PaperPlots_OLD.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib import gridspec\n",
    "from scipy.stats import norm \n",
    "import energyflow as ef #needed for the Geant4 datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/rikab/miniconda3/lib/python3.8/site-packages/requests/__init__.py:89: RequestsDependencyWarning: urllib3 (1.26.7) or chardet (3.0.4) doesn't match a supported version!\n",
      "  warnings.warn(\"urllib3 ({}) or chardet ({}) doesn't match a supported \"\n"
     ]
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "from keras.layers import Dense, Input\n",
    "from keras.models import Model\n",
    "import tensorflow.keras.backend as K\n",
    "\n",
    "import os\n",
    "# os.environ[\"CUDA_DEVICE_ORDER\"]=\"PCI_BUS_ID\"\n",
    "# os.environ['CUDA_VISIBLE_DEVICES']=\"3\"\n",
    "# gpus = tf.config.experimental.list_physical_devices('GPU')\n",
    "# tf.config.experimental.set_virtual_device_configuration(gpus[0], [tf.config.experimental.VirtualDeviceConfiguration(memory_limit=24220)]) #in MB"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Read in the Delphes Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "w10 = 4.568e-10\n",
    "w14 = 1.380e-11\n",
    "w18 = 4.347e-13"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "myfile = open(\"Datasets/dijets.txt\")\n",
    "delphes_v2 = []\n",
    "for line in myfile:\n",
    "    if (len(line.split()) < 20):\n",
    "        continue\n",
    "    delphes_v2+=[[float(line.split()[0]),float(line.split()[1]),\n",
    "               float(line.split()[2]),float(line.split()[3]),\n",
    "               float(line.split()[4]),float(line.split()[5]),\n",
    "               float(line.split()[6]),float(line.split()[7]),\n",
    "               float(line.split()[8]),float(line.split()[9]),\n",
    "               float(line.split()[10]),float(line.split()[11]),\n",
    "               float(line.split()[12]),float(line.split()[13]),\n",
    "               float(line.split()[14]),float(line.split()[15]),\n",
    "               float(line.split()[16]),float(line.split()[17]),\n",
    "               float(line.split()[18]),float(line.split()[19])]]\n",
    "delphes_v2 = np.array(delphes_v2)\n",
    "\n",
    "myfile = open(\"Datasets/dijets14.txt\")\n",
    "delphes14_v2 = []\n",
    "for line in myfile:\n",
    "    if (len(line.split()) < 20):\n",
    "        continue\n",
    "    delphes14_v2+=[[float(line.split()[0]),float(line.split()[1]),\n",
    "               float(line.split()[2]),float(line.split()[3]),\n",
    "               float(line.split()[4]),float(line.split()[5]),\n",
    "               float(line.split()[6]),float(line.split()[7]),\n",
    "               float(line.split()[8]),float(line.split()[9]),\n",
    "               float(line.split()[10]),float(line.split()[11]),\n",
    "               float(line.split()[12]),float(line.split()[13]),\n",
    "               float(line.split()[14]),float(line.split()[15]),\n",
    "               float(line.split()[16]),float(line.split()[17]),\n",
    "               float(line.split()[18]),float(line.split()[19])]]\n",
    "delphes14_v2 = np.array(delphes14_v2)\n",
    "\n",
    "myfile = open(\"Datasets/dijets18.txt\")\n",
    "delphes18_v2 = []\n",
    "for line in myfile:\n",
    "    if (len(line.split()) < 20):\n",
    "        continue\n",
    "    delphes18_v2+=[[float(line.split()[0]),float(line.split()[1]),\n",
    "               float(line.split()[2]),float(line.split()[3]),\n",
    "               float(line.split()[4]),float(line.split()[5]),\n",
    "               float(line.split()[6]),float(line.split()[7]),\n",
    "               float(line.split()[8]),float(line.split()[9]),\n",
    "               float(line.split()[10]),float(line.split()[11]),\n",
    "               float(line.split()[12]),float(line.split()[13]),\n",
    "               float(line.split()[14]),float(line.split()[15]),\n",
    "               float(line.split()[16]),float(line.split()[17]),\n",
    "               float(line.split()[18]),float(line.split()[19])]]\n",
    "delphes18_v2 = np.array(delphes18_v2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "delphes_v2_all = np.concatenate([delphes_v2,delphes14_v2,delphes18_v2])\n",
    "delphes_weights = np.concatenate([w10*np.ones(len(delphes_v2)),w14*np.ones(len(delphes14_v2)),w18*np.ones(len(delphes18_v2))])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, '$m_{jj}$ [GeV]')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAELCAYAAAA/cjqaAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAAZ+ElEQVR4nO3df5Dc9X3f8efLFx/ImMhY4octne5EpaEhJglmo6atmzJxsYWDIidhJsLuhIk0aHCttplMpoVxpriZ6Yj+SmsbjYlsFMDTSiYqdSQPqeI6TEWmBHRyMEgmss/KCZ0ic2dI5B8hoeB3//h+L6xWu3d7u/vd7372+3rM7Nzu57v7vfdXu9r3fX4rIjAzs2p6U9kBmJlZeZwEzMwqzEnAzKzCnATMzCrMScDMrMKcBMzMKuyHyg5gKVauXBkTExNlh2FmlpSjR49+OyIub3YsqSQwMTHB5ORk2WGYmSVF0qlWx5JoDpK0SdLuc+fOlR2KmdlQSSIJRMTBiNi+fPnyskMxMxsqSSQBMzMrhpOAmVmFOQmYmVVYqUlA0o2SnpB0v6Qby4zFzKyKOk4CkvZImpV0rKF8o6QTkqYk3bXIaQL4HnAxMNNpLGZm1plu5gk8CNwHPDxfIGkE2AXcRPalfkTSAWAE2Nnw+q3AExHxfyRdCfwW8OEu4jEzsyXquCYQEYeBlxuKNwBTEXEyIl4F9gGbI+K5iLil4TYbET/IX/cXwEXNfo+k7ZImJU3Ozc11Gq4lYmx8HEkX3DxT3KwYvZ4xvAo4Xfd4Bvh7rZ4s6ReA9wNvI6tVXCAidgO7AWq1mrdBG3IzL7zAt869ckH5VcuXlRCN2fArddmIiHgUeHSx50naBGxat25d8UFZX4yNjzPzwgsXlK9es6aEaMyqq9dJ4AwwVvd4dV5mdp5Wf/G3snrNGiQ1PTY+Ps709HSPIjOrll4ngSPAeklryb78twAf6vakEXEQOFir1e7o9lyWpsnnTrQ85qYis851M0R0L/AkcI2kGUnbIuI1YAdwCHgeeCQijncbpBeQMzMrRsc1gYi4rUX5Y8BjHUdkZmZ9k8SyEV5F1MysGEkkATcH2ULmO409t8Bs6RSRztD7Wq0W3llsOEha0uigTly1fBkpfb7NiiLpaETUmh1LoiZgZmbFSCIJuDnIzKwYSSQBdwybmRUjiSRg6Wq1IJyXhzAbDKWuHdQurx2UrqUuD9FLrZaa8DITZm/w6CArVD9GAS2VRw1Z1Xh0kJmZNeUkYGZWYU4CZmYVlkQS8DwB6yUvM2H2BncMW6EGsWO4FXcY27Byx7CZmTXlJGBmVmGlJgFJb5L07yR9StLtZcZi3fHMYLM0dTxjWNIe4BZgNiLeVVe+EfgEMAJ8NiLuXeA0m8k2o38JmOk0FitfmTODzaxz3dQEHgQ21hdIGgF2ATcD1wK3SbpW0nWSvthwuwK4Bvi/EfFrwEe6iMWsa+MtajMeNWTDrJs9hg9Lmmgo3gBMRcRJAEn7gM0RsZOs1nAeSTPAq/nD1zuNxawXnnr2T5uWX7V8WZ8jMeufXvcJrAJO1z2eyctaeRR4v6RPAYebPUHSdkmTkibn5uZ6F6mZmZW7imhE/BWwbZHn7JZ0Ftg0Ojp6Q38iMzOrhl7XBM4AY3WPV+dlXfGmMmZmxeh1EjgCrJe0VtIosAU40O1JvWyEmVkxOk4CkvYCTwLXSJqRtC0iXgN2AIeA54FHIuJ4b0K1srWaCzDs8wFajRryyCEbBl47yNqW0jpAvXTlD1/c8pgkrzdkAy/5tYPcHGRmVowkkoA7hs3MipFEEnBNwMysGKXOE2hXRBwEDtZqtTvKjsWq58Xv/HXZIZgVJomagNmg8npDlrokagKSNgGb1q1bV3YoZueZnp5uWi6pv4GYdSiJmoA7hvvLewOYVUcSNQHrL+8NYFYdSdQEPDrIzKwYSdQEPDrIyvSm731rgaMX9S0OsyIkUROwYrjtvzgeNWSpSKImYMVw239xPGrIUuGagJlZhSVRE/A8AUvOd86WHYFZW5KoCXiegJlZMZKoCZj1w8KjgJqb+97fNC2//K0eNWRpSKImYN3xKKDB4VFDNmhKrQlI+kfAh/M4ro2If1BmPMPKo4AGx/SzTzYt1/J39jkSs0w3ewzvkTQr6VhD+UZJJyRNSbproXNExBMRcSfwReChTmMxM7POdFMTeBC4D3h4vkDSCLALuAmYAY5IOgCMADsbXr81Imbz+x8CtnURi5mZdaDjJBARhyVNNBRvAKYi4iSApH3A5ojYCdzS7DyS1gDnIuK7ncZiZmad6XXH8CrgdN3jmbxsIduA32l1UNJ2SZOSJufm5noQotngcYexlaX0IaIRcc8ix3dLOgtsGh0dvaFPYZn1lZeZsLL0OgmcAcbqHq/Oy7riVURt6HmGsZWk181BR4D1ktZKGgW2AAe6Pan3E1hcq7kAng9gZgvpuCYgaS9wI7BS0gxwT0Q8IGkHcIhsRNCeiDjek0htQZ4L0L5OZgabDStFRNkxtK1Wq8Xk5GTZYQwkSU4CbepHEujVshFa/k5S+j9qg0nS0YioNTtWesdwO7yKqKXGawpZKpJYO8iriFpVtRo66uGj1iuuCZgNsFZrDYHXG7LecE3AzKzCkkgC9gYvC21mveTmoMR4KKiZ9VISNQE3B5ldyOsNWS8kURMw68SwTwrzekPWC0nUBLxshJlZMZJIAm4OMjMrhpuDzFLllUetB5KoCZiZWTGcBMzMKiyJ5iDPE7CFpDQKqB8Ly80PHW1W3mpEkVVXEjUBdwybtW96epqIuOB26tSpskOzAZREEqgiLw9hZv2QRHNQFXl5CDPrh1KTgKQ1wCeBl4GvR8S9ZcZjZlY1HTcHSdojaVbSsYbyjZJOSJqSdNcip7kO2B8RW4HrO43FzMw6001N4EHgPuDh+QJJI8Au4CZgBjgi6QDZpvM7G16/FfhjYL+krcDnuojFhlxKI4AGlUcNWTMdJ4GIOCxpoqF4AzAVEScBJO0DNkfETuCWxnNI+nXgnvxc+4Hf6TQes5T1dOhoi5nEXnDOmun16KBVwOm6xzN5WSv/C/gXku4Hpps9QdJ2SZOSJufm5noWqJmZldwxHBHHgFsXec5uSWeBTaOjozf0JzIzs2rodRI4A4zVPV6dl5lZn8y+8lLT8it++B19jsRS0OskcARYL2kt2Zf/FuBD3Z40Ig4CB2u12h3dnssGmzuAizP7V7NNy1t1GM8fc6fxcOtmiOhe4EngGkkzkrZFxGvADuAQ8DzwSEQc7zbIYd5UxjODrWytlpnwUhPVoIgoO4a21Wq1mJycLDuMnpLkmcF1XBM430Kjg1o1+7R06RVNi694S/NyyD6fKX1HWHOSjkZErdmxJNYOGuaagJlZmZJYO8h9AsPHf/EPjlZ9BVYNrgmYmVVYEknA+wmYmRUjiSTgmoCZWTHcJ2CFctv/APlui7b/FqOGwIvOVUESScCsqpY8DLTHnv7a003Lr7zkyj5HYkVxEuiTsfFxZl544YJyTwqzhbz0/eariy5kxSW927Tehl8SSUDSJmDTunXryg6lY94u0swGURIdwx4dZGZWjCRqAjb43AFslqYkagJmNljGWyx8ODExUXZotkRJ1ASGoU/AbJh41NDwSKImkEqfQKtlob00tJkNqiRqAqnwCCAzS42TgLXNnb/F+fZflzspzKorieYgMzMrRqk1AUnXAh8HXgK+HBH7y4zHrJJarSkEC64r1IzXGkpPx0lA0h7gFmA2It5VV74R+AQwAnw2Iu5d4DQ3A5+KiCckHQCcBAaAm32sUx41lJ5uagIPAvcBD88XSBoBdgE3ATPAkfzLfQTY2fD6rcDngHsk/RywootYzMysAx0ngYg4LGmioXgDMBURJwEk7QM2R8ROslpDMx/Nk8ejzQ5K2g5sB1jjYZZmZj3V647hVcDpusczeVlTkiYk7SarTfzHZs+JiN0RUYuI2uWXX97TYBczMTHhWZFmNtRK7RiOiGnyv/IXUtaM4VOnTjUd93/V8mV9jcOGxyAOBV1oueoVl/YxECtFr2sCZ4Cxuser8zIzMxtAvU4CR4D1ktZKGgW2AAe6PWkqy0aYWXNja8bctDqguhkiuhe4EVgpaQa4JyIekLQDOEQ2ImhPRBzvNkgvIFcMDwW1fpl8frJpuYeOlq+b0UG3tSh/DHis44jMzKxvklg2ws1BZmbF8AJyZgUYxFFAHWm1pMQSl5OwwZVETUDSJkm7z507V3YoZmZDJYmaQEQcBA7WarU7yo4lRe4ArpaFxv0PmlYLzs0f86JzxUsiCXh0kNlwarXgHHjkUL8k0RzkjmEzs2IkURMwG1RD0wFslZVETcAdw2ZmxUgiCbg5yGzAfHe2+c2Sk0QSMDOzYjgJmJlVmDuGzRbhzt9yeNP6/kgiCXiegFn1eNP6/kiiOcgdw2ZmxUgiCZiZWTGSaA6yxXl9oO657d+qqG81AUlXS3pA0v66skskPSTpM5I+3K9YzMws01YSkLRH0qykYw3lGyWdkDQl6a6FzhERJyNiW0PxLwD7I+IO4OeWFLmZmXWt3ZrAg8DG+gJJI8Au4GbgWuA2SddKuk7SFxturXagWA2czu+/vvTwzaxqvGl9b7XVJxARhyVNNBRvAKYi4iSApH3A5ojYCdzS5u+fIUsEz+BOausTt/2nzZvW91Y3X7yreOOveMi+0Fe1erKkFZLuB66XdHde/Cjwi5I+DRxs8brtkiYlTc7NzXURrpmZNerb6KCIeAm4s6Hs+8CvLPK63ZLOAptGR0dvKDBEGzL+i99scd3UBM4AY3WPV+dlPefJYmaJ8OqiyekmCRwB1ktaK2kU2AIc6E1Y5/N+AmZmxWh3iOhe4EngGkkzkrZFxGvADuAQ8DzwSEQcLy5UM7PWPGqoM4qIsmNoW61Wi8nJ5iMDiiCJb5175YLyq5Yvo9m/W6vnd6LVDOAfvPWqJT2/Ctz2X5wVl1zUmxNd2mqUePGuvOTKpv9fq0TS0YioNTuWxLBMNweZmRUjibWDIuIgcLBWq91Rdixlq/Jf/GbWe64JmJlVWBJJwENEzcyKkURzkJk7f82KkUQSGObtJd3Gb2ZlcnOQmQ01zx9YWBI1ATOzTnnV0YUlURPw6CAzs2IkURPwPIHqcAewWX8lURMwM7NiOAmYmVVYEs1BNnzc7GM2GJKoCbhj2MysGEkkAc8TMDMrRhJJwMys11pNIqvaRDL3CVih3PZvg6rVJDKo1kSyviUBSVcDHwOWR8StrcoGTbO1fcbXrEbSBeVjq1d5LSCzpehkE/oSdykbRu3uMbxH0qykYw3lGyWdkDQl6a6FzhERJyNi22JlKZh+7mni3J9fcDv61B+VHZqZ2ZK0WxN4ELgPeHi+QNIIsAu4CZgBjkg6AIwAOxtevzUiOkj5lgo3+5ilqa0kEBGHJU00FG8ApiLiJICkfcDmiNgJ3NLTKM3MrBDdjA5aBZyuezyTlzUlaYWk+4HrJd3dqqzJ67ZLmpQ0OTc310W4ZmbWqG8dwxHxEnDnYmVNXrdb0llg0+jo6A0FhmhmVjnd1ATOAGN1j1fnZT3nyWJmZsXoJgkcAdZLWitpFNgCHOhNWOfzshFmZsVoqzlI0l7gRmClpBngnoh4QNIO4BDZiKA9EXG8sEitbzzSx6w62h0ddFuL8seAx3oaUfPf401lzKxv5peUaDQ+Ps709HT/AypQEstGSNoEbFq3bl3ZoZhZBVRpX+IkFpBzx7CZWTGSSALuGDYzK0YSScA1ATOzYiSRBMzMBkGrPQhS3n/AHcNmZm0axg7jJGoCbg4yMytGEknAzMyK4eagCmg1A3jlxSv6HImZDZokagJuDjIzK0YSScDMzIrhJGBmVmFOAmZmBVkzvqbpvIJBmlvgjmEzs4KcfuE0L37/xabHBmVuQRI1AXcMm5kVI4kkYGZmxXASMDOrsL4mAUlXS3pA0v66sg9K+oykz0t6Xz/jMTOrurY7hiXtAW4BZiPiXXXlG4FPkO0z/NmIuLfVOSLiJLCtPglExBeAL0i6DPhPwB8s9SKG0VJn+XayL7D3EjbrjVbbUY6tGSshmqVZyuigB4H7gIfnCySNALuAm4AZ4IikA2QJYWfD67dGxOwC5/+N/FxmZklptbpoCtpOAhFxWNJEQ/EGYCr/Cx9J+4DNEbGTrNawKGXp817g9yPiK+3GY2Zm3eu2T2AVcLru8Uxe1pSkFZLuB66XdHde/M+BfwLcKunOJq/ZLmlS0uTc3FyX4ZqZWb2+ThaLiJeAOxvKPgl8coHX7JZ0Ftg0Ojp6Q8EhmplVSrc1gTNAfc/H6rzMzMwS0G0SOAKsl7RW0iiwBTjQfVjn84xhM7NitJ0EJO0FngSukTQjaVtEvAbsAA4BzwOPRMTxXgcpaZOk3efOnev1qc3MKm0po4Nua1H+GPBYzyJq/jsOAgdrtdodRf4eM7OqSWLZCNcEzMyKkUQScJ+AmVkxkkgCrgmYmRUjiSTgmoCZWTG8s1jJlrqImxd9MxsOrRada2V8fJzp6emex5FEEvDoIDMbNktddK6o7SiTaA4yM7NiOAmYmVVYEknAo4PMzIqRRBLw6CAzs2IkkQTMzKwYTgJmZhXmJGBmVmFJzBOYnywGfEfSN7o41Urg20t5wRWr1nbx6wqz5OsYYL6WwTMs1wFDdi2SOr2W8VYHFBEdnjM9kiYjolZ2HN0alusAX8sgGpbrAF9LO9wcZGZWYU4CZmYVVrUksLvsAHpkWK4DfC2DaFiuA3wti6pUn4CZmZ2vajUBMzOr4yRgZlZhlUgCkjZKOiFpStJdZcfTDknTkp6T9Iykybzs7ZK+JOkb+c/L8nJJ+mR+fc9KenfJse+RNCvpWF3ZkmOXdHv+/G9Iun1AruPjks7k78szkj5Qd+zu/DpOSHp/XXnpnz9JY5Iel/Q1Sccl/cu8PKn3ZYHrSO59kXSxpKclfTW/ln+bl6+V9FQe1+cljeblF+WPp/LjE4tdY1siYqhvwAjwTeBqYBT4KnBt2XG1Efc0sLKh7D8Ad+X37wL+fX7/A8DvAwJ+Cniq5Nh/Gng3cKzT2IG3Ayfzn5fl9y8bgOv4OPDrTZ57bf7ZughYm3/mRgbl8we8A3h3fv9S4Ot5zEm9LwtcR3LvS/5v+9b8/puBp/J/60eALXn5/cBH8vv/DLg/v78F+PxC19huHFWoCWwApiLiZES8CuwDNpccU6c2Aw/l9x8CPlhX/nBk/hh4m6R3lBAfABFxGHi5oXipsb8f+FJEvBwRfwF8CdhYePB1WlxHK5uBfRHxNxHxZ8AU2WdvID5/EXE2Ir6S3/8u8DywisTelwWuo5WBfV/yf9vv5Q/fnN8C+Blgf17e+J7Mv1f7gfdKEq2vsS1VSAKrgNN1j2dY+EMzKAL4A0lHJW3Py66MiLP5/W8B8/vNpXCNS419kK9pR95Esme++YSEriNvRrie7C/PZN+XhuuABN8XSSOSngFmyRLqN4G/jIjXmsT1tzHnx88BK+jyWqqQBFL1noh4N3Az8FFJP11/MLJ6YJLje1OOHfg08HeAnwDOAv+51GiWSNJbgf8B/GpEfKf+WErvS5PrSPJ9iYjXI+IngNVkf73/3X7HUIUkcAYYq3u8Oi8baBFxJv85C/xPsg/Ii/PNPPnP2fzpKVzjUmMfyGuKiBfz/7g/AD7DG9Xugb8OSW8m++L8bxHxaF6c3PvS7DpSfl8AIuIvgceBv0/W9Da/uGd9XH8bc358OfASXV5LFZLAEWB93uM+StahcqDkmBYk6RJJl87fB94HHCOLe340xu3A7+X3DwC/nI/o+CngXF0Vf1AsNfZDwPskXZZX7d+Xl5Wqoa/l58neF8iuY0s+gmMtsB54mgH5/OVtxw8Az0fEb9UdSup9aXUdKb4vki6X9Lb8/jLgJrI+jseBW/OnNb4n8+/VrcAf5rW3VtfYnn72hpd1Ixvp8HWy9raPlR1PG/FeTdbb/1Xg+HzMZO1/Xwa+Afxv4O3xxiiDXfn1PQfUSo5/L1mV/P+RtU9u6yR2YCtZJ9cU8CsDch2fy+N8Nv/P9466538sv44TwM2D9PkD3kPW1PMs8Ex++0Bq78sC15Hc+wL8GPAneczHgH+Tl19N9iU+BfwucFFefnH+eCo/fvVi19jOzctGmJlVWBWag8zMrAUnATOzCnMSMDOrMCcBM7MKcxIwM6swJwEzswpzEjBbIkkTkl7J13yZL7tS0n+XdDJf7+lJST+/wDkeb1zyV9KvSvq0pGX5csivSlpZ4KWYOQmYdeibka35Mj+L9QvA4Yi4OiJuIJuBunqB1+/Nn1NvC7A3Il7Jz/3nvQ7arJGTgFWOpN+VdJ+kP5J0StJ7JH1O0tclPdDBKX8GeDUi7p8viIhTEfGp/Pf903zzkGck/bakEbKlgH+2bsOQCeCdwBNdX6DZEjgJWBVdB5yMiPcAv022Fs2/Ituc42clXbTE8/0o8JVmByT9CPBLwD/M/7p/HfhwRLxMNvX/5vypW4BHwlP4rc+cBKxSJF0MvA34r3lRAA9EtlnJa2Rf0q9K+s38+b/Zwe/YpWzLwCPAe4EbgCN5H8J7ydaGgfObhLbkj8366ocWf4rZUPlR4CuRLTkM8ONka9EjaTVZO/zlwEWSriDbsm8xx4FfnH8QER/NO3QnyRZieygi7m7yut8D/ouy/XvfEhFHO7wms465JmBVcx3Z6qzzfoxsFUfIEsKzwE+SfYHP/1zMHwIXS/pIXdlb8p9fBm7NE8r8xu7jAJFtLfg4sAfXAqwkTgJWNdeRLT883zS0LLK9cuGNhLCkJJC3438Q+MeS/kzS02R7wf7riPga8BtkW4U+S7aFYP3a93vJko+TgJXCS0mbNZD0UETcPv+zyfEJ4IsR8a6C45gmW8f/20X+Hqs21wTM6kj6KDA1/7PF014HltdPFutxDMvyc78Z+MEiTzfrimsCZmYV5pqAmVmFOQmYmVWYk4CZWYU5CZiZVZiTgJlZhTkJmJlVmJOAmVmFOQmYmVWYk4CZWYX9f04LYCWY3uplAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(delphes_v2[:,1],weights=w10*np.ones(len(delphes_v2)),alpha=0.1,bins=np.linspace(0,3000,50))\n",
    "plt.hist(delphes14_v2[:,1],weights=w14*np.ones(len(delphes14_v2)),alpha=0.1,bins=np.linspace(0,3000,50))\n",
    "plt.hist(delphes18_v2[:,1],weights=w18*np.ones(len(delphes18_v2)),alpha=0.1,bins=np.linspace(0,3000,50))\n",
    "plt.hist(delphes_v2_all[:,1],weights=delphes_weights,histtype=\"step\",color=\"black\",bins=np.linspace(0,3000,50))\n",
    "plt.yscale(\"log\")\n",
    "plt.xlabel(\"$m_{jj}$ [GeV]\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.00000e+00  9.60283e+02  4.05912e-01 -2.75399e+00  4.55495e+01\n",
      "  1.00000e+00  9.87996e+02  4.06273e-01 -2.75291e+00  4.55495e+01\n",
      "  1.00000e+00  6.65562e+02 -2.74401e-01  4.01456e-01  2.22392e+01\n",
      "  0.00000e+00  1.01406e+03 -2.79573e-01  3.86054e-01  2.22392e+01]\n"
     ]
    }
   ],
   "source": [
    "#mjj\n",
    "\n",
    "mjjs_delphes = []\n",
    "mjjsg_delphes = []\n",
    "weights_delphes = []\n",
    "pTs_delphes = []\n",
    "pTsg_delphes = []\n",
    "etas_delphes = []\n",
    "etasg_delphes = []\n",
    "\n",
    "print(delphes_v2[0])\n",
    "\n",
    "for i in range(len(delphes_v2)):\n",
    "    \n",
    "    #Truth\n",
    "    pT1g = delphes_v2[i,1]\n",
    "    pT2g = delphes_v2[i,11]\n",
    "    \n",
    "    y1g = delphes_v2[i,2]\n",
    "    y2g = delphes_v2[i,12]\n",
    "    \n",
    "    phi1g = delphes_v2[i,3]\n",
    "    phi2g = delphes_v2[i,13]\n",
    "    \n",
    "    m1g = delphes_v2[i,4]\n",
    "    m2g = delphes_v2[i,14]\n",
    "    \n",
    "    ET1g = np.sqrt(pT1g**2+m1g**2)\n",
    "    ET2g = np.sqrt(pT2g**2+m2g**2)\n",
    "    \n",
    "    mJJg = m1g**2+m2g**2+2*(ET1g*ET2g*np.cosh(y1g-y2g)-pT1g*pT2g*np.cos(phi1g-phi2g))\n",
    "    mjjsg_delphes+=[mJJg**0.5]\n",
    "    mjjsg_delphes+=[mJJg**0.5]\n",
    "    pTsg_delphes+=[pT1g]\n",
    "    pTsg_delphes+=[pT2g]\n",
    "    etasg_delphes+=[y1g]\n",
    "    etasg_delphes+=[y2g]\n",
    "    \n",
    "    #Reco\n",
    "    pT1 = delphes_v2[i,6]\n",
    "    pT2 = delphes_v2[i,16]\n",
    "    \n",
    "    y1 = delphes_v2[i,7]\n",
    "    y2 = delphes_v2[i,17]\n",
    "    \n",
    "    phi1 = delphes_v2[i,8]\n",
    "    phi2 = delphes_v2[i,18]\n",
    "    \n",
    "    m1 = delphes_v2[i,9]\n",
    "    m2 = delphes_v2[i,19]\n",
    "    \n",
    "    ET1 = np.sqrt(pT1**2+m1**2)\n",
    "    ET2 = np.sqrt(pT2**2+m2**2)\n",
    "    \n",
    "    mJJ = m1**2+m2**2+2*(ET1*ET2*np.cosh(y1-y2)-pT1*pT2*np.cos(phi1-phi2))\n",
    "    mjjs_delphes+=[mJJ**0.5]\n",
    "    mjjs_delphes+=[mJJ**0.5]\n",
    "    pTs_delphes+=[pT1g]\n",
    "    pTs_delphes+=[pT2g]\n",
    "    etas_delphes+=[y1]\n",
    "    etas_delphes+=[y2]\n",
    "    \n",
    "    weights_delphes+=[delphes_weights[i]]\n",
    "    weights_delphes+=[delphes_weights[i]]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Read in the Geant4 Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "sim = ef.mod.load(dataset='sim',subdatasets=[\"SIM1000_Jet300_pT375-infGeV\"])\n",
    "gen = ef.mod.load(dataset='gen',subdatasets=[\"GEN1000_pT375-infGeV\"])\n",
    "\n",
    "sim14 = ef.mod.load(dataset='sim',subdatasets=[\"SIM1400_Jet300_pT375-infGeV\"])\n",
    "sim18 = ef.mod.load(dataset='sim',subdatasets=[\"SIM1800_Jet300_pT375-infGeV\"])\n",
    "\n",
    "gen14 = ef.mod.load(dataset='gen',subdatasets=[\"GEN1400_pT375-infGeV\"])\n",
    "gen18 = ef.mod.load(dataset='gen',subdatasets=[\"GEN1800_pT375-infGeV\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "simjets_f = sim.jets_f[0:20000]\n",
    "sim14jets_f = sim14.jets_f[0:20000]\n",
    "sim18jets_f = sim18.jets_f[0:20000]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "#mjj\n",
    "\n",
    "simjets = np.concatenate([sim.jets_f,sim14.jets_f,sim18.jets_f])\n",
    "\n",
    "mjjs = []\n",
    "mjjsg = []\n",
    "weights = []\n",
    "pTs = []\n",
    "pTsg = []\n",
    "etas = []\n",
    "etasg = []\n",
    "for i in range(int(len(simjets)/2.)):\n",
    "    \n",
    "    #Truth\n",
    "    pT1g = simjets[2*i,8]\n",
    "    pT2g = simjets[2*i+1,8]\n",
    "    \n",
    "    y1g = simjets[2*i,9]\n",
    "    y2g = simjets[2*i+1,9]\n",
    "    \n",
