracy_test: 76.6%\n",
      "Epoch: 796, loss_train: 513.22, accuracy_train: 96.1%\n",
      "\t  loss_test: 488.10, accuracy_test: 76.6%\n",
      "Epoch: 797, loss_train: 512.15, accuracy_train: 96.1%\n",
      "\t  loss_test: 488.16, accuracy_test: 76.6%\n",
      "Epoch: 798, loss_train: 511.08, accuracy_train: 96.1%\n",
      "\t  loss_test: 488.22, accuracy_test: 76.6%\n",
      "Epoch: 799, loss_train: 510.01, accuracy_train: 96.2%\n",
      "\t  loss_test: 488.29, accuracy_test: 76.6%\n",
      "Epoch: 800, loss_train: 508.95, accuracy_train: 96.2%\n",
      "\t  loss_test: 488.35, accuracy_test: 76.6%\n",
      "Epoch: 801, loss_train: 507.90, accuracy_train: 96.2%\n",
      "\t  loss_test: 488.41, accuracy_test: 76.6%\n",
      "Epoch: 802, loss_train: 506.85, accuracy_train: 96.2%\n",
      "\t  loss_test: 488.48, accuracy_test: 76.6%\n",
      "Epoch: 803, loss_train: 505.81, accuracy_train: 96.2%\n",
      "\t  loss_test: 488.54, accuracy_test: 76.6%\n",
      "Epoch: 804, loss_train: 504.77, accuracy_train: 96.2%\n",
      "\t  loss_test: 488.61, accuracy_test: 76.6%\n",
      "Epoch: 805, loss_train: 503.74, accuracy_train: 96.2%\n",
      "\t  loss_test: 488.67, accuracy_test: 76.6%\n",
      "Epoch: 806, loss_train: 502.71, accuracy_train: 96.3%\n",
      "\t  loss_test: 488.74, accuracy_test: 76.6%\n",
      "Epoch: 807, loss_train: 501.69, accuracy_train: 96.3%\n",
      "\t  loss_test: 488.81, accuracy_test: 76.7%\n",
      "Epoch: 808, loss_train: 500.67, accuracy_train: 96.3%\n",
      "\t  loss_test: 488.87, accuracy_test: 76.7%\n",
      "Epoch: 809, loss_train: 499.66, accuracy_train: 96.3%\n",
      "\t  loss_test: 488.94, accuracy_test: 76.6%\n",
      "Epoch: 810, loss_train: 498.66, accuracy_train: 96.3%\n",
      "\t  loss_test: 489.01, accuracy_test: 76.6%\n",
      "Epoch: 811, loss_train: 497.65, accuracy_train: 96.3%\n",
      "\t  loss_test: 489.08, accuracy_test: 76.6%\n",
      "Epoch: 812, loss_train: 496.66, accuracy_train: 96.3%\n",
      "\t  loss_test: 489.15, accuracy_test: 76.6%\n",
      "Epoch: 813, loss_train: 495.67, accuracy_train: 96.3%\n",
      "\t  loss_test: 489.22, accuracy_test: 76.6%\n",
      "Epoch: 814, loss_train: 494.68, accuracy_train: 96.3%\n",
      "\t  loss_test: 489.29, accuracy_test: 76.6%\n",
      "Epoch: 815, loss_train: 493.70, accuracy_train: 96.3%\n",
      "\t  loss_test: 489.36, accuracy_test: 76.6%\n",
      "Epoch: 816, loss_train: 492.73, accuracy_train: 96.3%\n",
      "\t  loss_test: 489.43, accuracy_test: 76.6%\n",
      "Epoch: 817, loss_train: 491.76, accuracy_train: 96.3%\n",
      "\t  loss_test: 489.51, accuracy_test: 76.6%\n",
      "Epoch: 818, loss_train: 490.79, accuracy_train: 96.3%\n",
      "\t  loss_test: 489.58, accuracy_test: 76.6%\n",
      "Epoch: 819, loss_train: 489.83, accuracy_train: 96.3%\n",
      "\t  loss_test: 489.65, accuracy_test: 76.6%\n",
      "Epoch: 820, loss_train: 488.88, accuracy_train: 96.3%\n",
      "\t  loss_test: 489.72, accuracy_test: 76.6%\n",
      "Epoch: 821, loss_train: 487.93, accuracy_train: 96.3%\n",
      "\t  loss_test: 489.80, accuracy_test: 76.6%\n",
      "Epoch: 822, loss_train: 486.98, accuracy_train: 96.3%\n",
      "\t  loss_test: 489.87, accuracy_test: 76.6%\n",
      "Epoch: 823, loss_train: 486.04, accuracy_train: 96.3%\n",
      "\t  loss_test: 489.95, accuracy_test: 76.6%\n",
      "Epoch: 824, loss_train: 485.11, accuracy_train: 96.3%\n",
      "\t  loss_test: 490.02, accuracy_test: 76.6%\n",
      "Epoch: 825, loss_train: 484.18, accuracy_train: 96.3%\n",
      "\t  loss_test: 490.10, accuracy_test: 76.6%\n",
      "Epoch: 826, loss_train: 483.25, accuracy_train: 96.3%\n",
      "\t  loss_test: 490.18, accuracy_test: 76.6%\n",
      "Epoch: 827, loss_train: 482.33, accuracy_train: 96.3%\n",
      "\t  loss_test: 490.25, accuracy_test: 76.6%\n",
      "Epoch: 828, loss_train: 481.41, accuracy_train: 96.3%\n",
      "\t  loss_test: 490.33, accuracy_test: 76.6%\n",
      "Epoch: 829, loss_train: 480.50, accuracy_train: 96.3%\n",
      "\t  loss_test: 490.41, accuracy_test: 76.6%\n",
      "Epoch: 830, loss_train: 479.60, accuracy_train: 96.3%\n",
      "\t  loss_test: 490.49, accuracy_test: 76.6%\n",
      "Epoch: 831, loss_train: 478.69, accuracy_train: 96.3%\n",
      "\t  loss_test: 490.56, accuracy_test: 76.6%\n",
      "Epoch: 832, loss_train: 477.80, accuracy_train: 96.3%\n",
      "\t  loss_test: 490.64, accuracy_test: 76.6%\n",
      "Epoch: 833, loss_train: 476.90, accuracy_train: 96.4%\n",
      "\t  loss_test: 490.72, accuracy_test: 76.6%\n",
      "Epoch: 834, loss_train: 476.02, accuracy_train: 96.4%\n",
      "\t  loss_test: 490.80, accuracy_test: 76.6%\n",
      "Epoch: 835, loss_train: 475.13, accuracy_train: 96.4%\n",
      "\t  loss_test: 490.88, accuracy_test: 76.6%\n",
      "Epoch: 836, loss_train: 474.25, accuracy_train: 96.4%\n",
      "\t  loss_test: 490.96, accuracy_test: 76.6%\n",
      "Epoch: 837, loss_train: 473.38, accuracy_train: 96.4%\n",
      "\t  loss_test: 491.05, accuracy_test: 76.5%\n",
      "Epoch: 838, loss_train: 472.51, accuracy_train: 96.4%\n",
      "\t  loss_test: 491.13, accuracy_test: 76.5%\n",
      "Epoch: 839, loss_train: 471.64, accuracy_train: 96.4%\n",
      "\t  loss_test: 491.21, accuracy_test: 76.5%\n",
      "Epoch: 840, loss_train: 470.78, accuracy_train: 96.4%\n",
      "\t  loss_test: 491.29, accuracy_test: 76.5%\n",
      "Epoch: 841, loss_train: 469.92, accuracy_train: 96.4%\n",
      "\t  loss_test: 491.38, accuracy_test: 76.5%\n",
      "Epoch: 842, loss_train: 469.07, accuracy_train: 96.4%\n",
      "\t  loss_test: 491.46, accuracy_test: 76.5%\n",
      "Epoch: 843, loss_train: 468.22, accuracy_train: 96.4%\n",
      "\t  loss_test: 491.54, accuracy_test: 76.5%\n",
      "Epoch: 844, loss_train: 467.38, accuracy_train: 96.4%\n",
      "\t  loss_test: 491.63, accuracy_test: 76.5%\n",
      "Epoch: 845, loss_train: 466.54, accuracy_train: 96.4%\n",
      "\t  loss_test: 491.71, accuracy_test: 76.5%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 846, loss_train: 465.71, accuracy_train: 96.4%\n",
      "\t  loss_test: 491.80, accuracy_test: 76.5%\n",
      "Epoch: 847, loss_train: 464.88, accuracy_train: 96.4%\n",
      "\t  loss_test: 491.88, accuracy_test: 76.5%\n",
      "Epoch: 848, loss_train: 464.05, accuracy_train: 96.4%\n",
      "\t  loss_test: 491.97, accuracy_test: 76.5%\n",
      "Epoch: 849, loss_train: 463.23, accuracy_train: 96.4%\n",
      "\t  loss_test: 492.05, accuracy_test: 76.5%\n",
      "Epoch: 850, loss_train: 462.41, accuracy_train: 96.4%\n",
      "\t  loss_test: 492.14, accuracy_test: 76.5%\n",
      "Epoch: 851, loss_train: 461.59, accuracy_train: 96.4%\n",
      "\t  loss_test: 492.23, accuracy_test: 76.5%\n",
      "Epoch: 852, loss_train: 460.79, accuracy_train: 96.4%\n",
      "\t  loss_test: 492.31, accuracy_test: 76.5%\n",
      "Epoch: 853, loss_train: 459.98, accuracy_train: 96.4%\n",
      "\t  loss_test: 492.40, accuracy_test: 76.5%\n",
      "Epoch: 854, loss_train: 459.18, accuracy_train: 96.4%\n",
      "\t  loss_test: 492.49, accuracy_test: 76.5%\n",
      "Epoch: 855, loss_train: 458.38, accuracy_train: 96.4%\n",
      "\t  loss_test: 492.58, accuracy_test: 76.5%\n",
      "Epoch: 856, loss_train: 457.59, accuracy_train: 96.4%\n",
      "\t  loss_test: 492.67, accuracy_test: 76.5%\n",
      "Epoch: 857, loss_train: 456.80, accuracy_train: 96.4%\n",
      "\t  loss_test: 492.76, accuracy_test: 76.5%\n",
      "Epoch: 858, loss_train: 456.01, accuracy_train: 96.4%\n",
      "\t  loss_test: 492.85, accuracy_test: 76.5%\n",
      "Epoch: 859, loss_train: 455.23, accuracy_train: 96.4%\n",
      "\t  loss_test: 492.94, accuracy_test: 76.5%\n",
      "Epoch: 860, loss_train: 454.46, accuracy_train: 96.4%\n",
      "\t  loss_test: 493.03, accuracy_test: 76.5%\n",
      "Epoch: 861, loss_train: 453.68, accuracy_train: 96.4%\n",
      "\t  loss_test: 493.12, accuracy_test: 76.5%\n",
      "Epoch: 862, loss_train: 452.91, accuracy_train: 96.4%\n",
      "\t  loss_test: 493.21, accuracy_test: 76.5%\n",
      "Epoch: 863, loss_train: 452.15, accuracy_train: 96.4%\n",
      "\t  loss_test: 493.30, accuracy_test: 76.5%\n",