    "    phi1g = simjets[2*i,10]\n",
    "    phi2g = simjets[2*i+1,10]\n",
    "    \n",
    "    m1g = simjets[2*i,11]\n",
    "    m2g = simjets[2*i+1,11]\n",
    "    \n",
    "    ET1g = np.sqrt(pT1g**2+m1g**2)\n",
    "    ET2g = np.sqrt(pT2g**2+m2g**2)\n",
    "    \n",
    "    mJJg = m1g**2+m2g**2+2*(ET1g*ET2g*np.cosh(y1g-y2g)-pT1g*pT2g*np.cos(phi1g-phi2g))\n",
    "    mjjsg+=[mJJg**0.5]\n",
    "    mjjsg+=[mJJg**0.5]\n",
    "    pTsg+=[pT1g]\n",
    "    pTsg+=[pT2g]\n",
    "    etasg+=[y1g]\n",
    "    etasg+=[y2g]\n",
    "    \n",
    "    #Reco\n",
    "    pT1 = simjets[2*i,0]\n",
    "    pT2 = simjets[2*i+1,0]\n",
    "    \n",
    "    y1 = simjets[2*i,1]\n",
    "    y2 = simjets[2*i+1,1]\n",
    "    \n",
    "    phi1 = simjets[2*i,2]\n",
    "    phi2 = simjets[2*i+1,2]\n",
    "    \n",
    "    m1 = simjets[2*i,3]\n",
    "    m2 = simjets[2*i+1,3]\n",
    "    \n",
    "    ET1 = np.sqrt(pT1**2+m1**2)\n",
    "    ET2 = np.sqrt(pT2**2+m2**2)\n",
    "    \n",
    "    mJJ = m1**2+m2**2+2*(ET1*ET2*np.cosh(y1-y2)-pT1*pT2*np.cos(phi1-phi2))\n",
    "    mjjs+=[mJJ**0.5]\n",
    "    mjjs+=[mJJ**0.5]\n",
    "    pTs+=[pT1g]\n",
    "    pTs+=[pT2g]\n",
    "    etas+=[y1]\n",
    "    etas+=[y2]\n",
    "    \n",
    "    weights+=[simjets[2*i,-1]]\n",
    "    weights+=[simjets[2*i,-1]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Divide in half\n",
    "\n",
    "inds1_d = np.random.choice(range(len(mjjs_delphes)), size=(int(0.5*len(mjjs_delphes)),), replace=False)\n",
    "inds2_d = list(set(range(len(inds1_d))) - set(inds1_d))\n",
    "\n",
    "mjjs_delphes = np.array(mjjs_delphes)\n",
    "mjjsg_delphes = np.array(mjjsg_delphes)\n",
    "weights_delphes = np.array(weights_delphes)\n",
    "mjjs_delphes_firsthalf = mjjs_delphes[inds1_d]\n",
    "mjjs_delphes_secondhalf = mjjs_delphes[inds2_d]\n",
    "mjjsg_delphes_firsthalf = mjjsg_delphes[inds1_d]\n",
    "mjjsg_delphes_secondhalf = mjjsg_delphes[inds2_d]\n",
    "weights_delphes_firsthalf = weights_delphes[inds1_d]\n",
    "weights_delphes_secondhalf = weights_delphes[inds2_d]\n",
    "\n",
    "inds1 = np.random.choice(range(len(mjjs)), size=(int(0.5*len(mjjs)),), replace=False)\n",
    "inds2 = list(set(range(len(inds1))) - set(inds1))\n",
    "\n",
    "mjjs = np.array(mjjs)\n",
    "mjjsg = np.array(mjjsg)\n",
    "weights = np.array(weights)\n",
    "mjjs_firsthalf = mjjs[inds1]\n",
    "mjjs_secondhalf = mjjs[inds2]\n",
    "mjjsg_firsthalf = mjjsg[inds1]\n",
    "mjjsg_secondhalf = mjjsg[inds2]\n",
    "weights_firsthalf = weights[inds1]\n",
    "weights_secondhalf = weights[inds2]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Simulation-based Calibration"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "inputs_prior = Input((1, ))\n",
    "hidden_layer_1_prior = Dense(50, activation='relu')(inputs_prior)\n",
    "hidden_layer_2_prior = Dense(50, activation='relu')(hidden_layer_1_prior)\n",
    "hidden_layer_3_prior = Dense(50, activation='relu')(hidden_layer_2_prior)\n",
    "outputs_prior = Dense(1, activation='linear')(hidden_layer_3_prior)\n",
    "model_prior = Model(inputs=inputs_prior, outputs=outputs_prior)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "150/150 [==============================] - 0s 3ms/step - loss: 1.2286e-08 - accuracy: 0.0000e+00\n",
      "Epoch 2/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 1.1920e-08 - accuracy: 0.0000e+00\n",
      "Epoch 3/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 1.1481e-08 - accuracy: 0.0000e+00\n",
      "Epoch 4/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 1.1007e-08 - accuracy: 0.0000e+00\n",
      "Epoch 5/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 1.0497e-08 - accuracy: 0.0000e+00\n",
      "Epoch 6/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 9.9950e-09 - accuracy: 0.0000e+00\n",
      "Epoch 7/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 9.4392e-09 - accuracy: 0.0000e+00\n",
      "Epoch 8/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 8.8621e-09 - accuracy: 0.0000e+00\n",
      "Epoch 9/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 8.3034e-09 - accuracy: 0.0000e+00\n",
      "Epoch 10/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 7.7698e-09 - accuracy: 0.0000e+00\n",
      "Epoch 11/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 7.2712e-09 - accuracy: 0.0000e+00\n",
      "Epoch 12/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 6.7992e-09 - accuracy: 0.0000e+00\n",
      "Epoch 13/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 6.3434e-09 - accuracy: 0.0000e+00\n",
      "Epoch 14/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 5.9031e-09 - accuracy: 0.0000e+00\n",
      "Epoch 15/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 5.4708e-09 - accuracy: 0.0000e+00\n",
      "Epoch 16/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 5.0424e-09 - accuracy: 0.0000e+00\n",
      "Epoch 17/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 4.6370e-09 - accuracy: 0.0000e+00\n",
      "Epoch 18/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 4.2501e-09 - accuracy: 0.0000e+00\n",
      "Epoch 19/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 3.8802e-09 - accuracy: 0.0000e+00\n",
      "Epoch 20/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 3.5250e-09 - accuracy: 0.0000e+00\n",
      "Epoch 21/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 3.1856e-09 - accuracy: 0.0000e+00\n",
      "Epoch 22/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 2.8661e-09 - accuracy: 0.0000e+00\n",
      "Epoch 23/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 2.5662e-09 - accuracy: 0.0000e+00\n",
      "Epoch 24/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 2.2859e-09 - accuracy: 0.0000e+00\n",
      "Epoch 25/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 2.0259e-09 - accuracy: 0.0000e+00\n",
      "Epoch 26/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 1.7861e-09 - accuracy: 0.0000e+00\n",
      "Epoch 27/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 1.5671e-09 - accuracy: 0.0000e+00\n",
      "Epoch 28/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 1.3685e-09 - accuracy: 0.0000e+00\n",
      "Epoch 29/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 1.1900e-09 - accuracy: 0.0000e+00\n",
      "Epoch 30/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 1.0310e-09 - accuracy: 0.0000e+00\n",
      "Epoch 31/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 8.9057e-10 - accuracy: 0.0000e+00\n",
      "Epoch 32/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 7.6771e-10 - accuracy: 0.0000e+00\n",
      "Epoch 33/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 6.6112e-10 - accuracy: 0.0000e+00\n",
      "Epoch 34/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 5.6939e-10 - accuracy: 0.0000e+00\n",
      "Epoch 35/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 4.9113e-10 - accuracy: 0.0000e+00\n",
      "Epoch 36/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 4.2487e-10 - accuracy: 0.0000e+00\n",
      "Epoch 37/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 3.6916e-10 - accuracy: 0.0000e+00\n",
      "Epoch 38/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 3.2269e-10 - accuracy: 0.0000e+00\n",
      "Epoch 39/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 2.8413e-10 - accuracy: 0.0000e+00\n",
      "Epoch 40/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 2.5236e-10 - accuracy: 0.0000e+00\n",
      "Epoch 41/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 2.2631e-10 - accuracy: 0.0000e+00\n",
      "Epoch 42/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 2.0505e-10 - accuracy: 0.0000e+00\n",
      "Epoch 43/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 1.8776e-10 - accuracy: 0.0000e+00\n",
      "Epoch 44/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 1.7376e-10 - accuracy: 0.0000e+00\n",
      "Epoch 45/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 1.6245e-10 - accuracy: 0.0000e+00\n",
      "Epoch 46/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 1.5332e-10 - accuracy: 0.0000e+00\n",
      "Epoch 47/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 1.4596e-10 - accuracy: 0.0000e+00\n",
      "Epoch 48/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 1.4003e-10 - accuracy: 0.0000e+00\n",
      "Epoch 49/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 1.3524e-10 - accuracy: 0.0000e+00\n",
      "Epoch 50/50\n",
      "150/150 [==============================] - 0s 2ms/step - loss: 1.3136e-10 - accuracy: 0.0000e+00\n"
     ]
    }
   ],
   "source": [
    "model_prior.compile(loss=\"mse\",optimizer='Adam',metrics=['accuracy'])\n",
    "h = model_prior.fit(mjjs_firsthalf/1000.,\n",
    "          mjjsg_firsthalf/1000.,\n",
    "          epochs=50,\n",
    "          sample_weight=weights_firsthalf,\n",
    "          batch_size=1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "pres_mass = model_prior.predict(mjjs_secondhalf/1000.,batch_size=1000).flatten()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "_,_,_=plt.hist(mjjs_secondhalf/mjjsg_secondhalf,weights=weights_secondhalf,bins=np.linspace(0.5,1.5,50),alpha=0.2,density=True,color=\"blue\",label=\"QCD, Delphes\")\n",
    "_,_,_=plt.hist(1000*pres_mass/mjjsg_secondhalf,weights=weights_secondhalf,bins=np.linspace(0.5,1.5,50),alpha=0.2,density=True,color=\"black\",histtype=\"step\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "mybins = np.logspace(np.log(2000)/np.log(10),np.log(5000)/np.log(10),50)\n",
    "mybinsB = np.linspace(0.8,1.4,50)\n",
    "\n",
    "resmass = 2800\n",
    "\n",
    "fig = plt.figure(figsize=(6, 6))\n",
    "gs = gridspec.GridSpec(1, 1) \n",
    "plt.subplots_adjust(hspace = .001)\n",
    "\n",
    "ax0 = plt.subplot(gs[0])\n",
    "ax0.yaxis.set_ticks_position('both')\n",
    "ax0.xaxis.set_ticks_position('both')\n",
    "ax0.tick_params(direction=\"in\",which=\"both\")\n",
    "plt.xticks(fontsize=20)\n",
    "plt.yticks(fontsize=20)\n",
    "ax0.minorticks_on()\n",
    "\n",
    "_,_,_=plt.hist(mjjs_secondhalf/mjjsg_secondhalf,weights=weights_secondhalf,bins=mybinsB,alpha=0.2,density=True,color=\"blue\",label=\"QCD, Geant4\")\n",
    "_,_,_=plt.hist(1000*pres_mass/mjjsg_secondhalf,weights=weights_secondhalf,bins=mybinsB,density=True,color=\"black\",histtype=\"step\",label=\"QCD, Geant4 + NN\",lw=3)\n",
    "\n",
    "ww,bins = np.histogram(mjjsg_secondhalf,bins=mybins)\n",
    "ww = np.interp(mjjsg_secondhalf,0.5*(bins[:-1]+bins[1:]),ww)\n",
    "\n",
    "_,_,_=plt.hist(mjjs_secondhalf/mjjsg_secondhalf,weights=norm.pdf(mjjsg_secondhalf,resmass,10)/ww,bins=mybinsB,alpha=0.7,density=True,color=\"orange\",label=\"BSM, Geant4\")\n",
    "_,_,_=plt.hist(1000*pres_mass/mjjsg_secondhalf,weights=norm.pdf(mjjsg_secondhalf,resmass,10)/ww,bins=mybinsB,density=True,color=\"black\",histtype=\"step\",label=\"BSM, Geant4 + NN\",lw=3,ls=\":\")\n",
    "\n",
    "plt.ylabel(\"Normalized to Unity\",fontsize=20)\n",
    "#plt.xlim([mymin,mymax])\n",
    "#plt.yscale(\"log\")\n",
    "#plt.ylim([1e-6,1e-1])\n",
    "#plt.locator_params(axis='y', nbins=6)\n",
    "plt.axvline(1,alpha=0.5)\n",
    "plt.xlabel(\"reconstructed / true $m_{jj}$\",fontsize=20)\n",
    "plt.legend(frameon=False,fontsize=15,ncol=1)\n",
    "_=plt.title(\"Pythia 8 Dijets\",loc=\"right\",fontsize=20, fontstyle=\"italic\")\n",
    "plt.savefig(\"NN_prior.pdf\",bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Simulated-based Calibration with GA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Start of epoch 0\n",
      "WARNING:tensorflow:Layer g_ifn is casting an input tensor from dtype float64 to the layer's dtype of float32, which is new behavior in TensorFlow 2.  The layer has dtype float32 because it's dtype defaults to floatx.\n",
      "\n",
      "If you intended to run this layer in float32, you can safely ignore this warning. If in doubt, this warning is likely only an issue if you are porting a TensorFlow 1.X model to TensorFlow 2.\n",
      "\n",
      "To change all layers to have dtype float64 by default, call `tf.keras.backend.set_floatx('float64')`. To change just this layer, pass dtype='float64' to the layer constructor. If you are the author of this layer, you can disable autocasting by passing autocast=False to the base Layer constructor.\n",
      "\n",
      "WARNING:tensorflow:Layer dnn_1 is casting an input tensor from dtype float64 to the layer's dtype of float32, which is new behavior in TensorFlow 2.  The layer has dtype float32 because it's dtype defaults to floatx.\n",
      "\n",
      "If you intended to run this layer in float32, you can safely ignore this warning. If in doubt, this warning is likely only an issue if you are porting a TensorFlow 1.X model to TensorFlow 2.\n",
      "\n",
      "To change all layers to have dtype float64 by default, call `tf.keras.backend.set_floatx('float64')`. To change just this layer, pass dtype='float64' to the layer constructor. If you are the author of this layer, you can disable autocasting by passing autocast=False to the base Layer constructor.\n",
      "\n",
      "WARNING:tensorflow:Layer dnn is casting an input tensor from dtype float64 to the layer's dtype of float32, which is new behavior in TensorFlow 2.  The layer has dtype float32 because it's dtype defaults to floatx.\n",
      "\n",
      "If you intended to run this layer in float32, you can safely ignore this warning. If in doubt, this warning is likely only an issue if you are porting a TensorFlow 1.X model to TensorFlow 2.\n",
      "\n",
      "To change all layers to have dtype float64 by default, call `tf.keras.backend.set_floatx('float64')`. To change just this layer, pass dtype='float64' to the layer constructor. If you are the author of this layer, you can disable autocasting by passing autocast=False to the base Layer constructor.\n",
      "\n",
      "WARNING:tensorflow:Layer dnn_2 is casting an input tensor from dtype float64 to the layer's dtype of float32, which is new behavior in TensorFlow 2.  The layer has dtype float32 because it's dtype defaults to floatx.\n",
      "\n",
      "If you intended to run this layer in float32, you can safely ignore this warning. If in doubt, this warning is likely only an issue if you are porting a TensorFlow 1.X model to TensorFlow 2.\n",
      "\n",
      "To change all layers to have dtype float64 by default, call `tf.keras.backend.set_floatx('float64')`. To change just this layer, pass dtype='float64' to the layer constructor. If you are the author of this layer, you can disable autocasting by passing autocast=False to the base Layer constructor.\n",
      "\n",
      "Pre-Train Epoch 0, Batch 0: Loss = 23.201, a = 7.662 b = 0.700, c = -2.515, Joint = -2.665, Marginal = -3.874, MI = -18.909\n",
      "Pre-Train Epoch 0, Batch 1: Loss = 14.114, a = 5.771 b = 0.076, c = -3.252, Joint = -2.094, Marginal = -3.298, MI = -13.712\n",
      "Pre-Train Epoch 0, Batch 2: Loss = 11.734, a = 4.282 b = 0.035, c = -1.146, Joint = -2.206, Marginal = -3.120, MI = -22.495\n",
      "Pre-Train Epoch 0, Batch 3: Loss = 7.745, a = 3.066 b = 0.032, c = -1.637, Joint = -1.789, Marginal = -2.255, MI = -20.115\n",
      "Pre-Train Epoch 0, Batch 4: Loss = 4.099, a = 2.034 b = 0.050, c = -1.502, Joint = -1.123, Marginal = -1.080, MI = -33.856\n",
      "Pre-Train Epoch 0, Batch 5: Loss = 2.165, a = 1.247 b = 0.150, c = -1.957, Joint = -0.412, Marginal = 0.214, MI = -27.604\n",
      "Pre-Train Epoch 0, Batch 6: Loss = 2.203, a = 0.722 b = 0.335, c = -1.320, Joint = 0.340, Marginal = 1.615, MI = -35.942\n",
      "Pre-Train Epoch 0, Batch 7: Loss = 2.820, a = 0.383 b = 0.405, c = -0.396, Joint = 0.776, Marginal = 2.124, MI = -25.586\n",
      "Pre-Train Epoch 0, Batch 8: Loss = 3.612, a = 0.191 b = 0.422, c = -0.280, Joint = 1.057, Marginal = 2.689, MI = -35.797\n",
      "Pre-Train Epoch 0, Batch 9: Loss = 3.808, a = 0.118 b = 0.352, c = -0.014, Joint = 1.077, Marginal = 2.882, MI = -42.110\n",
      "Pre-Train Epoch 0, Batch 10: Loss = 2.314, a = 0.090 b = 0.166, c = 0.012, Joint = 0.697, Marginal = 1.981, MI = -31.760\n",
      "Pre-Train Epoch 0, Batch 11: Loss = 1.404, a = 0.085 b = 0.053, c = -0.356, Joint = 0.316, Marginal = 1.240, MI = -20.500\n",
      "Pre-Train Epoch 0, Batch 12: Loss = 0.957, a = 0.089 b = 0.024, c = -0.346, Joint = -0.145, Marginal = 0.608, MI = -19.495\n",
      "Pre-Train Epoch 0, Batch 13: Loss = 1.147, a = 0.097 b = 0.086, c = 0.201, Joint = -0.555, Marginal = -0.095, MI = -13.275\n",
      "Pre-Train Epoch 0, Batch 14: Loss = 1.520, a = 0.123 b = 0.181, c = -0.168, Joint = -0.839, Marginal = -0.502, MI = -28.924\n",
      "Pre-Train Epoch 0, Batch 15: Loss = 1.815, a = 0.140 b = 0.253, c = -0.883, Joint = -0.978, Marginal = -0.775, MI = -13.867\n",
      "Pre-Train Epoch 0, Batch 16: Loss = 1.605, a = 0.178 b = 0.247, c = -0.406, Joint = -0.889, Marginal = -0.782, MI = -12.063\n",
      "Pre-Train Epoch 0, Batch 17: Loss = 1.294, a = 0.215 b = 0.209, c = -0.494, Joint = -0.710, Marginal = -0.671, MI = -8.977\n",
      "Pre-Train Epoch 0, Batch 18: Loss = 0.959, a = 0.280 b = 0.144, c = -1.352, Joint = -0.425, Marginal = -0.510, MI = -8.002\n",
      "Pre-Train Epoch 0, Batch 19: Loss = 0.818, a = 0.351 b = 0.082, c = -0.095, Joint = -0.096, Marginal = -0.264, MI = -7.184\n",
      "Pre-Train Epoch 0, Batch 20: Loss = 0.858, a = 0.397 b = 0.038, c = -1.129, Joint = 0.177, Marginal = -0.078, MI = -6.692\n",
      "Pre-Train Epoch 0, Batch 21: Loss = 1.057, a = 0.399 b = 0.033, c = -1.275, Joint = 0.303, Marginal = -0.018, MI = -5.689\n",
      "Pre-Train Epoch 0, Batch 22: Loss = 1.184, a = 0.417 b = 0.025, c = -1.012, Joint = 0.449, Marginal = 0.076, MI = -5.936\n",