      "Epoch: 864, loss_train: 451.39, accuracy_train: 96.4%\n",
      "\t  loss_test: 493.39, accuracy_test: 76.5%\n",
      "Epoch: 865, loss_train: 450.63, accuracy_train: 96.4%\n",
      "\t  loss_test: 493.48, accuracy_test: 76.5%\n",
      "Epoch: 866, loss_train: 449.88, accuracy_train: 96.4%\n",
      "\t  loss_test: 493.58, accuracy_test: 76.5%\n",
      "Epoch: 867, loss_train: 449.13, accuracy_train: 96.4%\n",
      "\t  loss_test: 493.67, accuracy_test: 76.5%\n",
      "Epoch: 868, loss_train: 448.38, accuracy_train: 96.4%\n",
      "\t  loss_test: 493.76, accuracy_test: 76.3%\n",
      "Epoch: 869, loss_train: 447.64, accuracy_train: 96.4%\n",
      "\t  loss_test: 493.85, accuracy_test: 76.3%\n",
      "Epoch: 870, loss_train: 446.90, accuracy_train: 96.4%\n",
      "\t  loss_test: 493.95, accuracy_test: 76.3%\n",
      "Epoch: 871, loss_train: 446.17, accuracy_train: 96.4%\n",
      "\t  loss_test: 494.04, accuracy_test: 76.3%\n",
      "Epoch: 872, loss_train: 445.43, accuracy_train: 96.4%\n",
      "\t  loss_test: 494.14, accuracy_test: 76.3%\n",
      "Epoch: 873, loss_train: 444.71, accuracy_train: 96.4%\n",
      "\t  loss_test: 494.23, accuracy_test: 76.3%\n",
      "Epoch: 874, loss_train: 443.98, accuracy_train: 96.4%\n",
      "\t  loss_test: 494.33, accuracy_test: 76.3%\n",
      "Epoch: 875, loss_train: 443.26, accuracy_train: 96.4%\n",
      "\t  loss_test: 494.42, accuracy_test: 76.5%\n",
      "Epoch: 876, loss_train: 442.55, accuracy_train: 96.4%\n",
      "\t  loss_test: 494.52, accuracy_test: 76.5%\n",
      "Epoch: 877, loss_train: 441.84, accuracy_train: 96.4%\n",
      "\t  loss_test: 494.61, accuracy_test: 76.5%\n",
      "Epoch: 878, loss_train: 441.13, accuracy_train: 96.4%\n",
      "\t  loss_test: 494.71, accuracy_test: 76.5%\n",
      "Epoch: 879, loss_train: 440.42, accuracy_train: 96.4%\n",
      "\t  loss_test: 494.80, accuracy_test: 76.5%\n",
      "Epoch: 880, loss_train: 439.72, accuracy_train: 96.4%\n",
      "\t  loss_test: 494.90, accuracy_test: 76.5%\n",
      "Epoch: 881, loss_train: 439.02, accuracy_train: 96.4%\n",
      "\t  loss_test: 495.00, accuracy_test: 76.3%\n",
      "Epoch: 882, loss_train: 438.33, accuracy_train: 96.4%\n",
      "\t  loss_test: 495.10, accuracy_test: 76.2%\n",
      "Epoch: 883, loss_train: 437.64, accuracy_train: 96.4%\n",
      "\t  loss_test: 495.19, accuracy_test: 76.2%\n",
      "Epoch: 884, loss_train: 436.95, accuracy_train: 96.4%\n",
      "\t  loss_test: 495.29, accuracy_test: 76.2%\n",
      "Epoch: 885, loss_train: 436.26, accuracy_train: 96.4%\n",
      "\t  loss_test: 495.39, accuracy_test: 76.2%\n",
      "Epoch: 886, loss_train: 435.58, accuracy_train: 96.4%\n",
      "\t  loss_test: 495.49, accuracy_test: 76.2%\n",
      "Epoch: 887, loss_train: 434.91, accuracy_train: 96.4%\n",
      "\t  loss_test: 495.59, accuracy_test: 76.2%\n",
      "Epoch: 888, loss_train: 434.23, accuracy_train: 96.4%\n",
      "\t  loss_test: 495.69, accuracy_test: 76.2%\n",
      "Epoch: 889, loss_train: 433.56, accuracy_train: 96.4%\n",
      "\t  loss_test: 495.78, accuracy_test: 76.2%\n",
      "Epoch: 890, loss_train: 432.89, accuracy_train: 96.4%\n",
      "\t  loss_test: 495.88, accuracy_test: 76.2%\n",
      "Epoch: 891, loss_train: 432.23, accuracy_train: 96.4%\n",
      "\t  loss_test: 495.98, accuracy_test: 76.2%\n",
      "Epoch: 892, loss_train: 431.57, accuracy_train: 96.4%\n",
      "\t  loss_test: 496.08, accuracy_test: 76.2%\n",
      "Epoch: 893, loss_train: 430.91, accuracy_train: 96.4%\n",
      "\t  loss_test: 496.18, accuracy_test: 76.2%\n",
      "Epoch: 894, loss_train: 430.26, accuracy_train: 96.4%\n",
      "\t  loss_test: 496.28, accuracy_test: 76.2%\n",
      "Epoch: 895, loss_train: 429.61, accuracy_train: 96.4%\n",
      "\t  loss_test: 496.39, accuracy_test: 76.2%\n",
      "Epoch: 896, loss_train: 428.96, accuracy_train: 96.4%\n",
      "\t  loss_test: 496.49, accuracy_test: 76.2%\n",
      "Epoch: 897, loss_train: 428.32, accuracy_train: 96.4%\n",
      "\t  loss_test: 496.59, accuracy_test: 76.2%\n",
      "Epoch: 898, loss_train: 427.68, accuracy_train: 96.4%\n",
      "\t  loss_test: 496.69, accuracy_test: 76.2%\n",
      "Epoch: 899, loss_train: 427.04, accuracy_train: 96.4%\n",
      "\t  loss_test: 496.79, accuracy_test: 76.2%\n",
      "Epoch: 900, loss_train: 426.41, accuracy_train: 96.5%\n",
      "\t  loss_test: 496.89, accuracy_test: 76.3%\n",
      "Epoch: 901, loss_train: 425.77, accuracy_train: 96.5%\n",
      "\t  loss_test: 497.00, accuracy_test: 76.5%\n",
      "Epoch: 902, loss_train: 425.15, accuracy_train: 96.5%\n",
      "\t  loss_test: 497.10, accuracy_test: 76.6%\n",
      "Epoch: 903, loss_train: 424.52, accuracy_train: 96.5%\n",
      "\t  loss_test: 497.20, accuracy_test: 76.6%\n",
      "Epoch: 904, loss_train: 423.90, accuracy_train: 96.5%\n",
      "\t  loss_test: 497.30, accuracy_test: 76.6%\n",
      "Epoch: 905, loss_train: 423.28, accuracy_train: 96.5%\n",
      "\t  loss_test: 497.41, accuracy_test: 76.6%\n",
      "Epoch: 906, loss_train: 422.67, accuracy_train: 96.5%\n",
      "\t  loss_test: 497.51, accuracy_test: 76.7%\n",
      "Epoch: 907, loss_train: 422.05, accuracy_train: 96.5%\n",
      "\t  loss_test: 497.62, accuracy_test: 76.7%\n",
      "Epoch: 908, loss_train: 421.44, accuracy_train: 96.5%\n",
      "\t  loss_test: 497.72, accuracy_test: 76.7%\n",
      "Epoch: 909, loss_train: 420.84, accuracy_train: 96.5%\n",
      "\t  loss_test: 497.82, accuracy_test: 76.7%\n",
      "Epoch: 910, loss_train: 420.23, accuracy_train: 96.5%\n",
      "\t  loss_test: 497.93, accuracy_test: 76.7%\n",
      "Epoch: 911, loss_train: 419.63, accuracy_train: 96.5%\n",
      "\t  loss_test: 498.03, accuracy_test: 76.7%\n",
      "Epoch: 912, loss_train: 419.04, accuracy_train: 96.5%\n",
      "\t  loss_test: 498.14, accuracy_test: 76.7%\n",
      "Epoch: 913, loss_train: 418.44, accuracy_train: 96.5%\n",
      "\t  loss_test: 498.24, accuracy_test: 76.7%\n",
      "Epoch: 914, loss_train: 417.85, accuracy_train: 96.5%\n",
      "\t  loss_test: 498.35, accuracy_test: 76.7%\n",
      "Epoch: 915, loss_train: 417.26, accuracy_train: 96.5%\n",
      "\t  loss_test: 498.45, accuracy_test: 76.7%\n",
      "Epoch: 916, loss_train: 416.68, accuracy_train: 96.5%\n",
      "\t  loss_test: 498.56, accuracy_test: 76.9%\n",
      "Epoch: 917, loss_train: 416.09, accuracy_train: 96.5%\n",
      "\t  loss_test: 498.66, accuracy_test: 76.9%\n",
      "Epoch: 918, loss_train: 415.51, accuracy_train: 96.5%\n",
      "\t  loss_test: 498.77, accuracy_test: 76.9%\n",
      "Epoch: 919, loss_train: 414.94, accuracy_train: 96.5%\n",
      "\t  loss_test: 498.88, accuracy_test: 76.9%\n",
      "Epoch: 920, loss_train: 414.36, accuracy_train: 96.5%\n",
      "\t  loss_test: 498.98, accuracy_test: 76.9%\n",
      "Epoch: 921, loss_train: 413.79, accuracy_train: 96.5%\n",
      "\t  loss_test: 499.09, accuracy_test: 77.0%\n",
      "Epoch: 922, loss_train: 413.22, accuracy_train: 96.5%\n",
      "\t  loss_test: 499.20, accuracy_test: 77.0%\n",
      "Epoch: 923, loss_train: 412.66, accuracy_train: 96.5%\n",
      "\t  loss_test: 499.30, accuracy_test: 77.0%\n",
      "Epoch: 924, loss_train: 412.09, accuracy_train: 96.5%\n",
      "\t  loss_test: 499.41, accuracy_test: 77.0%\n",
      "Epoch: 925, loss_train: 411.53, accuracy_train: 96.5%\n",
      "\t  loss_test: 499.52, accuracy_test: 77.0%\n",
      "Epoch: 926, loss_train: 410.98, accuracy_train: 96.5%\n",
      "\t  loss_test: 499.63, accuracy_test: 77.0%\n",
      "Epoch: 927, loss_train: 410.42, accuracy_train: 96.5%\n",
      "\t  loss_test: 499.73, accuracy_test: 77.0%\n",
      "Epoch: 928, loss_train: 409.87, accuracy_train: 96.5%\n",
      "\t  loss_test: 499.84, accuracy_test: 77.0%\n",
      "Epoch: 929, loss_train: 409.32, accuracy_train: 96.5%\n",
      "\t  loss_test: 499.95, accuracy_test: 77.0%\n",
      "Epoch: 930, loss_train: 408.77, accuracy_train: 96.5%\n",
      "\t  loss_test: 500.06, accuracy_test: 77.0%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 931, loss_train: 408.23, accuracy_train: 96.5%\n",
      "\t  loss_test: 500.17, accuracy_test: 77.0%\n",
      "Epoch: 932, loss_train: 407.69, accuracy_train: 96.5%\n",