      "Pre-Train Epoch 0, Batch 23: Loss = 1.246, a = 0.423 b = 0.024, c = -1.118, Joint = 0.496, Marginal = 0.046, MI = -3.529\n",
      "Pre-Train Epoch 0, Batch 24: Loss = 1.125, a = 0.365 b = 0.030, c = -2.063, Joint = 0.434, Marginal = -0.003, MI = -8.045\n",
      "Pre-Train Epoch 0, Batch 25: Loss = 0.943, a = 0.311 b = 0.023, c = -1.577, Joint = 0.376, Marginal = -0.071, MI = -2.829\n",
      "Pre-Train Epoch 0, Batch 26: Loss = 0.891, a = 0.300 b = 0.027, c = -0.616, Joint = 0.300, Marginal = -0.224, MI = -4.258\n",
      "Pre-Train Epoch 0, Batch 27: Loss = 0.729, a = 0.243 b = 0.030, c = -0.283, Joint = 0.190, Marginal = -0.303, MI = -2.662\n",
      "Pre-Train Epoch 0, Batch 28: Loss = 0.572, a = 0.184 b = 0.037, c = -1.244, Joint = 0.039, Marginal = -0.436, MI = -1.831\n",
      "Pre-Train Epoch 0, Batch 29: Loss = 0.504, a = 0.149 b = 0.048, c = -0.125, Joint = -0.068, Marginal = -0.543, MI = -1.832\n",
      "Pre-Train Epoch 0, Batch 30: Loss = 0.505, a = 0.119 b = 0.067, c = -1.456, Joint = -0.176, Marginal = -0.612, MI = -2.046\n",
      "Pre-Train Epoch 0, Batch 31: Loss = 0.523, a = 0.091 b = 0.079, c = -0.174, Joint = -0.265, Marginal = -0.635, MI = -1.807\n",
      "Pre-Train Epoch 0, Batch 32: Loss = 0.470, a = 0.074 b = 0.068, c = -0.634, Joint = -0.294, Marginal = -0.670, MI = -1.804\n",
      "Pre-Train Epoch 0, Batch 33: Loss = 0.436, a = 0.065 b = 0.064, c = -0.830, Joint = -0.289, Marginal = -0.659, MI = -1.601\n",
      "Pre-Train Epoch 0, Batch 34: Loss = 0.477, a = 0.063 b = 0.059, c = -1.072, Joint = -0.282, Marginal = -0.648, MI = -1.418\n",
      "Pre-Train Epoch 0, Batch 35: Loss = 0.442, a = 0.060 b = 0.048, c = -0.974, Joint = -0.248, Marginal = -0.611, MI = -1.755\n",
      "Pre-Train Epoch 0, Batch 36: Loss = 0.410, a = 0.063 b = 0.030, c = -0.560, Joint = -0.168, Marginal = -0.584, MI = -1.207\n",
      "Pre-Train Epoch 0, Batch 37: Loss = 0.351, a = 0.060 b = 0.026, c = -0.594, Joint = -0.100, Marginal = -0.481, MI = -1.061\n",
      "Pre-Train Epoch 0, Batch 38: Loss = 0.363, a = 0.059 b = 0.034, c = -0.761, Joint = -0.039, Marginal = -0.452, MI = -0.858\n",
      "Pre-Train Epoch 0, Batch 39: Loss = 0.369, a = 0.064 b = 0.038, c = -0.405, Joint = 0.015, Marginal = -0.502, MI = -0.711\n",
      "Pre-Train Epoch 0, Batch 40: Loss = 0.366, a = 0.069 b = 0.048, c = -1.150, Joint = 0.032, Marginal = -0.491, MI = -0.555\n",
      "Pre-Train Epoch 0, Batch 41: Loss = 0.354, a = 0.066 b = 0.060, c = -1.011, Joint = 0.067, Marginal = -0.537, MI = -0.407\n",
      "Pre-Train Epoch 0, Batch 42: Loss = 0.345, a = 0.065 b = 0.066, c = -0.943, Joint = 0.069, Marginal = -0.619, MI = -0.265\n",
      "Pre-Train Epoch 0, Batch 43: Loss = 0.329, a = 0.068 b = 0.068, c = -0.797, Joint = 0.062, Marginal = -0.656, MI = -0.158\n",
      "Pre-Train Epoch 0, Batch 44: Loss = 0.313, a = 0.068 b = 0.064, c = -0.511, Joint = 0.041, Marginal = -0.651, MI = -0.137\n",
      "Pre-Train Epoch 0, Batch 45: Loss = 0.336, a = 0.073 b = 0.068, c = -1.454, Joint = 0.026, Marginal = -0.814, MI = -0.026\n",
      "Pre-Train Epoch 0, Batch 46: Loss = 0.340, a = 0.072 b = 0.062, c = -1.254, Joint = 0.016, Marginal = -0.936, MI = 0.034\n",
      "Pre-Train Epoch 0, Batch 47: Loss = 0.298, a = 0.069 b = 0.048, c = -0.928, Joint = -0.017, Marginal = -0.880, MI = 0.023\n",
      "Pre-Train Epoch 0, Batch 48: Loss = 0.282, a = 0.059 b = 0.038, c = -0.862, Joint = -0.050, Marginal = -0.898, MI = 0.059\n",
      "Pre-Train Epoch 0, Batch 49: Loss = 0.255, a = 0.058 b = 0.033, c = -0.422, Joint = -0.057, Marginal = -0.912, MI = 0.022\n",
      "Pre-Train Epoch 0, Batch 50: Loss = 0.253, a = 0.054 b = 0.029, c = -0.722, Joint = -0.064, Marginal = -1.033, MI = 0.049\n",
      "Pre-Train Epoch 0, Batch 51: Loss = 0.225, a = 0.047 b = 0.023, c = -0.606, Joint = -0.070, Marginal = -0.959, MI = 0.042\n",
      "Pre-Train Epoch 0, Batch 52: Loss = 0.231, a = 0.042 b = 0.025, c = -1.121, Joint = -0.056, Marginal = -0.897, MI = -0.005\n",
      "Pre-Train Epoch 0, Batch 53: Loss = 0.220, a = 0.040 b = 0.022, c = -1.237, Joint = -0.050, Marginal = -0.957, MI = 0.046\n",
      "Pre-Train Epoch 0, Batch 54: Loss = 0.208, a = 0.038 b = 0.022, c = -0.997, Joint = -0.024, Marginal = -0.980, MI = 0.083\n",
      "Pre-Train Epoch 0, Batch 55: Loss = 0.221, a = 0.037 b = 0.027, c = -1.164, Joint = -0.023, Marginal = -0.796, MI = -0.014\n",
      "Pre-Train Epoch 0, Batch 56: Loss = 0.207, a = 0.041 b = 0.021, c = -0.904, Joint = 0.004, Marginal = -1.054, MI = 0.080\n",
      "Pre-Train Epoch 0, Batch 57: Loss = 0.195, a = 0.039 b = 0.024, c = -0.852, Joint = 0.008, Marginal = -0.881, MI = 0.054\n",
      "Pre-Train Epoch 0, Batch 58: Loss = 0.195, a = 0.041 b = 0.024, c = -0.958, Joint = 0.020, Marginal = -0.987, MI = 0.120\n",
      "Pre-Train Epoch 0, Batch 59: Loss = 0.190, a = 0.040 b = 0.023, c = -1.105, Joint = 0.022, Marginal = -1.157, MI = 0.199\n",
      "Pre-Train Epoch 0, Batch 60: Loss = 0.169, a = 0.039 b = 0.019, c = -1.003, Joint = 0.026, Marginal = -1.016, MI = 0.185\n",
      "Pre-Train Epoch 0, Batch 61: Loss = 0.160, a = 0.036 b = 0.018, c = -1.049, Joint = 0.031, Marginal = -1.097, MI = 0.243\n",
      "Pre-Train Epoch 0, Batch 62: Loss = 0.172, a = 0.033 b = 0.025, c = -0.763, Joint = 0.017, Marginal = -1.214, MI = 0.291\n",
      "Pre-Train Epoch 0, Batch 63: Loss = 0.170, a = 0.033 b = 0.027, c = -0.588, Joint = -0.005, Marginal = -1.199, MI = 0.268\n",
      "Pre-Train Epoch 0, Batch 64: Loss = 0.151, a = 0.032 b = 0.024, c = -0.870, Joint = -0.003, Marginal = -1.487, MI = 0.355\n",
      "Pre-Train Epoch 0, Batch 65: Loss = 0.136, a = 0.030 b = 0.023, c = -1.240, Joint = -0.020, Marginal = -1.538, MI = 0.395\n",
      "Pre-Train Epoch 0, Batch 66: Loss = 0.134, a = 0.031 b = 0.025, c = -0.713, Joint = -0.031, Marginal = -1.427, MI = 0.386\n",
      "Pre-Train Epoch 0, Batch 67: Loss = 0.141, a = 0.034 b = 0.029, c = -0.727, Joint = -0.041, Marginal = -1.571, MI = 0.421\n",
      "Pre-Train Epoch 0, Batch 68: Loss = 0.104, a = 0.030 b = 0.023, c = -0.924, Joint = -0.024, Marginal = -1.614, MI = 0.469\n",
      "Pre-Train Epoch 0, Batch 69: Loss = 0.111, a = 0.033 b = 0.027, c = -1.162, Joint = -0.034, Marginal = -1.613, MI = 0.494\n",
      "Pre-Train Epoch 0, Batch 70: Loss = 0.096, a = 0.030 b = 0.023, c = -1.057, Joint = -0.024, Marginal = -1.808, MI = 0.548\n",
      "Pre-Train Epoch 0, Batch 71: Loss = 0.102, a = 0.029 b = 0.025, c = -1.061, Joint = -0.033, Marginal = -1.794, MI = 0.528\n",
      "Pre-Train Epoch 0, Batch 72: Loss = 0.087, a = 0.028 b = 0.022, c = -0.885, Joint = -0.033, Marginal = -1.751, MI = 0.507\n",
      "Pre-Train Epoch 0, Batch 73: Loss = 0.080, a = 0.025 b = 0.021, c = -0.930, Joint = -0.028, Marginal = -1.672, MI = 0.516\n",
      "Pre-Train Epoch 0, Batch 74: Loss = 0.077, a = 0.024 b = 0.020, c = -0.920, Joint = -0.028, Marginal = -1.725, MI = 0.514\n",
      "Pre-Train Epoch 0, Batch 75: Loss = 0.082, a = 0.023 b = 0.019, c = -0.804, Joint = -0.026, Marginal = -1.882, MI = 0.535\n",
      "Pre-Train Epoch 0, Batch 76: Loss = 0.076, a = 0.021 b = 0.018, c = -0.846, Joint = -0.015, Marginal = -1.816, MI = 0.477\n",
      "Pre-Train Epoch 0, Batch 77: Loss = 0.058, a = 0.019 b = 0.015, c = -0.928, Joint = -0.002, Marginal = -1.848, MI = 0.553\n",
      "Pre-Train Epoch 0, Batch 78: Loss = 0.066, a = 0.017 b = 0.019, c = -1.038, Joint = -0.008, Marginal = -1.996, MI = 0.513\n",
      "Pre-Train Epoch 0, Batch 79: Loss = 0.066, a = 0.016 b = 0.018, c = -0.935, Joint = 0.003, Marginal = -1.831, MI = 0.581\n",
      "Pre-Train Epoch 0, Batch 80: Loss = 0.077, a = 0.014 b = 0.021, c = -1.118, Joint = -0.003, Marginal = -1.898, MI = 0.566\n",
      "Pre-Train Epoch 0, Batch 81: Loss = 0.058, a = 0.012 b = 0.018, c = -1.015, Joint = 0.014, Marginal = -1.947, MI = 0.577\n",
      "Pre-Train Epoch 0, Batch 82: Loss = 0.045, a = 0.010 b = 0.014, c = -1.027, Joint = 0.031, Marginal = -1.810, MI = 0.583\n",
      "Pre-Train Epoch 0, Batch 83: Loss = 0.054, a = 0.008 b = 0.018, c = -0.832, Joint = 0.036, Marginal = -1.884, MI = 0.583\n",
      "Pre-Train Epoch 0, Batch 84: Loss = 0.051, a = 0.008 b = 0.015, c = -0.905, Joint = 0.041, Marginal = -2.047, MI = 0.608\n",
      "Pre-Train Epoch 0, Batch 85: Loss = 0.046, a = 0.006 b = 0.016, c = -0.853, Joint = -0.005, Marginal = -1.741, MI = 0.504\n",
      "Pre-Train Epoch 0, Batch 86: Loss = 0.062, a = 0.003 b = 0.023, c = -0.942, Joint = -0.034, Marginal = -1.810, MI = 0.528\n",
      "Pre-Train Epoch 0, Batch 87: Loss = 0.047, a = 0.003 b = 0.018, c = -1.047, Joint = -0.012, Marginal = -1.916, MI = 0.609\n",
      "Pre-Train Epoch 0, Batch 88: Loss = 0.042, a = 0.002 b = 0.017, c = -1.141, Joint = 0.013, Marginal = -2.007, MI = 0.634\n",
      "Pre-Train Epoch 0, Batch 89: Loss = 0.041, a = 0.002 b = 0.017, c = -1.001, Joint = 0.024, Marginal = -1.751, MI = 0.587\n",