      "\t  loss_test: 500.28, accuracy_test: 77.0%\n",
      "Epoch: 933, loss_train: 407.15, accuracy_train: 96.5%\n",
      "\t  loss_test: 500.38, accuracy_test: 77.0%\n",
      "Epoch: 934, loss_train: 406.61, accuracy_train: 96.5%\n",
      "\t  loss_test: 500.49, accuracy_test: 77.0%\n",
      "Epoch: 935, loss_train: 406.08, accuracy_train: 96.5%\n",
      "\t  loss_test: 500.60, accuracy_test: 77.0%\n",
      "Epoch: 936, loss_train: 405.55, accuracy_train: 96.5%\n",
      "\t  loss_test: 500.71, accuracy_test: 77.0%\n",
      "Epoch: 937, loss_train: 405.02, accuracy_train: 96.5%\n",
      "\t  loss_test: 500.82, accuracy_test: 77.0%\n",
      "Epoch: 938, loss_train: 404.50, accuracy_train: 96.5%\n",
      "\t  loss_test: 500.93, accuracy_test: 77.0%\n",
      "Epoch: 939, loss_train: 403.97, accuracy_train: 96.5%\n",
      "\t  loss_test: 501.04, accuracy_test: 77.0%\n",
      "Epoch: 940, loss_train: 403.45, accuracy_train: 96.5%\n",
      "\t  loss_test: 501.15, accuracy_test: 77.0%\n",
      "Epoch: 941, loss_train: 402.93, accuracy_train: 96.5%\n",
      "\t  loss_test: 501.26, accuracy_test: 77.0%\n",
      "Epoch: 942, loss_train: 402.42, accuracy_train: 96.5%\n",
      "\t  loss_test: 501.37, accuracy_test: 77.0%\n",
      "Epoch: 943, loss_train: 401.91, accuracy_train: 96.5%\n",
      "\t  loss_test: 501.48, accuracy_test: 77.0%\n",
      "Epoch: 944, loss_train: 401.40, accuracy_train: 96.5%\n",
      "\t  loss_test: 501.59, accuracy_test: 77.0%\n",
      "Epoch: 945, loss_train: 400.89, accuracy_train: 96.5%\n",
      "\t  loss_test: 501.70, accuracy_test: 77.0%\n",
      "Epoch: 946, loss_train: 400.38, accuracy_train: 96.5%\n",
      "\t  loss_test: 501.81, accuracy_test: 77.0%\n",
      "Epoch: 947, loss_train: 399.88, accuracy_train: 96.5%\n",
      "\t  loss_test: 501.92, accuracy_test: 77.0%\n",
      "Epoch: 948, loss_train: 399.38, accuracy_train: 96.5%\n",
      "\t  loss_test: 502.04, accuracy_test: 77.0%\n",
      "Epoch: 949, loss_train: 398.88, accuracy_train: 96.5%\n",
      "\t  loss_test: 502.15, accuracy_test: 77.0%\n",
      "Epoch: 950, loss_train: 398.38, accuracy_train: 96.5%\n",
      "\t  loss_test: 502.26, accuracy_test: 77.0%\n",
      "Epoch: 951, loss_train: 397.89, accuracy_train: 96.5%\n",
      "\t  loss_test: 502.37, accuracy_test: 77.0%\n",
      "Epoch: 952, loss_train: 397.40, accuracy_train: 96.5%\n",
      "\t  loss_test: 502.48, accuracy_test: 77.0%\n",
      "Epoch: 953, loss_train: 396.91, accuracy_train: 96.5%\n",
      "\t  loss_test: 502.59, accuracy_test: 77.0%\n",
      "Epoch: 954, loss_train: 396.43, accuracy_train: 96.5%\n",
      "\t  loss_test: 502.71, accuracy_test: 77.0%\n",
      "Epoch: 955, loss_train: 395.94, accuracy_train: 96.5%\n",
      "\t  loss_test: 502.82, accuracy_test: 77.0%\n",
      "Epoch: 956, loss_train: 395.46, accuracy_train: 96.5%\n",
      "\t  loss_test: 502.93, accuracy_test: 77.0%\n",
      "Epoch: 957, loss_train: 394.98, accuracy_train: 96.5%\n",
      "\t  loss_test: 503.04, accuracy_test: 77.0%\n",
      "Epoch: 958, loss_train: 394.50, accuracy_train: 96.5%\n",
      "\t  loss_test: 503.15, accuracy_test: 77.0%\n",
      "Epoch: 959, loss_train: 394.03, accuracy_train: 96.5%\n",
      "\t  loss_test: 503.27, accuracy_test: 77.0%\n",
      "Epoch: 960, loss_train: 393.56, accuracy_train: 96.5%\n",
      "\t  loss_test: 503.38, accuracy_test: 77.0%\n",
      "Epoch: 961, loss_train: 393.09, accuracy_train: 96.5%\n",
      "\t  loss_test: 503.49, accuracy_test: 77.0%\n",
      "Epoch: 962, loss_train: 392.62, accuracy_train: 96.5%\n",
      "\t  loss_test: 503.61, accuracy_test: 77.0%\n",
      "Epoch: 963, loss_train: 392.15, accuracy_train: 96.5%\n",
      "\t  loss_test: 503.72, accuracy_test: 77.0%\n",
      "Epoch: 964, loss_train: 391.69, accuracy_train: 96.5%\n",
      "\t  loss_test: 503.83, accuracy_test: 77.0%\n",
      "Epoch: 965, loss_train: 391.23, accuracy_train: 96.5%\n",
      "\t  loss_test: 503.94, accuracy_test: 77.0%\n",
      "Epoch: 966, loss_train: 390.77, accuracy_train: 96.5%\n",
      "\t  loss_test: 504.06, accuracy_test: 77.0%\n",
      "Epoch: 967, loss_train: 390.31, accuracy_train: 96.5%\n",
      "\t  loss_test: 504.17, accuracy_test: 77.0%\n",
      "Epoch: 968, loss_train: 389.86, accuracy_train: 96.5%\n",
      "\t  loss_test: 504.28, accuracy_test: 77.0%\n",
      "Epoch: 969, loss_train: 389.40, accuracy_train: 96.5%\n",
      "\t  loss_test: 504.40, accuracy_test: 77.0%\n",
      "Epoch: 970, loss_train: 388.95, accuracy_train: 96.5%\n",
      "\t  loss_test: 504.51, accuracy_test: 77.0%\n",
      "Epoch: 971, loss_train: 388.51, accuracy_train: 96.5%\n",
      "\t  loss_test: 504.63, accuracy_test: 77.0%\n",
      "Epoch: 972, loss_train: 388.06, accuracy_train: 96.5%\n",
      "\t  loss_test: 504.74, accuracy_test: 77.0%\n",
      "Epoch: 973, loss_train: 387.62, accuracy_train: 96.5%\n",
      "\t  loss_test: 504.85, accuracy_test: 77.0%\n",
      "Epoch: 974, loss_train: 387.17, accuracy_train: 96.5%\n",
      "\t  loss_test: 504.97, accuracy_test: 77.0%\n",
      "Epoch: 975, loss_train: 386.73, accuracy_train: 96.5%\n",
      "\t  loss_test: 505.08, accuracy_test: 77.0%\n",
      "Epoch: 976, loss_train: 386.30, accuracy_train: 96.5%\n",
      "\t  loss_test: 505.20, accuracy_test: 77.0%\n",
      "Epoch: 977, loss_train: 385.86, accuracy_train: 96.5%\n",
      "\t  loss_test: 505.31, accuracy_test: 77.0%\n",
      "Epoch: 978, loss_train: 385.43, accuracy_train: 96.5%\n",
      "\t  loss_test: 505.43, accuracy_test: 77.0%\n",
      "Epoch: 979, loss_train: 384.99, accuracy_train: 96.5%\n",
      "\t  loss_test: 505.54, accuracy_test: 77.0%\n",
      "Epoch: 980, loss_train: 384.57, accuracy_train: 96.5%\n",
      "\t  loss_test: 505.65, accuracy_test: 77.0%\n",
      "Epoch: 981, loss_train: 384.14, accuracy_train: 96.5%\n",
      "\t  loss_test: 505.77, accuracy_test: 77.0%\n",
      "Epoch: 982, loss_train: 383.71, accuracy_train: 96.5%\n",
      "\t  loss_test: 505.88, accuracy_test: 77.0%\n",
      "Epoch: 983, loss_train: 383.29, accuracy_train: 96.5%\n",
      "\t  loss_test: 506.00, accuracy_test: 77.0%\n",
      "Epoch: 984, loss_train: 382.87, accuracy_train: 96.5%\n",
      "\t  loss_test: 506.11, accuracy_test: 77.0%\n",
      "Epoch: 985, loss_train: 382.45, accuracy_train: 96.5%\n",
      "\t  loss_test: 506.23, accuracy_test: 77.0%\n",
      "Epoch: 986, loss_train: 382.03, accuracy_train: 96.5%\n",
      "\t  loss_test: 506.34, accuracy_test: 77.0%\n",
      "Epoch: 987, loss_train: 381.61, accuracy_train: 96.5%\n",
      "\t  loss_test: 506.46, accuracy_test: 77.0%\n",
      "Epoch: 988, loss_train: 381.20, accuracy_train: 96.5%\n",
      "\t  loss_test: 506.57, accuracy_test: 77.0%\n",
      "Epoch: 989, loss_train: 380.79, accuracy_train: 96.5%\n",
      "\t  loss_test: 506.69, accuracy_test: 77.0%\n",
      "Epoch: 990, loss_train: 380.38, accuracy_train: 96.5%\n",
      "\t  loss_test: 506.81, accuracy_test: 77.1%\n",
      "Epoch: 991, loss_train: 379.97, accuracy_train: 96.5%\n",
      "\t  loss_test: 506.92, accuracy_test: 77.1%\n",
      "Epoch: 992, loss_train: 379.56, accuracy_train: 96.5%\n",
      "\t  loss_test: 507.04, accuracy_test: 77.1%\n",
      "Epoch: 993, loss_train: 379.16, accuracy_train: 96.5%\n",
      "\t  loss_test: 507.15, accuracy_test: 77.1%\n",
      "Epoch: 994, loss_train: 378.76, accuracy_train: 96.5%\n",
      "\t  loss_test: 507.27, accuracy_test: 77.1%\n",
      "Epoch: 995, loss_train: 378.36, accuracy_train: 96.5%\n",
      "\t  loss_test: 507.38, accuracy_test: 77.1%\n",
      "Epoch: 996, loss_train: 377.96, accuracy_train: 96.5%\n",
      "\t  loss_test: 507.50, accuracy_test: 77.1%\n",
      "Epoch: 997, loss_train: 377.56, accuracy_train: 96.5%\n",
      "\t  loss_test: 507.61, accuracy_test: 77.1%\n",
      "Epoch: 998, loss_train: 377.17, accuracy_train: 96.5%\n",
      "\t  loss_test: 507.73, accuracy_test: 77.1%\n",
      "Epoch: 999, loss_train: 376.77, accuracy_train: 96.5%\n",
      "\t  loss_test: 507.85, accuracy_test: 77.1%\n",
      "Epoch: 1000, loss_train: 376.38, accuracy_train: 96.5%\n",
      "\t  loss_test: 507.96, accuracy_test: 77.1%\n",
      "Epoch: 1001, loss_train: 375.99, accuracy_train: 96.5%\n",
      "\t  loss_test: 508.08, accuracy_test: 77.1%\n",
      "Epoch: 1002, loss_train: 375.61, accuracy_train: 96.5%\n",
      "\t  loss_test: 508.20, accuracy_test: 77.1%\n",
      "Epoch: 1003, loss_train: 375.22, accuracy_train: 96.5%\n",
      "\t  loss_test: 508.31, accuracy_test: 77.1%\n",
      "Epoch: 1004, loss_train: 374.84, accuracy_train: 96.5%\n",
      "\t  loss_test: 508.43, accuracy_test: 77.1%\n",
      "Epoch: 1005, loss_train: 374.45, accuracy_train: 96.5%\n",
      "\t  loss_test: 508.54, accuracy_test: 77.1%\n",
      "Epoch: 1006, loss_train: 374.07, accuracy_train: 96.5%\n",
      "\t  loss_test: 508.66, accuracy_test: 77.1%\n",
      "Epoch: 1007, loss_train: 373.69, accuracy_train: 96.5%\n",
      "\t  loss_test: 508.78, accuracy_test: 77.1%\n",
      "Epoch: 1008, loss_train: 373.32, accuracy_train: 96.5%\n",
      "\t  loss_test: 508.89, accuracy_test: 77.1%\n",
      "Epoch: 1009, loss_train: 372.94, accuracy_train: 96.5%\n",
      "\t  loss_test: 509.01, accuracy_test: 77.1%\n",
      "Epoch: 1010, loss_train: 372.57, accuracy_train: 96.5%\n",
      "\t  loss_test: 509.13, accuracy_test: 77.1%\n",
      "Epoch: 1011, loss_train: 372.20, accuracy_train: 96.5%\n",
      "\t  loss_test: 509.24, accuracy_test: 77.1%\n",
      "Epoch: 1012, loss_train: 371.82, accuracy_train: 96.5%\n",
      "\t  loss_test: 509.36, accuracy_test: 77.1%\n",
      "Epoch: 1013, loss_train: 371.46, accuracy_train: 96.5%\n",
      "\t  loss_test: 509.48, accuracy_test: 77.1%\n",
      "Epoch: 1014, loss_train: 371.09, accuracy_train: 96.5%\n",
      "\t  loss_test: 509.59, accuracy_test: 77.1%\n",
      "Epoch: 1015, loss_train: 370.72, accuracy_train: 96.5%\n",
      "\t  loss_test: 509.71, accuracy_test: 77.1%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 1016, loss_train: 370.36, accuracy_train: 96.5%\n",
      "\t  loss_test: 509.83, accuracy_test: 77.1%\n",
      "Epoch: 1017, loss_train: 370.00, accuracy_train: 96.5%\n",
      "\t  loss_test: 509.94, accuracy_test: 77.1%\n",
      "Epoch: 1018, loss_train: 369.64, accuracy_train: 96.5%\n",
      "\t  loss_test: 510.06, accuracy_test: 77.1%\n",
      "Epoch: 1019, loss_train: 369.28, accuracy_train: 96.5%\n",
      "\t  loss_test: 510.18, accuracy_test: 77.1%\n",
      "Epoch: 1020, loss_train: 368.92, accuracy_train: 96.5%\n",
      "\t  loss_test: 510.29, accuracy_test: 77.1%\n",
      "Epoch: 1021, loss_train: 368.57, accuracy_train: 96.5%\n",
      "\t  loss_test: 510.41, accuracy_test: 77.3%\n",
      "Epoch: 1022, loss_train: 368.21, accuracy_train: 96.5%\n",
      "\t  loss_test: 510.53, accuracy_test: 77.3%\n",
      "Epoch: 1023, loss_train: 367.86, accuracy_train: 96.5%\n",
      "\t  loss_test: 510.64, accuracy_test: 77.3%\n",
      "Epoch: 1024, loss_train: 367.51, accuracy_train: 96.5%\n",
      "\t  loss_test: 510.76, accuracy_test: 77.3%\n",
      "Epoch: 1025, loss_train: 367.16, accuracy_train: 96.5%\n",
      "\t  loss_test: 510.88, accuracy_test: 77.3%\n",
      "Epoch: 1026, loss_train: 366.81, accuracy_train: 96.5%\n",
      "\t  loss_test: 510.99, accuracy_test: 77.3%\n",
      "Epoch: 1027, loss_train: 366.47, accuracy_train: 96.5%\n",
      "\t  loss_test: 511.11, accuracy_test: 77.3%\n",
      "Epoch: 1028, loss_train: 366.12, accuracy_train: 96.5%\n",
      "\t  loss_test: 511.23, accuracy_test: 77.3%\n",
      "Epoch: 1029, loss_train: 365.78, accuracy_train: 96.5%\n",
      "\t  loss_test: 511.35, accuracy_test: 77.3%\n",
      "Epoch: 1030, loss_train: 365.44, accuracy_train: 96.5%\n",
      "\t  loss_test: 511.46, accuracy_test: 77.3%\n",
      "Epoch: 1031, loss_train: 365.10, accuracy_train: 96.5%\n",
      "\t  loss_test: 511.58, accuracy_test: 77.3%\n",
      "Epoch: 1032, loss_train: 364.76, accuracy_train: 96.5%\n",
      "\t  loss_test: 511.70, accuracy_test: 77.3%\n",
      "Epoch: 1033, loss_train: 364.42, accuracy_train: 96.5%\n",
      "\t  loss_test: 511.81, accuracy_test: 77.3%\n",
      "Epoch: 1034, loss_train: 364.09, accuracy_train: 96.5%\n",
      "\t  loss_test: 511.93, accuracy_test: 77.3%\n",
      "Epoch: 1035, loss_train: 363.76, accuracy_train: 96.5%\n",
      "\t  loss_test: 512.05, accuracy_test: 77.3%\n",
      "Epoch: 1036, loss_train: 363.42, accuracy_train: 96.5%\n",
      "\t  loss_test: 512.17, accuracy_test: 77.4%\n",
      "Epoch: 1037, loss_train: 363.09, accuracy_train: 96.5%\n",
      "\t  loss_test: 512.28, accuracy_test: 77.4%\n",
      "Epoch: 1038, loss_train: 362.76, accuracy_train: 96.5%\n",
      "\t  loss_test: 512.40, accuracy_test: 77.4%\n",
      "Epoch: 1039, loss_train: 362.43, accuracy_train: 96.5%\n",
      "\t  loss_test: 512.52, accuracy_test: 77.4%\n",
      "Epoch: 1040, loss_train: 362.11, accuracy_train: 96.5%\n",
      "\t  loss_test: 512.64, accuracy_test: 77.4%\n",
      "Epoch: 1041, loss_train: 361.78, accuracy_train: 96.5%\n",
      "\t  loss_test: 512.75, accuracy_test: 77.4%\n",
      "Epoch: 1042, loss_train: 361.46, accuracy_train: 96.5%\n",
      "\t  loss_test: 512.87, accuracy_test: 77.4%\n",
      "Epoch: 1043, loss_train: 361.14, accuracy_train: 96.5%\n",
      "\t  loss_test: 512.99, accuracy_test: 77.4%\n",
      "Epoch: 1044, loss_train: 360.82, accuracy_train: 96.5%\n",
      "\t  loss_test: 513.10, accuracy_test: 77.4%\n",
      "Epoch: 1045, loss_train: 360.50, accuracy_train: 96.5%\n",
      "\t  loss_test: 513.22, accuracy_test: 77.4%\n",
      "Epoch: 1046, loss_train: 360.18, accuracy_train: 96.5%\n",
      "\t  loss_test: 513.34, accuracy_test: 77.4%\n",
      "Epoch: 1047, loss_train: 359.86, accuracy_train: 96.5%\n",
      "\t  loss_test: 513.46, accuracy_test: 77.4%\n",
      "Epoch: 1048, loss_train: 359.55, accuracy_train: 96.5%\n",
      "\t  loss_test: 513.57, accuracy_test: 77.4%\n",
      "Epoch: 1049, loss_train: 359.23, accuracy_train: 96.5%\n",
      "\t  loss_test: 513.69, accuracy_test: 77.4%\n",
      "Epoch: 1050, loss_train: 358.92, accuracy_train: 96.5%\n",
      "\t  loss_test: 513.81, accuracy_test: 77.4%\n",
      "Epoch: 1051, loss_train: 358.61, accuracy_train: 96.5%\n",
      "\t  loss_test: 513.93, accuracy_test: 77.4%\n",
      "Epoch: 1052, loss_train: 358.30, accuracy_train: 96.5%\n",
      "\t  loss_test: 514.04, accuracy_test: 77.4%\n",
      "Epoch: 1053, loss_train: 357.99, accuracy_train: 96.5%\n",
      "\t  loss_test: 514.16, accuracy_test: 77.4%\n",
      "Epoch: 1054, loss_train: 357.69, accuracy_train: 96.5%\n",
      "\t  loss_test: 514.28, accuracy_test: 77.4%\n",
      "Epoch: 1055, loss_train: 357.38, accuracy_train: 96.5%\n",
      "\t  loss_test: 514.39, accuracy_test: 77.4%\n",
      "Epoch: 1056, loss_train: 357.08, accuracy_train: 96.5%\n",
      "\t  loss_test: 514.51, accuracy_test: 77.4%\n",
      "Epoch: 1057, loss_train: 356.77, accuracy_train: 96.5%\n",
      "\t  loss_test: 514.63, accuracy_test: 77.4%\n",
      "Epoch: 1058, loss_train: 356.47, accuracy_train: 96.5%\n",
      "\t  loss_test: 514.75, accuracy_test: 77.4%\n",
      "Epoch: 1059, loss_train: 356.17, accuracy_train: 96.5%\n",
      "\t  loss_test: 514.86, accuracy_test: 77.5%\n",
      "Epoch: 1060, loss_train: 355.87, accuracy_train: 96.5%\n",
      "\t  loss_test: 514.98, accuracy_test: 77.5%\n",
      "Epoch: 1061, loss_train: 355.57, accuracy_train: 96.5%\n",
      "\t  loss_test: 515.10, accuracy_test: 77.5%\n",
      "Epoch: 1062, loss_train: 355.28, accuracy_train: 96.5%\n",
      "\t  loss_test: 515.22, accuracy_test: 77.5%\n",
      "Epoch: 1063, loss_train: 354.98, accuracy_train: 96.5%\n",
      "\t  loss_test: 515.33, accuracy_test: 77.5%\n",
      "Epoch: 1064, loss_train: 354.69, accuracy_train: 96.5%\n",
      "\t  loss_test: 515.45, accuracy_test: 77.5%\n",
      "Epoch: 1065, loss_train: 354.40, accuracy_train: 96.5%\n",
      "\t  loss_test: 515.57, accuracy_test: 77.5%\n",
      "Epoch: 1066, loss_train: 354.11, accuracy_train: 96.5%\n",
      "\t  loss_test: 515.68, accuracy_test: 77.5%\n",
      "Epoch: 1067, loss_train: 353.82, accuracy_train: 96.5%\n",
      "\t  loss_test: 515.80, accuracy_test: 77.5%\n",
      "Epoch: 1068, loss_train: 353.53, accuracy_train: 96.5%\n",
      "\t  loss_test: 515.92, accuracy_test: 77.5%\n",