      "Pre-Train Epoch 0, Batch 90: Loss = 0.053, a = 0.002 b = 0.021, c = -1.140, Joint = 0.021, Marginal = -1.629, MI = 0.580\n",
      "Pre-Train Epoch 0, Batch 91: Loss = 0.047, a = 0.002 b = 0.021, c = -1.026, Joint = 0.019, Marginal = -1.815, MI = 0.625\n",
      "Pre-Train Epoch 0, Batch 92: Loss = 0.036, a = 0.001 b = 0.016, c = -0.878, Joint = 0.021, Marginal = -1.976, MI = 0.634\n",
      "Pre-Train Epoch 0, Batch 93: Loss = 0.037, a = 0.001 b = 0.015, c = -0.854, Joint = 0.007, Marginal = -1.670, MI = 0.561\n",
      "Pre-Train Epoch 0, Batch 94: Loss = 0.041, a = 0.001 b = 0.016, c = -0.982, Joint = -0.010, Marginal = -1.772, MI = 0.613\n",
      "Pre-Train Epoch 0, Batch 95: Loss = 0.037, a = 0.000 b = 0.015, c = -0.918, Joint = -0.013, Marginal = -2.020, MI = 0.650\n",
      "Pre-Train Epoch 0, Batch 96: Loss = 0.034, a = 0.001 b = 0.014, c = -1.078, Joint = -0.009, Marginal = -1.665, MI = 0.561\n",
      "Pre-Train Epoch 0, Batch 97: Loss = 0.045, a = 0.001 b = 0.018, c = -0.968, Joint = -0.001, Marginal = -1.756, MI = 0.569\n",
      "Pre-Train Epoch 0, Batch 98: Loss = 0.038, a = 0.001 b = 0.013, c = -1.196, Joint = 0.015, Marginal = -1.439, MI = 0.542\n",
      "Pre-Train Epoch 0, Batch 99: Loss = 0.041, a = 0.001 b = 0.016, c = -1.107, Joint = -0.008, Marginal = -1.547, MI = 0.538\n",
      "Pre-Train Epoch 0, Batch 100: Loss = 0.040, a = 0.000 b = 0.016, c = -0.909, Joint = -0.007, Marginal = -1.761, MI = 0.567\n",
      "Pre-Train Epoch 0, Batch 101: Loss = 0.029, a = 0.001 b = 0.013, c = -0.944, Joint = -0.004, Marginal = -1.415, MI = 0.529\n",
      "Pre-Train Epoch 0, Batch 102: Loss = 0.037, a = 0.001 b = 0.015, c = -0.939, Joint = -0.002, Marginal = -1.742, MI = 0.599\n",
      "Pre-Train Epoch 0, Batch 103: Loss = 0.041, a = 0.001 b = 0.018, c = -1.012, Joint = 0.007, Marginal = -1.492, MI = 0.537\n",
      "Pre-Train Epoch 0, Batch 104: Loss = 0.039, a = 0.000 b = 0.018, c = -0.952, Joint = 0.011, Marginal = -1.693, MI = 0.585\n",
      "Pre-Train Epoch 0, Batch 105: Loss = 0.028, a = 0.000 b = 0.013, c = -1.004, Joint = 0.014, Marginal = -1.454, MI = 0.539\n",
      "Pre-Train Epoch 0, Batch 106: Loss = 0.035, a = 0.000 b = 0.015, c = -1.052, Joint = 0.001, Marginal = -1.541, MI = 0.569\n",
      "Pre-Train Epoch 0, Batch 107: Loss = 0.036, a = 0.000 b = 0.016, c = -0.908, Joint = -0.005, Marginal = -1.702, MI = 0.570\n",
      "Pre-Train Epoch 0, Batch 108: Loss = 0.033, a = 0.000 b = 0.015, c = -1.006, Joint = -0.012, Marginal = -1.651, MI = 0.584\n",
      "Pre-Train Epoch 0, Batch 109: Loss = 0.028, a = 0.000 b = 0.013, c = -0.952, Joint = -0.001, Marginal = -1.504, MI = 0.552\n",
      "Pre-Train Epoch 0, Batch 110: Loss = 0.049, a = 0.000 b = 0.021, c = -0.992, Joint = -0.003, Marginal = -1.567, MI = 0.545\n",
      "Pre-Train Epoch 0, Batch 111: Loss = 0.030, a = 0.000 b = 0.013, c = -1.023, Joint = 0.012, Marginal = -1.573, MI = 0.568\n",
      "Pre-Train Epoch 0, Batch 112: Loss = 0.027, a = 0.000 b = 0.013, c = -0.964, Joint = 0.008, Marginal = -1.351, MI = 0.538\n",
      "Pre-Train Epoch 0, Batch 113: Loss = 0.024, a = 0.000 b = 0.012, c = -0.957, Joint = 0.003, Marginal = -1.444, MI = 0.528\n",
      "Pre-Train Epoch 0, Batch 114: Loss = 0.040, a = 0.001 b = 0.018, c = -0.931, Joint = -0.012, Marginal = -1.498, MI = 0.555\n",
      "Pre-Train Epoch 0, Batch 115: Loss = 0.026, a = 0.000 b = 0.012, c = -0.992, Joint = -0.002, Marginal = -1.517, MI = 0.568\n",
      "Pre-Train Epoch 0, Batch 116: Loss = 0.035, a = 0.000 b = 0.016, c = -0.997, Joint = -0.003, Marginal = -1.642, MI = 0.612\n",
      "Pre-Train Epoch 0, Batch 117: Loss = 0.031, a = 0.000 b = 0.015, c = -0.961, Joint = 0.012, Marginal = -1.514, MI = 0.581\n",
      "Pre-Train Epoch 0, Batch 118: Loss = 0.042, a = 0.000 b = 0.018, c = -1.063, Joint = 0.003, Marginal = -1.554, MI = 0.532\n",
      "Pre-Train Epoch 0, Batch 119: Loss = 0.041, a = 0.000 b = 0.018, c = -0.970, Joint = 0.004, Marginal = -1.366, MI = 0.549\n",
      "Pre-Train Epoch 0, Batch 120: Loss = 0.034, a = 0.000 b = 0.016, c = -0.983, Joint = -0.005, Marginal = -1.651, MI = 0.606\n",
      "Pre-Train Epoch 0, Batch 121: Loss = 0.029, a = 0.000 b = 0.014, c = -0.966, Joint = -0.013, Marginal = -1.520, MI = 0.565\n",
      "Pre-Train Epoch 0, Batch 122: Loss = 0.030, a = 0.000 b = 0.014, c = -0.995, Joint = -0.008, Marginal = -1.431, MI = 0.530\n",
      "Pre-Train Epoch 0, Batch 123: Loss = 0.025, a = 0.000 b = 0.012, c = -0.979, Joint = 0.004, Marginal = -1.469, MI = 0.580\n",
      "Pre-Train Epoch 0, Batch 124: Loss = 0.023, a = 0.000 b = 0.012, c = -0.967, Joint = 0.004, Marginal = -1.601, MI = 0.584\n",
      "Pre-Train Epoch 0, Batch 125: Loss = 0.029, a = 0.000 b = 0.015, c = -0.875, Joint = -0.002, Marginal = -1.503, MI = 0.575\n",
      "Pre-Train Epoch 0, Batch 126: Loss = 0.030, a = 0.000 b = 0.016, c = -0.907, Joint = -0.002, Marginal = -1.366, MI = 0.562\n",
      "Pre-Train Epoch 0, Batch 127: Loss = 0.023, a = 0.000 b = 0.012, c = -1.032, Joint = 0.004, Marginal = -1.611, MI = 0.569\n",
      "Pre-Train Epoch 0, Batch 128: Loss = 0.025, a = 0.000 b = 0.012, c = -1.004, Joint = 0.000, Marginal = -1.379, MI = 0.579\n",
      "Pre-Train Epoch 0, Batch 129: Loss = 0.033, a = 0.000 b = 0.016, c = -1.009, Joint = -0.012, Marginal = -1.527, MI = 0.584\n",
      "Pre-Train Epoch 0, Batch 130: Loss = 0.025, a = 0.000 b = 0.013, c = -1.016, Joint = -0.002, Marginal = -1.546, MI = 0.584\n",
      "Pre-Train Epoch 0, Batch 131: Loss = 0.039, a = 0.000 b = 0.018, c = -1.025, Joint = -0.002, Marginal = -1.484, MI = 0.565\n",
      "Pre-Train Epoch 0, Batch 132: Loss = 0.026, a = 0.000 b = 0.013, c = -1.016, Joint = 0.008, Marginal = -1.313, MI = 0.514\n",
      "Pre-Train Epoch 0, Batch 133: Loss = 0.037, a = 0.000 b = 0.018, c = -1.008, Joint = -0.000, Marginal = -1.469, MI = 0.556\n",
      "Pre-Train Epoch 0, Batch 134: Loss = 0.032, a = 0.000 b = 0.016, c = -0.984, Joint = -0.002, Marginal = -1.228, MI = 0.512\n",
      "Pre-Train Epoch 0, Batch 135: Loss = 0.034, a = 0.000 b = 0.016, c = -0.917, Joint = -0.008, Marginal = -1.416, MI = 0.526\n",
      "Pre-Train Epoch 0, Batch 136: Loss = 0.032, a = 0.000 b = 0.016, c = -0.928, Joint = -0.003, Marginal = -1.329, MI = 0.543\n",
      "Pre-Train Epoch 0, Batch 137: Loss = 0.033, a = 0.000 b = 0.017, c = -1.004, Joint = 0.010, Marginal = -1.520, MI = 0.589\n",
      "Pre-Train Epoch 0, Batch 138: Loss = 0.037, a = 0.000 b = 0.018, c = -1.038, Joint = 0.001, Marginal = -1.434, MI = 0.597\n",
      "Pre-Train Epoch 0, Batch 139: Loss = 0.027, a = 0.000 b = 0.014, c = -0.998, Joint = 0.004, Marginal = -1.718, MI = 0.592\n",
      "Pre-Train Epoch 0, Batch 140: Loss = 0.034, a = 0.000 b = 0.017, c = -1.011, Joint = -0.009, Marginal = -1.569, MI = 0.575\n",
      "Epoch 1/150\n",
      "147/147 - 1s - loss: -1.6712e+00 - MI: 1.6718\n",
      "Epoch 2/150\n",
      "147/147 - 1s - loss: -1.8962e+00 - MI: 1.8957\n",
      "Epoch 3/150\n",
      "147/147 - 1s - loss: -1.9533e+00 - MI: 1.9538\n",
      "Epoch 4/150\n",
      "147/147 - 1s - loss: -1.9915e+00 - MI: 1.9915\n",
      "Epoch 5/150\n",
      "147/147 - 1s - loss: -2.0850e+00 - MI: 2.0847\n",
      "Epoch 6/150\n",
      "147/147 - 1s - loss: -2.1433e+00 - MI: 2.1434\n",
      "Epoch 7/150\n",
      "147/147 - 1s - loss: -2.1640e+00 - MI: 2.1646\n",
      "Epoch 8/150\n",
      "147/147 - 1s - loss: -2.1792e+00 - MI: 2.1795\n",
      "Epoch 9/150\n",
      "147/147 - 1s - loss: -2.2045e+00 - MI: 2.2047\n",
      "Epoch 10/150\n",
      "147/147 - 1s - loss: -2.1820e+00 - MI: 2.1829\n",
      "Epoch 11/150\n",
      "147/147 - 1s - loss: -2.2185e+00 - MI: 2.2176\n",
      "Epoch 12/150\n",
      "147/147 - 1s - loss: -2.2436e+00 - MI: 2.2435\n",
      "Epoch 13/150\n",
      "147/147 - 1s - loss: -2.2227e+00 - MI: 2.2228\n",
      "Epoch 14/150\n",
      "147/147 - 1s - loss: -2.2008e+00 - MI: 2.2001\n",
      "Epoch 15/150\n",
      "147/147 - 1s - loss: -2.2228e+00 - MI: 2.2224\n",
      "Epoch 16/150\n",
      "147/147 - 1s - loss: -2.2457e+00 - MI: 2.2468\n",
      "Epoch 17/150\n",
      "147/147 - 1s - loss: -2.2464e+00 - MI: 2.2463\n",
      "Epoch 18/150\n",
      "147/147 - 1s - loss: -2.2496e+00 - MI: 2.2486\n",
      "Epoch 19/150\n",
      "147/147 - 1s - loss: -2.2709e+00 - MI: 2.2710\n",
      "Epoch 20/150\n",
      "147/147 - 1s - loss: -2.2344e+00 - MI: 2.2343\n",
      "Epoch 21/150\n",
      "147/147 - 1s - loss: -2.2675e+00 - MI: 2.2676\n",
      "Epoch 22/150\n",
      "147/147 - 1s - loss: -2.2314e+00 - MI: 2.2313\n",
      "Epoch 23/150\n",
      "147/147 - 1s - loss: -2.1928e+00 - MI: 2.1914\n",
      "Epoch 24/150\n",
      "147/147 - 1s - loss: -2.2429e+00 - MI: 2.2436\n",
      "Epoch 25/150\n",
      "147/147 - 1s - loss: -2.2398e+00 - MI: 2.2395\n",
      "Epoch 26/150\n",
      "147/147 - 1s - loss: -2.2714e+00 - MI: 2.2718\n",
      "Epoch 27/150\n",
      "147/147 - 1s - loss: -2.2719e+00 - MI: 2.2722\n",
      "Epoch 28/150\n",
      "147/147 - 1s - loss: -2.2340e+00 - MI: 2.2335\n",
      "Epoch 29/150\n",
      "147/147 - 1s - loss: -2.2317e+00 - MI: 2.2319\n",
      "Epoch 30/150\n",
      "147/147 - 1s - loss: -2.2669e+00 - MI: 2.2678\n",
      "Epoch 31/150\n",
      "147/147 - 1s - loss: -2.2663e+00 - MI: 2.2666\n",
      "Epoch 32/150\n",
      "147/147 - 1s - loss: -2.2665e+00 - MI: 2.2667\n",
      "Epoch 33/150\n",
      "147/147 - 1s - loss: -2.2521e+00 - MI: 2.2510\n",
      "Epoch 34/150\n",
      "147/147 - 1s - loss: -2.2636e+00 - MI: 2.2640\n",
      "Epoch 35/150\n",
      "147/147 - 1s - loss: -2.2628e+00 - MI: 2.2624\n",
      "Epoch 36/150\n",
      "147/147 - 1s - loss: -2.2433e+00 - MI: 2.2420\n",
      "Epoch 37/150\n",
      "147/147 - 1s - loss: -2.2514e+00 - MI: 2.2514\n",
      "Epoch 38/150\n",
      "147/147 - 1s - loss: -2.2702e+00 - MI: 2.2710\n",
      "Epoch 39/150\n",
      "147/147 - 1s - loss: -2.2649e+00 - MI: 2.2634\n",
      "Epoch 40/150\n",
      "147/147 - 1s - loss: -2.2664e+00 - MI: 2.2666\n",
      "Epoch 41/150\n",
      "147/147 - 1s - loss: -2.2766e+00 - MI: 2.2763\n",
      "Epoch 42/150\n",
      "147/147 - 1s - loss: -2.2593e+00 - MI: 2.2603\n",
      "Epoch 43/150\n",
      "147/147 - 1s - loss: -2.2630e+00 - MI: 2.2631\n",
      "Epoch 44/150\n",
      "147/147 - 1s - loss: -2.2669e+00 - MI: 2.2666\n",
      "Epoch 45/150\n",
      "147/147 - 1s - loss: -2.2683e+00 - MI: 2.2686\n",
      "Epoch 46/150\n",
      "147/147 - 1s - loss: -2.2713e+00 - MI: 2.2715\n",
      "Epoch 47/150\n",
      "147/147 - 1s - loss: -2.2831e+00 - MI: 2.2812\n",
      "Epoch 48/150\n",
      "147/147 - 1s - loss: -2.2612e+00 - MI: 2.2601\n",
      "Epoch 49/150\n",
      "147/147 - 1s - loss: -2.3042e+00 - MI: 2.3050\n",
      "Epoch 50/150\n",
      "147/147 - 1s - loss: -2.2875e+00 - MI: 2.2880\n",
      "Epoch 51/150\n",
      "147/147 - 1s - loss: -2.2745e+00 - MI: 2.2727\n",
      "Epoch 52/150\n",
      "147/147 - 1s - loss: -2.2822e+00 - MI: 2.2821\n",
      "Epoch 53/150\n",
      "147/147 - 1s - loss: -2.3069e+00 - MI: 2.3072\n",
      "Epoch 54/150\n",
      "147/147 - 1s - loss: -2.2871e+00 - MI: 2.2874\n",
      "Epoch 55/150\n",
      "147/147 - 1s - loss: -2.3081e+00 - MI: 2.3081\n",
      "Epoch 56/150\n",
      "147/147 - 1s - loss: -2.3289e+00 - MI: 2.3304\n",
      "Epoch 57/150\n",
      "147/147 - 1s - loss: -2.3072e+00 - MI: 2.3069\n",
      "Epoch 58/150\n",
      "147/147 - 1s - loss: -2.2936e+00 - MI: 2.2931\n",
      "Epoch 59/150\n",
      "147/147 - 1s - loss: -2.3021e+00 - MI: 2.3018\n",
      "Epoch 60/150\n",
      "147/147 - 1s - loss: -2.3108e+00 - MI: 2.3102\n",
      "Epoch 61/150\n",
      "147/147 - 1s - loss: -2.3116e+00 - MI: 2.3106\n",
      "Epoch 62/150\n",
      "147/147 - 1s - loss: -2.2916e+00 - MI: 2.2923\n",
      "Epoch 63/150\n",
      "147/147 - 1s - loss: -2.3093e+00 - MI: 2.3096\n",
      "Epoch 64/150\n",
      "147/147 - 1s - loss: -2.2735e+00 - MI: 2.2736\n",
      "Epoch 65/150\n",
      "147/147 - 1s - loss: -2.3170e+00 - MI: 2.3171\n",
      "Epoch 66/150\n",
      "147/147 - 1s - loss: -2.3140e+00 - MI: 2.3140\n",
      "Epoch 67/150\n",
      "147/147 - 1s - loss: -2.2821e+00 - MI: 2.2813\n",
      "Epoch 68/150\n",
      "147/147 - 1s - loss: -2.2992e+00 - MI: 2.2989\n",
      "Epoch 69/150\n",
      "147/147 - 1s - loss: -2.3184e+00 - MI: 2.3181\n",
      "Epoch 70/150\n",
      "147/147 - 1s - loss: -2.2875e+00 - MI: 2.2881\n",
      "Epoch 71/150\n",
      "147/147 - 1s - loss: -2.3131e+00 - MI: 2.3126\n",
      "Epoch 72/150\n",
      "147/147 - 1s - loss: -2.3226e+00 - MI: 2.3223\n",
      "Epoch 73/150\n",
      "147/147 - 1s - loss: -2.3273e+00 - MI: 2.3274\n",
      "Epoch 74/150\n",
      "147/147 - 1s - loss: -1.6246e+00 - MI: 1.6272\n",
      "Epoch 75/150\n",
      "147/147 - 1s - loss: -2.2636e+00 - MI: 2.2633\n",
      "Epoch 76/150\n",
      "147/147 - 1s - loss: -2.2988e+00 - MI: 2.2996\n",
      "Epoch 77/150\n",
      "147/147 - 1s - loss: -2.3070e+00 - MI: 2.3070\n",
      "Epoch 78/150\n",
      "147/147 - 1s - loss: -2.3309e+00 - MI: 2.3315\n",
      "Epoch 79/150\n",
      "147/147 - 1s - loss: -2.2970e+00 - MI: 2.2955\n",
      "Epoch 80/150\n",
      "147/147 - 1s - loss: -2.3206e+00 - MI: 2.3201\n",
      "Epoch 81/150\n",
      "147/147 - 1s - loss: -2.3170e+00 - MI: 2.3164\n",
      "Epoch 82/150\n",
      "147/147 - 1s - loss: -2.3142e+00 - MI: 2.3136\n",
      "Epoch 83/150\n",
      "147/147 - 1s - loss: -2.3272e+00 - MI: 2.3268\n",
      "Epoch 84/150\n",
      "147/147 - 1s - loss: -2.3185e+00 - MI: 2.3196\n",
      "Epoch 85/150\n",
      "147/147 - 1s - loss: -2.3209e+00 - MI: 2.3212\n",
      "Epoch 86/150\n",
      "147/147 - 1s - loss: -2.3213e+00 - MI: 2.3219\n",
      "Epoch 87/150\n",
      "147/147 - 1s - loss: -2.3223e+00 - MI: 2.3221\n",
      "Epoch 88/150\n",
      "147/147 - 1s - loss: -2.3178e+00 - MI: 2.3184\n",
      "Epoch 89/150\n",
      "147/147 - 1s - loss: -2.3292e+00 - MI: 2.3302\n",
      "Epoch 90/150\n",
      "147/147 - 1s - loss: -2.3354e+00 - MI: 2.3375\n",
      "Epoch 91/150\n",
      "147/147 - 1s - loss: -2.3192e+00 - MI: 2.3202\n",
      "Epoch 92/150\n",
      "147/147 - 1s - loss: -2.3085e+00 - MI: 2.3096\n",
      "Epoch 93/150\n",
      "147/147 - 1s - loss: -2.3059e+00 - MI: 2.3050\n",
      "Epoch 94/150\n",
      "147/147 - 1s - loss: -2.3308e+00 - MI: 2.3306\n",
      "Epoch 95/150\n",
      "147/147 - 1s - loss: -2.3170e+00 - MI: 2.3166\n",
      "Epoch 96/150\n",
      "147/147 - 1s - loss: -2.3295e+00 - MI: 2.3287\n",
      "Epoch 97/150\n",
      "147/147 - 1s - loss: -2.3069e+00 - MI: 2.3072\n",
      "Epoch 98/150\n",
      "147/147 - 1s - loss: -2.3112e+00 - MI: 2.3118\n",
      "Epoch 99/150\n",
      "147/147 - 1s - loss: -2.3064e+00 - MI: 2.3060\n",
      "Epoch 100/150\n",
      "147/147 - 1s - loss: -2.2984e+00 - MI: 2.2982\n",
      "Epoch 101/150\n",
      "147/147 - 1s - loss: -2.3336e+00 - MI: 2.3336\n",
      "Epoch 102/150\n",
      "147/147 - 1s - loss: -2.3387e+00 - MI: 2.3388\n",
      "Epoch 103/150\n",
      "147/147 - 1s - loss: -2.3420e+00 - MI: 2.3422\n",
      "Epoch 104/150\n",
      "147/147 - 1s - loss: -2.3180e+00 - MI: 2.3177\n",
      "Epoch 105/150\n",
      "147/147 - 1s - loss: -2.3199e+00 - MI: 2.3207\n",
      "Epoch 106/150\n",
      "147/147 - 1s - loss: -2.3305e+00 - MI: 2.3315\n",
      "Epoch 107/150\n",
      "147/147 - 1s - loss: -2.3192e+00 - MI: 2.3191\n",
      "Epoch 108/150\n",
      "147/147 - 1s - loss: -2.3204e+00 - MI: 2.3206\n",
      "Epoch 109/150\n",
      "147/147 - 1s - loss: -2.3272e+00 - MI: 2.3269\n",
      "Epoch 110/150\n",
      "147/147 - 1s - loss: -2.3306e+00 - MI: 2.3309\n",
      "Epoch 111/150\n",
      "147/147 - 1s - loss: -2.3244e+00 - MI: 2.3250\n",
      "Epoch 112/150\n",
      "147/147 - 1s - loss: -2.3576e+00 - MI: 2.3579\n",
      "Epoch 113/150\n",
      "147/147 - 1s - loss: -2.3291e+00 - MI: 2.3286\n",
      "Epoch 114/150\n",
      "147/147 - 1s - loss: -2.3258e+00 - MI: 2.3270\n",
      "Epoch 115/150\n",
      "147/147 - 1s - loss: -2.3271e+00 - MI: 2.3274\n",
      "Epoch 116/150\n",
      "147/147 - 1s - loss: -2.3416e+00 - MI: 2.3405\n",
      "Epoch 117/150\n",
      "147/147 - 1s - loss: -2.3129e+00 - MI: 2.3132\n",
      "Epoch 118/150\n",
      "147/147 - 1s - loss: -2.3303e+00 - MI: 2.3300\n",
      "Epoch 119/150\n",
      "147/147 - 1s - loss: -2.3657e+00 - MI: 2.3662\n",
      "Epoch 120/150\n",
      "147/147 - 1s - loss: -2.3372e+00 - MI: 2.3383\n",
      "Epoch 121/150\n",
      "147/147 - 1s - loss: -1.9907e+00 - MI: 1.9910\n",
      "Epoch 122/150\n",
      "147/147 - 1s - loss: -2.3101e+00 - MI: 2.3109\n",
      "Epoch 123/150\n",
      "147/147 - 1s - loss: -2.3128e+00 - MI: 2.3127\n",
      "Epoch 124/150\n",
      "147/147 - 1s - loss: -2.3115e+00 - MI: 2.3112\n",
      "Epoch 125/150\n",
      "147/147 - 1s - loss: -2.3328e+00 - MI: 2.3331\n",
      "Epoch 126/150\n",
      "147/147 - 1s - loss: -2.3504e+00 - MI: 2.3497\n",
      "Epoch 127/150\n",
      "147/147 - 1s - loss: -2.3354e+00 - MI: 2.3324\n",
      "Epoch 128/150\n",
      "147/147 - 1s - loss: -2.3283e+00 - MI: 2.3293\n",
      "Epoch 129/150\n",
      "147/147 - 1s - loss: -2.3584e+00 - MI: 2.3579\n",
      "Epoch 130/150\n",
      "147/147 - 1s - loss: -2.3484e+00 - MI: 2.3488\n",
      "Epoch 131/150\n",
      "147/147 - 1s - loss: -2.3327e+00 - MI: 2.3324\n",
      "Epoch 132/150\n",
      "147/147 - 1s - loss: -2.3467e+00 - MI: 2.3466\n",
      "Epoch 133/150\n",
      "147/147 - 1s - loss: -2.3326e+00 - MI: 2.3327\n",
      "Epoch 134/150\n",
      "147/147 - 1s - loss: -2.3488e+00 - MI: 2.3484\n",
      "Epoch 135/150\n",