      "Epoch: 1069, loss_train: 353.24, accuracy_train: 96.5%\n",
      "\t  loss_test: 516.04, accuracy_test: 77.5%\n",
      "Epoch: 1070, loss_train: 352.95, accuracy_train: 96.5%\n",
      "\t  loss_test: 516.15, accuracy_test: 77.5%\n",
      "Epoch: 1071, loss_train: 352.67, accuracy_train: 96.5%\n",
      "\t  loss_test: 516.27, accuracy_test: 77.5%\n",
      "Epoch: 1072, loss_train: 352.38, accuracy_train: 96.5%\n",
      "\t  loss_test: 516.39, accuracy_test: 77.5%\n",
      "Epoch: 1073, loss_train: 352.10, accuracy_train: 96.5%\n",
      "\t  loss_test: 516.50, accuracy_test: 77.5%\n",
      "Epoch: 1074, loss_train: 351.82, accuracy_train: 96.5%\n",
      "\t  loss_test: 516.62, accuracy_test: 77.5%\n",
      "Epoch: 1075, loss_train: 351.54, accuracy_train: 96.5%\n",
      "\t  loss_test: 516.74, accuracy_test: 77.5%\n",
      "Epoch: 1076, loss_train: 351.26, accuracy_train: 96.5%\n",
      "\t  loss_test: 516.86, accuracy_test: 77.5%\n",
      "Epoch: 1077, loss_train: 350.98, accuracy_train: 96.5%\n",
      "\t  loss_test: 516.97, accuracy_test: 77.5%\n",
      "Epoch: 1078, loss_train: 350.71, accuracy_train: 96.5%\n",
      "\t  loss_test: 517.09, accuracy_test: 77.5%\n",
      "Epoch: 1079, loss_train: 350.43, accuracy_train: 96.5%\n",
      "\t  loss_test: 517.21, accuracy_test: 77.5%\n",
      "Epoch: 1080, loss_train: 350.16, accuracy_train: 96.5%\n",
      "\t  loss_test: 517.32, accuracy_test: 77.5%\n",
      "Epoch: 1081, loss_train: 349.88, accuracy_train: 96.5%\n",
      "\t  loss_test: 517.44, accuracy_test: 77.5%\n",
      "Epoch: 1082, loss_train: 349.61, accuracy_train: 96.5%\n",
      "\t  loss_test: 517.56, accuracy_test: 77.5%\n",
      "Epoch: 1083, loss_train: 349.34, accuracy_train: 96.5%\n",
      "\t  loss_test: 517.67, accuracy_test: 77.5%\n",
      "Epoch: 1084, loss_train: 349.07, accuracy_train: 96.5%\n",
      "\t  loss_test: 517.79, accuracy_test: 77.5%\n",
      "Epoch: 1085, loss_train: 348.80, accuracy_train: 96.5%\n",
      "\t  loss_test: 517.91, accuracy_test: 77.5%\n",
      "Epoch: 1086, loss_train: 348.53, accuracy_train: 96.5%\n",
      "\t  loss_test: 518.02, accuracy_test: 77.5%\n",
      "Epoch: 1087, loss_train: 348.27, accuracy_train: 96.5%\n",
      "\t  loss_test: 518.14, accuracy_test: 77.5%\n",
      "Epoch: 1088, loss_train: 348.00, accuracy_train: 96.5%\n",
      "\t  loss_test: 518.26, accuracy_test: 77.5%\n",
      "Epoch: 1089, loss_train: 347.74, accuracy_train: 96.5%\n",
      "\t  loss_test: 518.37, accuracy_test: 77.5%\n",
      "Epoch: 1090, loss_train: 347.47, accuracy_train: 96.5%\n",
      "\t  loss_test: 518.49, accuracy_test: 77.5%\n",
      "Epoch: 1091, loss_train: 347.21, accuracy_train: 96.5%\n",
      "\t  loss_test: 518.61, accuracy_test: 77.5%\n",
      "Epoch: 1092, loss_train: 346.95, accuracy_train: 96.5%\n",
      "\t  loss_test: 518.72, accuracy_test: 77.5%\n",
      "Epoch: 1093, loss_train: 346.69, accuracy_train: 96.5%\n",
      "\t  loss_test: 518.84, accuracy_test: 77.5%\n",
      "Epoch: 1094, loss_train: 346.43, accuracy_train: 96.5%\n",
      "\t  loss_test: 518.96, accuracy_test: 77.5%\n",
      "Epoch: 1095, loss_train: 346.18, accuracy_train: 96.5%\n",
      "\t  loss_test: 519.07, accuracy_test: 77.5%\n",
      "Epoch: 1096, loss_train: 345.92, accuracy_train: 96.5%\n",
      "\t  loss_test: 519.19, accuracy_test: 77.5%\n",
      "Epoch: 1097, loss_train: 345.66, accuracy_train: 96.5%\n",
      "\t  loss_test: 519.31, accuracy_test: 77.5%\n",
      "Epoch: 1098, loss_train: 345.41, accuracy_train: 96.5%\n",
      "\t  loss_test: 519.42, accuracy_test: 77.5%\n",
      "Epoch: 1099, loss_train: 345.16, accuracy_train: 96.5%\n",
      "\t  loss_test: 519.54, accuracy_test: 77.5%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 1100, loss_train: 344.90, accuracy_train: 96.5%\n",
      "\t  loss_test: 519.66, accuracy_test: 77.5%\n",
      "Epoch: 1101, loss_train: 344.65, accuracy_train: 96.5%\n",
      "\t  loss_test: 519.77, accuracy_test: 77.5%\n",
      "Epoch: 1102, loss_train: 344.40, accuracy_train: 96.5%\n",
      "\t  loss_test: 519.89, accuracy_test: 77.5%\n",
      "Epoch: 1103, loss_train: 344.15, accuracy_train: 96.5%\n",
      "\t  loss_test: 520.00, accuracy_test: 77.4%\n",
      "Epoch: 1104, loss_train: 343.90, accuracy_train: 96.5%\n",
      "\t  loss_test: 520.12, accuracy_test: 77.4%\n",
      "Epoch: 1105, loss_train: 343.66, accuracy_train: 96.5%\n",
      "\t  loss_test: 520.24, accuracy_test: 77.4%\n",
      "Epoch: 1106, loss_train: 343.41, accuracy_train: 96.5%\n",
      "\t  loss_test: 520.35, accuracy_test: 77.4%\n",
      "Epoch: 1107, loss_train: 343.17, accuracy_train: 96.5%\n",
      "\t  loss_test: 520.47, accuracy_test: 77.4%\n",
      "Epoch: 1108, loss_train: 342.92, accuracy_train: 96.5%\n",
      "\t  loss_test: 520.59, accuracy_test: 77.4%\n",
      "Epoch: 1109, loss_train: 342.68, accuracy_train: 96.5%\n",
      "\t  loss_test: 520.70, accuracy_test: 77.4%\n",
      "Epoch: 1110, loss_train: 342.44, accuracy_train: 96.5%\n",
      "\t  loss_test: 520.82, accuracy_test: 77.4%\n",
      "Epoch: 1111, loss_train: 342.20, accuracy_train: 96.5%\n",
      "\t  loss_test: 520.93, accuracy_test: 77.4%\n",
      "Epoch: 1112, loss_train: 341.96, accuracy_train: 96.5%\n",
      "\t  loss_test: 521.05, accuracy_test: 77.4%\n",
      "Epoch: 1113, loss_train: 341.72, accuracy_train: 96.5%\n",
      "\t  loss_test: 521.16, accuracy_test: 77.4%\n",
      "Epoch: 1114, loss_train: 341.48, accuracy_train: 96.5%\n",
      "\t  loss_test: 521.28, accuracy_test: 77.4%\n",
      "Epoch: 1115, loss_train: 341.24, accuracy_train: 96.5%\n",
      "\t  loss_test: 521.40, accuracy_test: 77.4%\n",
      "Epoch: 1116, loss_train: 341.01, accuracy_train: 96.5%\n",
      "\t  loss_test: 521.51, accuracy_test: 77.4%\n",
      "Epoch: 1117, loss_train: 340.77, accuracy_train: 96.5%\n",
      "\t  loss_test: 521.63, accuracy_test: 77.4%\n",
      "Epoch: 1118, loss_train: 340.54, accuracy_train: 96.5%\n",
      "\t  loss_test: 521.74, accuracy_test: 77.4%\n",
      "Epoch: 1119, loss_train: 340.30, accuracy_train: 96.5%\n",
      "\t  loss_test: 521.86, accuracy_test: 77.4%\n",
      "Epoch: 1120, loss_train: 340.07, accuracy_train: 96.5%\n",
      "\t  loss_test: 521.97, accuracy_test: 77.4%\n",
      "Epoch: 1121, loss_train: 339.84, accuracy_train: 96.5%\n",
      "\t  loss_test: 522.09, accuracy_test: 77.4%\n",
      "Epoch: 1122, loss_train: 339.61, accuracy_train: 96.5%\n",
      "\t  loss_test: 522.20, accuracy_test: 77.4%\n",
      "Epoch: 1123, loss_train: 339.38, accuracy_train: 96.5%\n",
      "\t  loss_test: 522.32, accuracy_test: 77.4%\n",
      "Epoch: 1124, loss_train: 339.15, accuracy_train: 96.5%\n",
      "\t  loss_test: 522.44, accuracy_test: 77.4%\n",
      "Epoch: 1125, loss_train: 338.92, accuracy_train: 96.5%\n",
      "\t  loss_test: 522.55, accuracy_test: 77.4%\n",
      "Epoch: 1126, loss_train: 338.69, accuracy_train: 96.5%\n",
      "\t  loss_test: 522.67, accuracy_test: 77.4%\n",
      "Epoch: 1127, loss_train: 338.47, accuracy_train: 96.5%\n",
      "\t  loss_test: 522.78, accuracy_test: 77.4%\n",
      "Epoch: 1128, loss_train: 338.24, accuracy_train: 96.5%\n",
      "\t  loss_test: 522.90, accuracy_test: 77.4%\n",
      "Epoch: 1129, loss_train: 338.02, accuracy_train: 96.5%\n",
      "\t  loss_test: 523.01, accuracy_test: 77.4%\n",
      "Epoch: 1130, loss_train: 337.80, accuracy_train: 96.5%\n",
      "\t  loss_test: 523.13, accuracy_test: 77.4%\n",
      "Epoch: 1131, loss_train: 337.57, accuracy_train: 96.5%\n",
      "\t  loss_test: 523.24, accuracy_test: 77.4%\n",
      "Epoch: 1132, loss_train: 337.35, accuracy_train: 96.5%\n",
      "\t  loss_test: 523.36, accuracy_test: 77.4%\n",
      "Epoch: 1133, loss_train: 337.13, accuracy_train: 96.5%\n",
      "\t  loss_test: 523.47, accuracy_test: 77.4%\n",
      "Epoch: 1134, loss_train: 336.91, accuracy_train: 96.5%\n",
      "\t  loss_test: 523.59, accuracy_test: 77.4%\n",
      "Epoch: 1135, loss_train: 336.69, accuracy_train: 96.5%\n",
      "\t  loss_test: 523.70, accuracy_test: 77.4%\n",
      "Epoch: 1136, loss_train: 336.48, accuracy_train: 96.5%\n",
      "\t  loss_test: 523.81, accuracy_test: 77.4%\n",
      "Epoch: 1137, loss_train: 336.26, accuracy_train: 96.5%\n",
      "\t  loss_test: 523.93, accuracy_test: 77.4%\n",
      "Epoch: 1138, loss_train: 336.04, accuracy_train: 96.5%\n",
      "\t  loss_test: 524.04, accuracy_test: 77.4%\n",
      "Epoch: 1139, loss_train: 335.83, accuracy_train: 96.5%\n",
      "\t  loss_test: 524.16, accuracy_test: 77.4%\n",
      "Epoch: 1140, loss_train: 335.61, accuracy_train: 96.5%\n",
      "\t  loss_test: 524.27, accuracy_test: 77.4%\n",
      "Epoch: 1141, loss_train: 335.40, accuracy_train: 96.5%\n",
      "\t  loss_test: 524.39, accuracy_test: 77.4%\n",
      "Epoch: 1142, loss_train: 335.19, accuracy_train: 96.5%\n",
      "\t  loss_test: 524.50, accuracy_test: 77.4%\n",
      "Epoch: 1143, loss_train: 334.97, accuracy_train: 96.5%\n",
      "\t  loss_test: 524.62, accuracy_test: 77.4%\n",
      "Epoch: 1144, loss_train: 334.76, accuracy_train: 96.5%\n",
      "\t  loss_test: 524.73, accuracy_test: 77.4%\n",
      "Epoch: 1145, loss_train: 334.55, accuracy_train: 96.5%\n",
      "\t  loss_test: 524.84, accuracy_test: 77.4%\n",
      "Epoch: 1146, loss_train: 334.34, accuracy_train: 96.5%\n",
      "\t  loss_test: 524.96, accuracy_test: 77.4%\n",
      "Epoch: 1147, loss_train: 334.13, accuracy_train: 96.5%\n",
      "\t  loss_test: 525.07, accuracy_test: 77.4%\n",
      "Epoch: 1148, loss_train: 333.93, accuracy_train: 96.5%\n",
      "\t  loss_test: 525.19, accuracy_test: 77.4%\n",
      "Epoch: 1149, loss_train: 333.72, accuracy_train: 96.5%\n",
      "\t  loss_test: 525.30, accuracy_test: 77.4%\n",
      "Epoch: 1150, loss_train: 333.51, accuracy_train: 96.5%\n",
      "\t  loss_test: 525.41, accuracy_test: 77.4%\n",
      "Epoch: 1151, loss_train: 333.31, accuracy_train: 96.5%\n",
      "\t  loss_test: 525.53, accuracy_test: 77.4%\n",
      "Epoch: 1152, loss_train: 333.10, accuracy_train: 96.5%\n",
      "\t  loss_test: 525.64, accuracy_test: 77.4%\n",
      "Epoch: 1153, loss_train: 332.90, accuracy_train: 96.5%\n",
      "\t  loss_test: 525.76, accuracy_test: 77.4%\n",
      "Epoch: 1154, loss_train: 332.70, accuracy_train: 96.5%\n",
      "\t  loss_test: 525.87, accuracy_test: 77.4%\n",
      "Epoch: 1155, loss_train: 332.49, accuracy_train: 96.5%\n",
      "\t  loss_test: 525.98, accuracy_test: 77.4%\n",
      "Epoch: 1156, loss_train: 332.29, accuracy_train: 96.5%\n",
      "\t  loss_test: 526.10, accuracy_test: 77.4%\n",
      "Epoch: 1157, loss_train: 332.09, accuracy_train: 96.5%\n",
      "\t  loss_test: 526.21, accuracy_test: 77.3%\n",
      "Epoch: 1158, loss_train: 331.89, accuracy_train: 96.5%\n",
      "\t  loss_test: 526.32, accuracy_test: 77.3%\n",
      "Epoch: 1159, loss_train: 331.69, accuracy_train: 96.5%\n",
      "\t  loss_test: 526.44, accuracy_test: 77.3%\n",
      "Epoch: 1160, loss_train: 331.49, accuracy_train: 96.5%\n",
      "\t  loss_test: 526.55, accuracy_test: 77.3%\n",
      "Epoch: 1161, loss_train: 331.30, accuracy_train: 96.5%\n",
      "\t  loss_test: 526.66, accuracy_test: 77.3%\n",
      "Epoch: 1162, loss_train: 331.10, accuracy_train: 96.5%\n",
      "\t  loss_test: 526.78, accuracy_test: 77.3%\n",
      "Epoch: 1163, loss_train: 330.90, accuracy_train: 96.5%\n",
      "\t  loss_test: 526.89, accuracy_test: 77.3%\n",
      "Epoch: 1164, loss_train: 330.71, accuracy_train: 96.5%\n",
      "\t  loss_test: 527.00, accuracy_test: 77.3%\n",
      "Epoch: 1165, loss_train: 330.51, accuracy_train: 96.5%\n",
      "\t  loss_test: 527.12, accuracy_test: 77.3%\n",
      "Epoch: 1166, loss_train: 330.32, accuracy_train: 96.5%\n",
      "\t  loss_test: 527.23, accuracy_test: 77.3%\n",
      "Epoch: 1167, loss_train: 330.13, accuracy_train: 96.5%\n",
      "\t  loss_test: 527.34, accuracy_test: 77.3%\n",
      "Epoch: 1168, loss_train: 329.93, accuracy_train: 96.5%\n",
      "\t  loss_test: 527.45, accuracy_test: 77.3%\n",
      "Epoch: 1169, loss_train: 329.74, accuracy_train: 96.5%\n",
      "\t  loss_test: 527.57, accuracy_test: 77.3%\n",
      "Epoch: 1170, loss_train: 329.55, accuracy_train: 96.5%\n",
      "\t  loss_test: 527.68, accuracy_test: 77.3%\n",
      "Epoch: 1171, loss_train: 329.36, accuracy_train: 96.5%\n",
      "\t  loss_test: 527.79, accuracy_test: 77.3%\n",
      "Epoch: 1172, loss_train: 329.17, accuracy_train: 96.5%\n",
      "\t  loss_test: 527.91, accuracy_test: 77.3%\n",
      "Epoch: 1173, loss_train: 328.98, accuracy_train: 96.5%\n",
      "\t  loss_test: 528.02, accuracy_test: 77.3%\n",
      "Epoch: 1174, loss_train: 328.80, accuracy_train: 96.5%\n",
      "\t  loss_test: 528.13, accuracy_test: 77.3%\n",
      "Epoch: 1175, loss_train: 328.61, accuracy_train: 96.5%\n",
      "\t  loss_test: 528.24, accuracy_test: 77.3%\n",
      "Epoch: 1176, loss_train: 328.42, accuracy_train: 96.5%\n",
      "\t  loss_test: 528.35, accuracy_test: 77.3%\n",
      "Epoch: 1177, loss_train: 328.24, accuracy_train: 96.5%\n",
      "\t  loss_test: 528.47, accuracy_test: 77.3%\n",
      "Epoch: 1178, loss_train: 328.05, accuracy_train: 96.5%\n",
      "\t  loss_test: 528.58, accuracy_test: 77.3%\n",
      "Epoch: 1179, loss_train: 327.87, accuracy_train: 96.5%\n",
      "\t  loss_test: 528.69, accuracy_test: 77.3%\n",
      "Epoch: 1180, loss_train: 327.68, accuracy_train: 96.5%\n",
      "\t  loss_test: 528.80, accuracy_test: 77.3%\n",
      "Epoch: 1181, loss_train: 327.50, accuracy_train: 96.5%\n",
      "\t  loss_test: 528.92, accuracy_test: 77.3%\n",
      "Epoch: 1182, loss_train: 327.32, accuracy_train: 96.5%\n",
      "\t  loss_test: 529.03, accuracy_test: 77.3%\n",
      "Epoch: 1183, loss_train: 327.14, accuracy_train: 96.5%\n",
      "\t  loss_test: 529.14, accuracy_test: 77.1%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 1184, loss_train: 326.96, accuracy_train: 96.5%\n",
      "\t  loss_test: 529.25, accuracy_test: 77.1%\n",
      "Epoch: 1185, loss_train: 326.77, accuracy_train: 96.5%\n",
      "\t  loss_test: 529.36, accuracy_test: 77.1%\n",
      "Epoch: 1186, loss_train: 326.60, accuracy_train: 96.5%\n",
      "\t  loss_test: 529.47, accuracy_test: 77.1%\n",
      "Epoch: 1187, loss_train: 326.42, accuracy_train: 96.5%\n",
      "\t  loss_test: 529.59, accuracy_test: 77.1%\n",
      "Epoch: 1188, loss_train: 326.24, accuracy_train: 96.5%\n",
      "\t  loss_test: 529.70, accuracy_test: 77.1%\n",
      "Epoch: 1189, loss_train: 326.06, accuracy_train: 96.5%\n",
      "\t  loss_test: 529.81, accuracy_test: 77.1%\n",
      "Epoch: 1190, loss_train: 325.88, accuracy_train: 96.5%\n",
      "\t  loss_test: 529.92, accuracy_test: 77.1%\n",
      "Epoch: 1191, loss_train: 325.71, accuracy_train: 96.5%\n",
      "\t  loss_test: 530.03, accuracy_test: 77.1%\n",
      "Epoch: 1192, loss_train: 325.53, accuracy_train: 96.5%\n",
      "\t  loss_test: 530.14, accuracy_test: 77.1%\n",
      "Epoch: 1193, loss_train: 325.36, accuracy_train: 96.5%\n",
      "\t  loss_test: 530.25, accuracy_test: 77.1%\n",
      "Epoch: 1194, loss_train: 325.18, accuracy_train: 96.5%\n",
      "\t  loss_test: 530.36, accuracy_test: 77.1%\n",
      "Epoch: 1195, loss_train: 325.01, accuracy_train: 96.5%\n",
      "\t  loss_test: 530.48, accuracy_test: 77.1%\n",
      "Epoch: 1196, loss_train: 324.83, accuracy_train: 96.5%\n",
      "\t  loss_test: 530.59, accuracy_test: 77.1%\n",
      "Epoch: 1197, loss_train: 324.66, accuracy_train: 96.5%\n",
      "\t  loss_test: 530.70, accuracy_test: 77.1%\n",
      "Epoch: 1198, loss_train: 324.49, accuracy_train: 96.5%\n",
      "\t  loss_test: 530.81, accuracy_test: 77.1%\n",
      "Epoch: 1199, loss_train: 324.32, accuracy_train: 96.5%\n",
      "\t  loss_test: 530.92, accuracy_test: 77.1%\n",
      "Epoch: 1200, loss_train: 324.15, accuracy_train: 96.5%\n",
      "\t  loss_test: 531.03, accuracy_test: 77.1%\n"
     ]
    }
   ],
   "source": [
    "mlp.train_minibatch(x=x_train, y_true=y_train, x_test=x_test, y_test=y_test, epochs=1200, batch_size=256)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "accuracies_train = [x[0] for x in mlp.accuracies]\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax.plot(accuracies_train)\n",
    "\n",
    "ax.set(xlabel='epochs', ylabel='accuracy [%]', title='Accuracies on Train Set')\n",
    "ax.grid()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "loss_train = [x[0] for x in mlp.losses]\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax.plot(loss_train)\n",
    "\n",
    "ax.set(xlabel='epochs', ylabel='loss', title='Loss on Train set')\n",
    "ax.grid()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "accuracies_test = [x[1] for x in mlp.accuracies]\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax.plot(accuracies_test)\n",
    "\n",
    "ax.set(xlabel='epochs', ylabel='accuracy [%]', title='Accuracies on Test Set')\n",
    "ax.grid()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAIABJREFUeJzt3Xl8VfWd//HXJ3vIQgJZScCwR2STIO5UtC7QsVptq13cplOdme6d6ahtp+10mXHaTlv766atzujUSju41rWKYN0VEFD2HROWhJ0QCFk+vz/uiaY0hBByc+7yfj4e93HvPfeck8+XE+475/s9i7k7IiIiR0oJuwAREYlNCggREemSAkJERLqkgBARkS4pIEREpEsKCBER6ZICQkREuqSAkJhlZhvN7P1h19FbZvakmTUGjxYzO9zp/a9OYL23mdlv+rLWYL3VZtba1+uV+JUWdgEiicrdZ3a8NrP/AWrd/evhVSRyfLQHIXHJzD5tZmvNbJeZPWpmQ4LpZmY/NrN6M9trZkvNbHzw2SwzW25m+82szsz++SjrTjGzr5vZpmA995rZwOCzKjNzM7vOzDab2Q4z+9oJtONDQY17zOwFMxvX6bN/NbOtZrbPzFaY2blmdjnwZeC6YE/k9aOs96+WDaanBp+tD2q/z8wKgsX+DKR22ss5tbftksSggJC4Y2bnA/8BfBQoBzYBs4OPLwKmA2OAAuAqYGfw2V3ATe6eB4wHnjvKj7g+eMwARgC5wM+OmOccYCxwAfANMzu5F+04A/gFcAMwGPhf4GEzSzOzScH0ycBA4ANE9kAeBn4E3OPuue4+rYv1drls8PFXiPwbnQNUAi3Aj4PPpgNtwXpz3f3N422TJBYFhMSjTwB3u/sid28GbgXONLMqIl94eUA1YO6+wt23Bsu1AOPMLN/dd7v7om7W/yN3X+/ujcH6rzazzl2y/+buB919CbAEmNSLdtwE/MzdF7p7m7vfCWQCNUArkA2MA1KDWjb0cL3dLXsTcIu7b3H3Q8C/AVeZmfWifklwCgiJR0OI7DUAEHyJ7wQq3P05In/t/xzYbmZ3mll+MOuVwCxgk5k9b2Zn9mT9wes0oLTTtG2dXjcR2cs4XicBXw26l/aY2R6gOGjHMuAW4HtAfdAVVNrdyjocbdkgBIYCT3T6eW8S+R4Y3Iv6JcEpICQebSHy5QqAmeUQ+YKrA3D3n7p7DXAKka6mrwTT33D3y4AS4GHgDz1ZPzCMyF/l2/u2GbwDfMPdCzo9Brj7g0G997j7WUS6ubKA7wbLHfMSzF0t65FLN9cB5x/xM7PcfUdP1ivJRQEhsS7dzLI6PdKA3wE3mNlkM8sE/h14zd03mtlpZna6maUDB4BDQJuZZZjZJ8xsoLu3APuAtqP8zPuBL5nZcDPLDdb/e3fv60NA7wQ+Z2ZTg8H1XDP7oJkNMLNxZva+oH0Hg0dHvduB4UfrFjrGsr8CbjOzocG8JWZ2afBZPZFB6mF93E6JUwoIiXVP8N6X3EHgW+4+F/hX4AFgKzASuDqYPx/4NbCbSNfQTuCHwWfXABvNbB/w98Anj/Iz7yYyYPxnYAORkPlcn7YKcPeXgM8DdwB7gNXAx4n8JZ8N/Bewg0gbc4FvBIvOBgYAu8zs5S5W3d2y3weeBZ4zs/3Ay8CUoJ7dwecLgy6oyX3ZXok/phsGiYhIV7QHISIiXVJAiIhIlxQQIiLSJQWEiIh0Ka4v1ldUVORVVVW9WvbAgQPk5OT0bUEhUVtiU6K0JVHaAWpLh4ULF+5w9+JjzRfXAVFVVcWCBQt6tez8+fM577zz+ragkKgtsSlR2pIo7QC1pYOZbTr2XOpiEhGRo1BAiIhIlxQQIiLSJQWEiIh0SQEhIiJdUkCIiEiXFBAiItKluD4PoreaW9t4dlMLS+euITXFyEhNISMthcy0FDLTU8hMSyUj9b3XHdMj0yLvczPTyExLQXdqFJFElZQB8fCbdfx2xWFYsfqE1pORmkJ+djoDs9OC53QG52RSUZDFkIJshhRkUzU4h8rCbFJSFCQiEl+SMiA+UjOUvbVr+NRl59PS1k5LWzuHW9s53NZOc0s7za3tNLe2cbj1vdfNLe99fqi1jQPNbew92MLegy3sO9TCvoMt7DpwmFXb9rN93yHaO91mIy8zjZPL85k0dCDnji5m2vBBZKWnhvcPICLSA0kZECkpxpjCVFJTjNSU1D7/sm5pa6d+fzN1uw+yvqGRZVv2sWzLXu55eRO/fmED2empzJpQzsemDaXmpEJ1U4lITErKgIi29NQUKgqyqSjIZtrwQe9OP3i4jVc37ORPy7bxxyVbeWBRLWeOGMzNM6uZPLQgxIpFRP6ajmLqR9kZqcwYW8J/XDGR1792Ad/4m3Gs3r6fD/3iJf7jyRU0t7YdeyUiIv1EARGSARlp/O05w3n+X2Zw9WnDuOP59Vx71+vsPdgSdmkiIoACInS5mWn8xxUTuP3qySzavJur7niFvU0KCREJnwIiRlw2uYK7rz+N9Q0H+PS9CzjUou4mEQmXAiKGnDu6mP/66CRe37iL255cGXY5IpLkFBAx5tJJQ7j+rCr+5+WNzF9VH3Y5IpLEFBAx6JaZ1YwqyeUbjyxTV5OIhEYBEYOy0lP55qXj2Lyribte3BB2OSKSpBQQMerc0cW8/+RSfvX8OvYd0lFNItL/FBAx7AsXjGb/oVZ+++qmsEsRkSSkgIhhEyoHcu7oIu56YYPGIkSk3ykgYtyN00ew88Bh/rR8e9iliEiSUUDEuLNHFlFZmM3v39gcdikikmQUEDEuJcW4aupQXlq7k807m8IuR0SSiAIiDlxZUwnAo0vqQq5ERJJJ1ALCzMaa2eJOj31m9kUzG2Rmz5jZmuC5MJjfzOynZrbWzJaa2ZRo1RZvhhRkc+qwAp58e1vYpYhIEolaQLj7Knef7O6TgRqgCXgIuAWY6+6jgbnBe4CZwOjgcSPwy2jVFo9mji9j2ZZ96mYSkX7TX11MFwDr3H0TcBlwTzD9HuDy4PVlwL0e8SpQYGbl/VRfzJs5PvJP8dSyrSFXIiLJwtw9+j/E7G5gkbv/zMz2uHtBp892u3uhmT0G3ObuLwbT5wI3u/uCI9Z1I5E9DEpLS2tmz57dq5oaGxvJzc3tZYvC8a8vHSQ3HW6elv0X0+OxLUejtsSeRGkHqC0dZsyYsdDdpx5zRneP6gPIAHYApcH7PUd8vjt4fhw4p9P0uUBNd+uuqanx3po3b16vlw3Lvz++3Ed99XE/0NzyF9PjsS1Ho7bEnkRph7va0gFY4D34/u6PLqaZRPYeOs702t7RdRQ8d1zTuhYY2mm5SmBLP9QXN84dXUxLm/Pa+l1hlyIiSaA/AuJjwP2d3j8KXBe8vg54pNP0a4Ojmc4A9rq7Otw7mVpVSGZaCn9e0xB2KSKSBNKiuXIzGwBcCNzUafJtwB/M7FPAZuAjwfQngFnAWiJHPN0QzdriUVZ6KqePGMyLa3aEXYqIJIGoBoS7NwGDj5i2k8hRTUfO68BnollPIjhjxCC+/9Qqdh04zKCcjLDLEZEEpjOp48xpVYMAWLhpd8iViEiiU0DEmQkVA8lITWHBRg1Ui0h0KSDiTFZ6KhMrB/KGAkJEokwBEYemVg3irbq9uomQiESVAiIO1ZxUSEubs2zL3rBLEZEEpoCIQxMqBgLwVq0CQkSiRwERh0rzMynOy2RpnQJCRKJHARGHzIwJFQO1ByEiUaWAiFMTKgayrqGRA82tYZciIglKARGnJlQMpN1h+dZ9YZciIglKARGnJlRqoFpEoksBEadK87MoycvkLQ1Ui0iUKCDi2Lgh+axQF5OIRIkCIo6NLctjXUMjre3Rv22siCQfBUQcqy7Lo6XN2X5AASEifU8BEcfGluYD8E5je8iViEgiUkDEsZElOaSmGLX7FRAi0vcUEHEsMy2VEUU5CggRiQoFRJwbW5ZHrbqYRCQKFBBxrrosjx0HnUZdckNE+pgCIs6NLYsMVK/atj/kSkQk0Sgg4lx1WR6ggBCRvqeAiHMVBdlkpMLa+sawSxGRBKOAiHMpKUZ5TgrrGhQQItK3FBAJoDzHtAchIn1OAZEAhuSmULfnIE2HdSSTiPQdBUQCKM+JbMb1DQdCrkREEokCIgEMCQJC4xAi0pcUEAmgJMdIMR3JJCJ9SwGRANJTjJMG52gPQkT6lAIiQYwsztEehIj0KQVEghhZksvGHU20tunCfSLSN6IaEGZWYGZzzGylma0wszPN7FtmVmdmi4PHrE7z32pma81slZldHM3aEs3I4lwOt7Xzzu6DYZciIgki2nsQtwNPuXs1MAlYEUz/sbtPDh5PAJjZOOBq4BTgEuAXZpYa5foSxqiSXEAD1SLSd6IWEGaWD0wH7gJw98PuvqebRS4DZrt7s7tvANYC06JVX6IZWRwJCA1Ui0hfMffo3PDezCYDdwLLiew9LAS+AHwFuB7YBywA/sndd5vZz4BX3f23wfJ3AU+6+5wj1nsjcCNAaWlpzezZs3tVX2NjI7m5ub1aNtZ0tOUL85qYUJTK303IDLukXkvE7RLvEqUdoLZ0mDFjxkJ3n3rMGd09Kg9gKtAKnB68vx34DlAKpBLZe/kecHfw+c+BT3Za/i7gyu5+Rk1NjffWvHnzer1srOloy1V3vOyX/ezFcIs5QYm4XeJdorTDXW3pACzwHnyPR3MMohaodffXgvdzgCnuvt3d29y9Hfg173Uj1QJDOy1fCWyJYn0JZ1RJLusbGjsCVkTkhEQtINx9G/COmY0NJl0ALDez8k6zfQh4O3j9KHC1mWWa2XBgNPB6tOpLRCOLc9l3qJWGxuawSxGRBJAW5fV/DrjPzDKA9cANwE+D8QkHNgI3Abj7MjP7A5Exi1bgM+7eFuX6Esq7A9X1ByjJywq5GhGJd1ENCHdfTGQsorNrupn/e0TGJaQXRpa8dyTTmSMHh1yNiMQ7nUmdQMrzsxiQkapDXUWkTyggEkhKijGiOId1ui+EiPQBBUSCGVmcyzqdTS0ifUABkWBGFufq9qMi0icUEAmm40gm3X5URE6UAiLBjCrRNZlEpG8oIBLMSYMHkGJooFpETpgCIsFkpacydNAADVSLyAlTQCSgkcW56mISkROmgEhAo0pyWb/jAG3tumifiPSeAiIBjSzO4XBrO3W6/aiInAAFRALS3eVEpC8oIBJQR0Do/tQiciIUEAmoMCeDQTkZ2oMQkROigEhQo3Qkk4icIAVEghpZoqu6isiJUUAkqJHFuew6cJhdBw6HXYqIxCkFRILSkUwicqIUEAnqvftTKyBEpHcUEAmqojCbzLQU7UGISK8pIBJUaooxvEgD1SLSewqIBDayRIe6ikjvKSAS2MjiXN7Z1cShlrawSxGROKSASGBjS/Nod11yQ0R6p0cBYWZfMLN8i7jLzBaZ2UXRLk5OTHV5HgArt+0PuRIRiUc93YP4W3ffB1wEFAM3ALdFrSrpE1WDc8hMS2Hl1n1hlyIicainAWHB8yzgv919SadpEqNSU4wxpXnagxCRXulpQCw0sz8RCYinzSwPaI9eWdJXqssUECLSOz0NiE8BtwCnuXsTkE6km0liXHV5Pjsam2nY3xx2KSISZ3oaEGcCq9x9j5l9Evg6sDd6ZUlfqS6LDFSv0l6EiBynngbEL4EmM5sE/AuwCbg3alVJn+kIiJXbNFAtIsenpwHR6u4OXAbc7u63A3nRK0v6yuDcTIrzMjUOISLHracBsd/MbgWuAR43s1Qi4xDdMrMCM5tjZivNbIWZnWlmg8zsGTNbEzwXBvOamf3UzNaa2VIzm9L7ZklnkYFq7UGIyPHpaUBcBTQTOR9iG1AB/KAHy90OPOXu1cAkYAWRwe657j4amBu8B5gJjA4eNxLp1pI+UF2Wx5rtjbS26cAzEem5HgVEEAr3AQPN7G+AQ+7e7RiEmeUD04G7gnUcdvc9RLqp7glmuwe4PHh9GXCvR7wKFJhZ+fE2SP7auCH5NLe2s36HruwqIj1nkaGFY8xk9lEiewzziZwgdy7wFXef080yk4E7geVE9h4WAl8A6ty9oNN8u9290MweA25z9xeD6XOBm919wRHrvZHIHgalpaU1s2fP7nlrO2lsbCQ3N7dXy8aaY7VlS2M7X33xIH83IYNzKo7ZMxiqZNou8SJR2gFqS4cZM2YsdPepx5zR3Y/5AJYAJZ3eFwNLjrHMVKAVOD14fzvwHWDPEfPtDp4fB87pNH0uUNPdz6ipqfHemjdvXq+XjTXHaktbW7uP+9cn/RsPv9U/BZ2AZNou8SJR2uGutnQAFngPvvt7OgaR4u71nd7v5NjdU7VArbu/FryfA0wBtnd0HQXP9Z3mH9pp+UpgSw/rk26kpBjjKwaytE6nrohIz/U0IJ4ys6fN7Hozu57IX/tPdLeAR8Yt3jGzscGkC4h0Nz0KXBdMuw54JHj9KHBtcDTTGcBed9/a86ZIdyZWDmT5ln20aKBaRHoorSczuftXzOxK4GwiYxB3uvtDPVj0c8B9ZpYBrCdyeY4U4A9m9ilgM/CRYN4niFzraS3QhC7l0acmVBbQ3LqBNdsbGTckP+xyRCQO9CggANz9AeCB41m5uy8mMhZxpAu6mNeBzxzP+qXnJlYMBOCtuj0KCBHpkW67mMxsv5nt6+Kx38x05lUcOWnwAPKy0lhaq3EIEemZbvcg3F2X00gQZsbEyoEKCBHpMd2TOolMHlrA8q37aDrcGnYpIhIHFBBJZGrVINrancXv7Am7FBGJAwqIJDJlWCFmsGDj7rBLEZE4oIBIIgOz0xlbmscbG3eFXYqIxAEFRJKpOamQNzfvoa392NfgEpHkpoBIMqdVDaKxuVX3hxCRY1JAJJmakwoBWLhJ4xAi0j0FRJKpLMymLD+L19ZrHEJEuqeASDJmxlmjBvPyuh20axxCRLqhgEhC544uYndTC8u3ahxCRI5OAZGEzh5ZBMALa3aEXImIxDIFRBIqyc9ibGkeL61VQIjI0SkgktQ5o4t4feMuDrW0hV2KiMQoBUSSOmd0EYdb23VWtYgclQIiSZ0+fBAZaSnMW9kQdikiEqMUEElqQEYa54wq4pkV24jczE9E5C8pIJLYheNKeWfXQVZt3x92KSISgxQQSeyCk0swg2eWbQ+7FBGJQQqIJFaSl8XkoQU8s0IBISJ/TQGR5C4cV8rS2r1s3Xsw7FJEJMYoIJLczPHlADy+dGvIlYhIrFFAJLnhRTlMrBzIw4vrwi5FRGKMAkK4bHIFb9ftY219Y9iliEgMUUAIl04sxwwe1V6EiHSigBBK8rM4a+RgHl68RSfNici7FBACwBWnVrJ5VxOvrN8ZdikiEiMUEALAByaWk5+Vxu9e2xx2KSISIxQQAkBWeipXTKnk6WXb2NnYHHY5IhIDFBDyro+fPoyWNueBRbVhlyIiMUABIe8aU5rHaVWF3PvKJlrb2sMuR0RCFtWAMLONZvaWmS02swXBtG+ZWV0wbbGZzeo0/61mttbMVpnZxdGsTbr2qXOGU7v7IE/rAn4iSS+tH37GDHc/8ubHP3b3H3aeYGbjgKuBU4AhwLNmNsbddU/MfnThuDKqBg/gzj+vY9aEMsws7JJEJCSx1MV0GTDb3ZvdfQOwFpgWck1JJzXF+NS5I1hSu5c3Nu4OuxwRCZFF88QoM9sA7AYcuMPd7zSzbwHXA/uABcA/uftuM/sZ8Kq7/zZY9i7gSXefc8Q6bwRuBCgtLa2ZPXt2r2prbGwkNze3V8vGmr5uS3Ob80/zmxhZkMqXarL6bL09oe0SexKlHaC2dJgxY8ZCd596zBndPWoPYEjwXAIsAaYDpUAqkb2X7wF3B/P8HPhkp2XvAq7sbv01NTXeW/Pmzev1srEmGm356bOr/aSbH/PFm3f3+bq7o+0SexKlHe5qSwdggffgOzyqXUzuviV4rgceAqa5+3Z3b3P3duDXvNeNVAsM7bR4JbAlmvXJ0d1wznAKB6Tzo2dWh12KiIQkagFhZjlmltfxGrgIeNvMyjvN9iHg7eD1o8DVZpZpZsOB0cDr0apPupebmcbfv28kz69uYMHGXWGXIyIhiOYeRCnwopktIfJF/7i7PwV8Pzj0dSkwA/gSgLsvA/4ALAeeAj7jOoIpVNeeWUVRbibff3qVLuInkoSidpiru68HJnUx/ZpulvkekXEJiQHZGal88f2j+frDb/PU29uYOaH82AuJSMKIpcNcJQZ9bNowqsvy+O7jKzjUoh06kWSigJBupaYY37z0FOr2HOTXf14fdjki0o8UEHJMZ44czKwJZfxi/jre2dUUdjki0k8UENIjX/vAOFIMbn3wLQ1YiyQJBYT0SEVBNrfMOpkX1+7g/xbocuAiyUABIT32iWnDmDZ8EN95fDnb9h4KuxwRiTIFhPRYSorxn1dOpLXN+fIfFtPWrq4mkUSmgJDjMrwoh299cBwvr9vJL+evDbscEYkiBYQct49OHcqlk4bw42fX6DIcIv2svd15Zd1OtjRG/66PCgg5bmbG9z40noqCbP7xvkUajxDpB5t2HuBHz6xm+g/m8bFfv8qzm1ui/jP7445ykoDys9K589oarvjFy9z0vwv4/U1nkpWeGnZZIgll/6EWnnhrK3MW1vLGxt2YwTmjivjKxWPJ2hn9Ky0rIKTXqsvy+fFVk7npfxdy8wNL+clVk3WLUpET1NbuvLxuBw8srOWpZds41NLOiOIcvnLxWK6YUkH5wGwA5s9fE/VaFBByQi4+pYx/vmgMP/zTaioLs/nKxdVhlyQSl9Y1NPLAwloeerOOrXsPkZ+VxpVTKvlwTSWThxaE8seXAkJO2GdmjKJuz0F+Pm8dhQMy+LtzR4Rdkkhc2NvUwh+XbmHOwloWv7OH1BRj+ugivv6BcVxwckno3bYKCDlhZsZ3L5/A3oMtfPfxFeRnp/PRqUOPvaBIEmpta+eFNTuYs7CWZ1Zs53BrO2NL8/jarJO57NQhlOT1733gu6OAkD6RmmL8+KrJ7D+0gJsfWEprm/Px04eFXZZIzFi1bT8PLIp0ITXsb6ZwQDofnzaMD9dUcsqQ/Jgcv1NASJ/JTEvl19dO5R9+u5CvPvQWh1ra+Ntzhoddlkhodh04zKOL65izqJa36/aRlmLMqC7hwzWVzBhbQkZabJ9poICQPpWVnsod10zl8/e/ybcfW87egy188f2jY/KvI5FoONzazvxV9cxZWMu8VfW0tDmnDMnnm5eO44OThjA4NzPsEntMASF9LiMthZ99/FRuffAtbp+7hg07DvD9D08MfcBNJFrcnWVb9jFnYS2PLtnCrgOHKcrN5PqzqriyppLqsvywS+wVBYRERVpqCt//8ESGF+fw/adWUbu7iV99soaS/NgZgBM5UQ37m3lkcR1zFtayctt+MlJTuHBcKVfWVDB9dDFpqbHdhXQsCgiJGjPjH88bxfDBOXz5D0uY9dMX+MlVp3LO6KKwSxPptebWNuauiHQhPb+6gbZ2Z9LQAr5z+XgunVhOwYCMsEvsMwoIibqZE8oZWZLLZ+5bxDV3v8ZnZ4zi8xeMJj3O/7qS5OHuLKndywNBF9Legy2U5mdy4/QRXDmlglEleWGXGBUKCOkXY0rzeOSzZ/ONR5bx/55by9wV9Xz/wxMZXzEw7NJEjmrb3kM89GYdcxa+w7qGA2SmpXDJ+DKunFLJ2aOKSE1J7IMvFBDSbwZkpPHDj0zi/SeX8vWH3+byn7/E379vJJ+ZMYrsDA1gS2w41NLG08u2MWdhLS+t3UG7w2lVhXz63BHMmlhOflZ62CX2GwWE9LtLxpdxxohBfPux5fxs3loeXFTLLbNO5tKJ5WGXJknK3Vm4aTdzFtby+NKt7G9upaIgm8/OGMUVUyqpKsoJu8RQKCAkFAUDMvjRRyfzkZqhfOex5Xz+/je59+WNXFzWxnlhFydJo3Z3Ew8uquPBRbVs3NnEgIxUZo4v58qaCs4YPpiUBO9COhYFhITqzJGD+ePnzmHOwnf4wdOr+d6mZubteJXPXzCaM0YMDrs8SUAHmlt56u1IF9Ir63cCcOaIwXz2/NHMHF9GTqa+FjvoX0JCl5piXHXaMD44qYJv/+45nq1r5Oo7X2Va1SBuOLuKC8eVxv3x5BKu9nbntQ27+PXSZv7xuWdpOtzGSYMH8OULx/ChUysYOmhA2CXGJAWExIzsjFQurkrnm584l/tf38xvXtjAP9y3iPKBWXzyjJO46rShFMXRZQokfBt2HOChRbU8+GYdtbsPkpUKl08ZypU1lUw9qVCXgDkGBYTEnKz0VG44ezjXnlnF3BXbueeVjfzg6VX8+JnVnDe2mCumVHJ+dfjXypfYtKfpMH9cupUHF9Xy5uY9pBic3ek2nRdfMDHsEuOGAkJiVmqKcdEpZVx0Shlrtu9nTnC3rWdX1JOflcYHJg5h1oQyzhgxWCfdJbnDre3MW1XPQ4vqmLtyOy1tztjSPL46q5rLJldQGlzipT9u05lIFBASF0aX5nHrrJP5l0uqeWntDh56s45HFtdx/+ubyc9K4/0nl3LRKWW8b0yxzqlIEh1nNz+4qJY/LtnC7qYWinIzufbMKq6YUsG48ti8x0I8iWpAmNlGYD/QBrS6+1QzGwT8HqgCNgIfdffdFtmStwOzgCbgendfFM36JP6kphjTxxQzfUwxh1raeGHNDp56extzV27nwTfryEpP4eyRRZxXXcJ5Y4o1+JiAanc38fCbdTy4qI71OyJnN190ShlXTKng3FFFOqChD/XHHsQMd9/R6f0twFx3v83Mbgne3wzMBEYHj9OBXwbPIl3KSk/lwnGlXDiulJa2dl7fsIs/LdvGc6vqmbuyHoCRxTnMGFvCeWNLOG14IZlp2ruIRzsam3nira08ungLCzbtBuD04YO46X0jmDkhuc5u7k9hdDFdBu+eC3UPMJ9IQFwG3OvuDrxqZgVmVu7uW0OoUeJMemoKZ48q4uxRRXzLnfU7DjB/VQPzV9Vz7yub+M2LGxiQkcpZIwdz3tgSZlSXUFGQHXbZ0o29B1t4etk2/rhky7uXvBhbmsc/XzSGyybr0NT+YJHv4yit3GwDsBtw4A53v9PM9rh7Qad5drt7oZlNPWl6AAANE0lEQVQ9Btzm7i8G0+cCN7v7giPWeSNwI0BpaWnN7Nmze1VbY2Mjubm5vVo21qgt3WtudVbsamNpQxtLd7Sx42Dkd74y15hYnMak4lRGFqSQ1sdnzSbKdunPdjS3OYvr23h1aytvNbTR6lCcbZxensYZ5WlU5p1Y91GibBM4sbbMmDFjobtPPdZ80d6DONvdt5hZCfCMma3sZt6u/nf+VXq5+53AnQBTp0718847r1eFzZ8/n94uG2vUlmO7OHh2d9Y1NDJvZQPzVtXzpw27eGJDC3lZaUwfUxx0RxX3yfkWibJdot2OA82tPL+6gSff3sbcFdtpOtxGaX4m1509jEsnDWFS5cA+G2xOlG0C/dOWqAaEu28JnuvN7CFgGrC9o+vIzMqB+mD2WmBop8UrgS3RrE+Sj5kxqiSPUSV5fHr6CPYfauGltTt4bmU981Y18PjSrZjBxIqBzKguYcbYEiZUDEz6a/L0tT1Nh5m7op6nlm3jz6sbaG5tZ3BOBpdNruCDk4YwbfighL+UdjyIWkCYWQ6Q4u77g9cXAd8GHgWuA24Lnh8JFnkU+KyZzSYyOL1X4w8SbXlZ6VwyvpxLxpfT3u4s37qPeSvrmbeqntvnruEnz66hKDeD940p4fzqEs4dU6QB0V6q33eIPy3fztPLtvHKup20tjvlA7P42LRhXDK+jNOqFAqxJpp7EKXAQ8GuYRrwO3d/yszeAP5gZp8CNgMfCeZ/gsghrmuJHOZ6QxRrE/krKSnG+IqBjK8YyOcuGM2uA4f58+oGnltZz7MrtvPAolrSUozTqgZxfnVkoHtkcY6OtT+KtnZnae2eIHAbeKtuLwDDi3L49PQRXHJKGRP7sPtI+l7UAsLd1wOTupi+E7igi+kOfCZa9Ygcr0E5GVx+agWXn1pBa1s7b76zh+dW1vPcinq+98QKvvfECoYNGsD51ZG9i9NHDEr6w2j3NrXw/JoG5q+sZ/7qBnYdOEyKwZRhhfzzRWO4cFwZY0pzFQpxQmdSi/RAWmoKp1UN4rSqQdx8STW1u5uYt6qBeSvruf/1zfzPyxsZkJHK2aOKInsXY0vCLrlfHGhu5Y2Nu3hl3U5eWb+Tt+v20u5QOCCd940pZkZ1CdNHF1OYkxF2qdILCgiRXqgsHMA1Z5zENWecxMHDbbyyPhjoXtnAM8u3AzAsL4WZTSs4Y8RgplYVkpcAYxd7m1pYXLuHBRt38fK6nSx5Zw+t7U56qnHqsEI+d/5o3je2mEmVBRpPSAAKCJETlJ2RyvnVpZxfXYq7s3p7I8+trOfh11Zz90sbuOPP60kxGF8xkNOHD2La8MFMGjqQkryssEvv1uHWdtbU72fe5hYe+78lvLl5N+saDgCQYjCxsoAbp4/gzJGDmXrSIF0DKwEpIET6kJkxtiyPsWV5nMw7nH7Wuby5eTevrt/Jqxt2cc/Lm/j1CxsAKM3PZEJFARMrBzK+Ip9RxXlUFGb3+1/ebe3O1r0HWddwgJVb97Fy235WbN3HuoZGWtoipyINyqlnyrACrphSyalDC5g4tIBc3Xkt4WkLi0RRdkYqZ40q4qxRRQAcamljae1e3qrby9t1e1lau4e5K7fTcUGDjLQUqgYPYGRxLsMGDaAkP4uy/CzKBmZSkpdFfnY6uZlpPQoRd6e5tZ19h1rYsf8wDY3N7NjfzI7GZrbsOcimXU1s3tnEO7ub3g0CgPKBWVSX5TGjuoTqsjyat6ziIzNnaGA5CSkgRPpRVnoq04YPYtrwQe9Oa2xuZcXWfaxvaGR9wwHWNTSyatt+5q6s53Bre5fryU5PJTcrjcy0FMzAOl2IoLm1jabmNg4cbqX9KFfSyctMY9jgAVSX53HRKWWcNHgAw4tyqC7Lo2DAXw4oz9+zRuGQpBQQIiHLzUx79wipztydPU0tbNt3iG37DtGwr5l9h1o40NxGY3MLjc2tNLe0v3s9GnfHiYRHdkYqORlpZGekkpeVRnFuJkV5mRTnZlKcl0mOuoekB/RbIhKjzIzCnAwKczI4uTw/7HIkCenOGiIi0iUFhIiIdEkBISIiXVJAiIhIlxQQIiLSJQWEiIh0SQEhIiJdUkCIiEiXzP0o5+LHATNrADb1cvEiYEcflhMmtSU2JUpbEqUdoLZ0OMndi481U1wHxIkwswXuPjXsOvqC2hKbEqUtidIOUFuOl7qYRESkSwoIERHpUjIHxJ1hF9CH1JbYlChtSZR2gNpyXJJ2DEJERLqXzHsQIiLSDQWEiIh0KSkDwswuMbNVZrbWzG4Ju55jMbOhZjbPzFaY2TIz+0IwfZCZPWNma4LnwmC6mdlPg/YtNbMp4bbgL5lZqpm9aWaPBe+Hm9lrQTt+b2YZwfTM4P3a4POqMOs+kpkVmNkcM1sZbJsz43ibfCn43XrbzO43s6x42S5mdreZ1ZvZ252mHfd2MLPrgvnXmNl1MdSWHwS/Y0vN7CEzK+j02a1BW1aZ2cWdpvfNd5y7J9UDSAXWASOADGAJMC7suo5RczkwJXidB6wGxgHfB24Jpt8C/GfwehbwJGDAGcBrYbfhiPZ8Gfgd8Fjw/g/A1cHrXwH/ELz+R+BXweurgd+HXfsR7bgH+LvgdQZQEI/bBKgANgDZnbbH9fGyXYDpwBTg7U7Tjms7AIOA9cFzYfC6MEbachGQFrz+z05tGRd8f2UCw4PvtdS+/I4L/ZczhA1wJvB0p/e3AreGXddxtuER4EJgFVAeTCsHVgWv7wA+1mn+d+cL+wFUAnOB84HHgv+oOzr9B3h3+wBPA2cGr9OC+SzsNgT15AdfqnbE9HjcJhXAO8GXY1qwXS6Op+0CVB3xpXpc2wH4GHBHp+l/MV+YbTnisw8B9wWv/+K7q2O79OV3XDJ2MXX8Z+hQG0yLC8Hu/KnAa0Cpu28FCJ5LgtliuY0/Af4FaA/eDwb2uHtr8L5zre+2I/h8bzB/LBgBNAD/HXSX/cbMcojDbeLudcAPgc3AViL/zguJz+3S4Xi3Q8xunyP8LZE9IOiHtiRjQFgX0+LiWF8zywUeAL7o7vu6m7WLaaG30cz+Bqh394WdJ3cxq/fgs7ClEekK+KW7nwocINKVcTQx25agf/4yIt0UQ4AcYGYXs8bDdjmWo9Ue820ys68BrcB9HZO6mK1P25KMAVELDO30vhLYElItPWZm6UTC4T53fzCYvN3MyoPPy4H6YHqstvFs4INmthGYTaSb6SdAgZmlBfN0rvXddgSfDwR29WfB3agFat39teD9HCKBEW/bBOD9wAZ3b3D3FuBB4Czic7t0ON7tEMvbh2DQ/G+AT3jQb0Q/tCUZA+INYHRwhEYGkUG2R0OuqVtmZsBdwAp3/1Gnjx4FOo62uI7I2ETH9GuDIzbOAPZ27G6Hyd1vdfdKd68i8u/+nLt/ApgHfDiY7ch2dLTvw8H8MfFXnbtvA94xs7HBpAuA5cTZNglsBs4wswHB71pHW+Juu3RyvNvhaeAiMysM9qguCqaFzswuAW4GPujuTZ0+ehS4OjiqbDgwGnidvvyOC3NgKawHkSMZVhMZ6f9a2PX0oN5ziOwiLgUWB49ZRPp95wJrgudBwfwG/Dxo31vA1LDb0EWbzuO9o5hGBL/Ya4H/AzKD6VnB+7XB5yPCrvuINkwGFgTb5WEiR7/E5TYB/g1YCbwN/C+RI2PiYrsA9xMZO2kh8tfzp3qzHYj0768NHjfEUFvWEhlT6Pi//6tO838taMsqYGan6X3yHadLbYiISJeSsYtJRER6QAEhIiJdUkCIiEiXFBAiItIlBYSIiHRJASHSj8zsPAuuYisS6xQQIiLSJQWESBfM7JNm9rqZLTazOyxyD4tGM/svM1tkZnPNrDiYd7KZvdrpev0d9x4YZWbPmtmSYJmRwepz7b37SNwXnL2Mmd1mZsuD9fwwpKaLvEsBIXIEMzsZuAo4290nA23AJ4hcxG6Ru08Bnge+GSxyL3Czu08kcnZux/T7gJ+7+yQi1zbquLTGqcAXiVzPfwRwtpkNInIp51OC9Xw3uq0UOTYFhMhfuwCoAd4ws8XB+xFELlH++2Ce3wLnmNlAoMDdnw+m3wNMN7M8oMLdHwJw90P+3nV0Xnf3WndvJ3LphCpgH3AI+I2ZXQF0vuaOSCgUECJ/zYB73H1y8Bjr7t/qYr7urlPT1SWXOzR3et1G5KY8rcA0IlfsvRx46jhrFulzCgiRvzYX+LCZlcC79zc+icj/l46rm34ceNHd9wK7zezcYPo1wPMeuV9HrZldHqwj08wGHO0HBvf6GOjuTxDpfpocjYaJHI+0Y88iklzcfbmZfR34k5mlELmy5meI3BToFDNbSOQualcFi1wH/CoIgPXADcH0a4A7zOzbwTo+0s2PzQMeMbMsInsfX+rjZokcN13NVaSHzKzR3XPDrkOkv6iLSUREuqQ9CBER6ZL2IEREpEsKCBER6ZICQkREuqSAEBGRLikgRESkS/8fRy1a3c5afZgAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "loss_test = [x[1] for x in mlp.losses]\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax.plot(loss_test)\n",
    "\n",
    "ax.set(xlabel='epochs', ylabel='loss', title='Loss on Test set')\n",
    "ax.grid()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "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": "3.7.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}