      "147/147 - 1s - loss: -2.1184e+00 - MI: 2.1193\n",
      "Epoch 136/150\n",
      "147/147 - 1s - loss: -2.3337e+00 - MI: 2.3324\n",
      "Epoch 137/150\n",
      "147/147 - 1s - loss: -2.3117e+00 - MI: 2.3120\n",
      "Epoch 138/150\n",
      "147/147 - 1s - loss: -2.3378e+00 - MI: 2.3381\n",
      "Epoch 139/150\n",
      "147/147 - 1s - loss: -2.3171e+00 - MI: 2.3151\n",
      "Epoch 140/150\n",
      "147/147 - 1s - loss: -2.3449e+00 - MI: 2.3445\n",
      "Epoch 141/150\n",
      "147/147 - 1s - loss: -2.3403e+00 - MI: 2.3411\n",
      "Epoch 142/150\n",
      "147/147 - 1s - loss: -2.3345e+00 - MI: 2.3335\n",
      "Epoch 143/150\n",
      "147/147 - 1s - loss: -2.3297e+00 - MI: 2.3298\n",
      "Epoch 144/150\n",
      "147/147 - 1s - loss: -2.3421e+00 - MI: 2.3404\n",
      "Epoch 145/150\n",
      "147/147 - 1s - loss: -2.3424e+00 - MI: 2.3419\n",
      "Epoch 146/150\n",
      "147/147 - 1s - loss: -2.3317e+00 - MI: 2.3313\n",
      "Epoch 147/150\n",
      "147/147 - 1s - loss: -2.3543e+00 - MI: 2.3541\n",
      "Epoch 148/150\n",
      "147/147 - 1s - loss: -2.3354e+00 - MI: 2.3355\n",
      "Epoch 149/150\n",
      "147/147 - 1s - loss: -2.3418e+00 - MI: 2.3413\n",
      "Epoch 150/150\n",
      "147/147 - 1s - loss: -2.3358e+00 - MI: 2.3352\n"
     ]
    }
   ],
   "source": [
    "from ifn.Architectures.utils import build_gIFN_DNN\n",
    "from ifn.Architectures.ifn import IFN, mine_loss, MI\n",
    "\n",
    "\n",
    "X, Y = np.expand_dims(mjjs_firsthalf/1000.0,1) , np.expand_dims(mjjsg_firsthalf/1000.0,1)\n",
    "\n",
    "# Build model\n",
    "ifn = build_gIFN_DNN(1, 1, [32, 32, 32], l2_reg = 0, d_l1_reg= 1e-3, )\n",
    "ifn.pre_train([X, Y], epochs = 1, batch_size = 1024, verbose = True)\n",
    "opt = tf.keras.optimizers.Adam(clipnorm = 1.0)\n",
    "ifn.compile(loss=mine_loss, optimizer=opt, metrics = [MI,])\n",
    "\n",
    "\n",
    "# Fit\n",
    "history = ifn.fit([X, Y],\n",
    "            batch_size= 1024,\n",
    "            epochs = 150,\n",
    "            sample_weight= weights_firsthalf,\n",
    "            shuffle=True, verbose= 2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pres_mass = np.squeeze(ifn.maximum_likelihood(np.expand_dims(mjjs_secondhalf/1000.0, 1)))\n",
    "\n",
    "\n",
    "mybins = np.logspace(np.log(2000)/np.log(10),np.log(5000)/np.log(10),50)\n",
    "mybinsB = np.linspace(0.8,1.4,50)\n",
    "\n",
    "resmass = 2800\n",
    "\n",
    "fig = plt.figure(figsize=(6, 6))\n",
    "gs = gridspec.GridSpec(1, 1) \n",
    "plt.subplots_adjust(hspace = .001)\n",
    "\n",
    "ax0 = plt.subplot(gs[0])\n",
    "ax0.yaxis.set_ticks_position('both')\n",
    "ax0.xaxis.set_ticks_position('both')\n",
    "ax0.tick_params(direction=\"in\",which=\"both\")\n",
    "plt.xticks(fontsize=20)\n",
    "plt.yticks(fontsize=20)\n",
    "ax0.minorticks_on()\n",
    "\n",
    "_,_,_=plt.hist(mjjs_secondhalf/mjjsg_secondhalf,weights=weights_secondhalf,bins=mybinsB,alpha=0.2,density=True,color=\"blue\",label=\"QCD, Geant4\")\n",
    "_,_,_=plt.hist(1000*pres_mass/mjjsg_secondhalf,weights=weights_secondhalf,bins=mybinsB,density=True,color=\"black\",histtype=\"step\",label=\"QCD, Geant4 + GA\",lw=3)\n",
    "\n",
    "ww,bins = np.histogram(mjjsg_secondhalf,bins=mybins)\n",
    "ww = np.interp(mjjsg_secondhalf,0.5*(bins[:-1]+bins[1:]),ww)\n",
    "\n",
    "_,_,_=plt.hist(mjjs_secondhalf/mjjsg_secondhalf,weights=norm.pdf(mjjsg_secondhalf,resmass,10)/ww,bins=mybinsB,alpha=0.7,density=True,color=\"orange\",label=\"BSM, Geant4\")\n",
    "_,_,_=plt.hist(1000*pres_mass/mjjsg_secondhalf,weights=norm.pdf(mjjsg_secondhalf,resmass,10)/ww,bins=mybinsB,density=True,color=\"black\",histtype=\"step\",label=\"BSM, Geant4 + GA\",lw=3,ls=\":\")\n",
    "\n",
    "plt.ylabel(\"Normalized to Unity\",fontsize=20)\n",
    "#plt.xlim([mymin,mymax])\n",
    "#plt.yscale(\"log\")\n",
    "#plt.ylim([1e-6,1e-1])\n",
    "#plt.locator_params(axis='y', nbins=6)\n",
    "plt.axvline(1,alpha=0.5)\n",
    "plt.xlabel(\"reconstructed / true $m_{jj}$\",fontsize=20)\n",
    "plt.legend(frameon=False,fontsize=15,ncol=1)\n",
    "_=plt.title(\"Pythia 8 Dijets\",loc=\"right\",fontsize=20, fontstyle=\"italic\")\n",
    "plt.savefig(\"NN_prior_GA.pdf\",bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data-based Optimal Transport Calibration"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "x0_x = np.sort(mjjs_delphes_firsthalf)\n",
    "w0_x = [x for _, x in sorted(zip(mjjs_delphes_firsthalf, weights_delphes_firsthalf))]\n",
    "x0_ecdf = np.cumsum(w0_x)/np.sum(w0_x)\n",
    "\n",
    "x1_x = np.sort(mjjs_firsthalf)\n",
    "w1_x = [x for _, x in sorted(zip(mjjs_firsthalf, weights_firsthalf))]\n",
    "x1_ecdf = np.cumsum(w1_x)/np.sum(w1_x)\n",
    "\n",
    "def my_mjj_calib(x):\n",
    "    firstpart = np.interp(x,x0_x,x0_ecdf)\n",
    "    return np.interp(firstpart,x1_ecdf,x1_x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "<ipython-input-20-a5e09e17bcb3>:59: RuntimeWarning: divide by zero encountered in true_divide\n",
      "  plt.plot(0.5*(b[1:]+b[:-1]),nt0/(nt1+0),color='black',ls=\":\")\n",
      "<ipython-input-20-a5e09e17bcb3>:59: RuntimeWarning: invalid value encountered in true_divide\n",
      "  plt.plot(0.5*(b[1:]+b[:-1]),nt0/(nt1+0),color='black',ls=\":\")\n",
      "<ipython-input-20-a5e09e17bcb3>:60: RuntimeWarning: divide by zero encountered in true_divide\n",
      "  plt.plot(0.5*(b[1:]+b[:-1]),nt/(nt1+0),color='blue',lw=4,alpha=0.7)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhoAAAIKCAYAAABlb+a9AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAADNr0lEQVR4nOzdeVxN+f8H8NdtLyGRFhFKhWSraRBKliwJUyYlw1iHGuuUbGXGMvYZkWXsZBo7IfsyGEtkyTdbSlqUtGqv2/39cX73cOtW99a93W69n49HD+ee9X2uu7zvZ+XweDweCCGEEEKkQEHWARBCCCGk/qJEgxBCCCFSQ4kGIYQQQqSGEg1CCCGESA0lGoQQQgiRGko0CCGEECI1lGgQQgipc9LT02FmZoaZM2dWuW9RURG6du2KcePG1UJktWPChAmwsLBAfn4+ACAiIgJmZmZYv369jCMTn5KsAyCEEFL3OTs74+XLlwLrVFRUYGBgAAsLC0yZMgUdO3as1rkzMjJw6NAhTJkyBerq6gDAXkuUcxYVFWHDhg1o1apVta5fkezsbBw4cACXL19GYmIiOBwO2rVrB3d3dzg7O4PD4VR6fFZWFr755huBdWpqatDR0UGPHj3w/fffo2fPnkKPnTx5MmbMmME+H3p6eti6dSs6d+5crXsR9hzXFko0CCGEVKqoqAhv376FkZERfHx82PWFhYWIjY1FSEgIzp07hxUrVsDFxUXs8584cQL79++Ht7c3u+7FixcAREs0NDU1MXDgQLGvW5m0tDS4urri48eP+P777/Hjjz+ioKAAp0+fhq+vL16+fImFCxdWeg7+PQwcOBCjR48GAOTl5eHt27c4c+YMTp8+DU9PTyxevLhc0tK/f3+BxwYGBjAwMKj2/Qh7jmsLJRqEEEIq9ebNGxQXF6Nr165Cv9DHjBmD4cOHY9WqVRgxYgTU1NREPjePx8ORI0dgamoqsJ7/JW1ubl6z4Ktp3759SExMxK+//orvv/+eXT969Gg4Ojri0KFDmDNnTqX3yr8HOzu7cs/bzJkzMX/+fBw8eBBaWlrw8vKSzo2g4ue4tlAbDUIIIZWKiooCUHHpAr/6JDc3F+/evcP+/fthZmaG8+fPl9u3tLQUTk5OsLS0xJYtW2Bubo53797h0aNHMDMzw4YNGwAwVSfa2tpQVlbG/PnzYWNjgx49emDixImIjY0VOOfkyZPRsWNHtj0DwLTxWL16NYYMGYLu3bvDxsYGEyZMwIMHD0S657dv3wIAHBwcBNarqKigTZs2KC4uRkFBQaXn4D9vwpIlVVVVrFu3Djo6Ovjrr7+Qnp7Obtu5cyfMzMzw77//AmAShR49euC7774TOEdycjJ+/fVXDBw4EF27dsXgwYOxdOlSJCUlsfv8/fffFT7Hd+/exU8//QQ7Ozt06dIFtra28Pb2LldFVlOUaBBCCKmUKNUYnz9/BgBoaGiw7QhevXpVbr/Q0FC8fv0akyZNgqOjI9zc3AAAEydOxNatW/H999+zVTL6+voYP348tLW1sXjxYnz33Xe4e/euQPUNwCQlRkZGbNuD/Px8uLm54fDhwxg4cCB+++03zJ07Fx8/fsTkyZNF+iLlV1N8+vRJYH1ubi5evXqFtm3bQktLq9JzvHjxAoqKihWWJKirq2PEiBEoKCjAnTt3BI4DvjzfcXFxyM3NRadOndh9oqKiMHr0aNy5cwdubm5YuXIlnJyccOHCBYwZMwYpKSkAAGtra6HP8ZUrVzBp0iSkpKRg6tSpWLNmDSZMmIBHjx5hwoQJSE1NrfI5EhVVnRBCCKkU/5e5mZmZ0O0vX77EmzdvYGRkhDZt2qB58+ZQUFAo94VeVFSEzZs3Q0dHB9OmTUOjRo2grKwMAHBycoKFhQUAIDIyEiUlJXj58iX27dvHNqgcOXIkHj58iOfPn6OwsBCqqqr49OkTPn36JNDo8u3btzA3N4eXlxdGjhzJrm/fvj08PT1x5cqVKqtkfvjhB5w9exYLFy6En58f2rdvj6SkJGzevBn5+fnYuHFjpccXFBQgJiYGbdu2haqqaoX7GRsbA2CSia+fzxYtWkBHRwdA+RKlgoICeHl5wcTEBLt374aKigp7rImJCebMmYNjx45h1qxZMDExEfocr1q1ChwOBwcOHICmpiZ7fI8ePbB//358+PCBvX5NUaJBCCGkQqWlpXj16hV0dHSgpKSE7Oxsdn1qairCw8Oxbds2KCgoYPny5QCARo0aoV27duUSjX/++QcJCQlYsWIFGjVqBID5UlVSUhL41c8/zs3NrVyvDQ0NDfB4PLbxpLC2HBYWFti8eTMAsFUcPB6Pjb2wsLDK+9bV1YW3tzdWrlyJCRMmsOvHjBmDkydPol27dpUe//r1a3C5XJHbmBQVFQFgkoi4uDj06tWL3ca/R36JRkhICBITE7F27VoUFBQIVOG0bdsWANgSDUD4c6yiooLS0lI8efIEtra27HorKytYWVmJFLOoKNEghBBSobi4OOTl5SEvLw/W1tbltisoKMDGxgbz589Hly5d2PWdOnVCaGgoMjMzoaWlhdzcXGzbtg3m5uYCbQ1evnyJ9u3bC/wq53+xDh8+vNz13r17B0NDQ3Z/YYnGkydPsGfPHjx69Khc1QfAlGxU5u3bt5g1axYKCgrg4+ODdu3aIT09HcHBwTh79iz69u1bZaIhamNWfkLQvHlzAEx1U9kE5cWLF1BQUGBLlC5dugQA8PDwqPC8enp67LKw53jatGm4ffs2pk2bhn79+mHEiBFwcHCQStdXSjQIIYRUiP+F6e7uDkdHR3Y9h8NBo0aNYGRkJFD0zmdhYYHQ0FC8evUKNjY22LdvH9LS0rBx40YoKDDNAxMSEvD582fY29uXu6aKigosLS0F1qempuLTp08YPHgwu47fDoRfrXD9+nXMmjULjRo1gru7O7p06QJNTU1wOBxs2bIFDx48qLStSUlJCWbPno3MzEycP38e2tra7LahQ4di5MiRWLp0Kezs7KChoVHheSprCPq1hw8fAgDbrkXY+CEvXrxAu3bt2CTgxYsX6Nq1K+bPn1/hefklGxU9x506dcLZs2exZ88ehIaG4vr162jcuDG8vLwwceLESmMWFyUahBBCKsT/wrSzs4ONjY3Ix/G/OF+/fg1TU1Ps2bMHDg4O+Pbbb9l9+F+qX38Z83g8vHr1Ch06dGDbFvAJKyV48eIFmjdvjpYtWwIANm7cCC6Xi927d5dLVJYtWwZVVVWYmJhUer9v3rzBuHHjBJIMgBlsq0ePHjh58iTev39faRIhSolGfHw87t+/j1atWqFHjx5Cj/v48SM+ffrEPm/80iU9PT2R/j+EPcd8enp6WLRoEXx9fXH79m38+eefWL16NTQ1Nas1HkpFqNcJIYSQCvG/+CpqCFqRjh07gsPh4O3bt9i5cycKCwuF9hbh78v3/v175ObmCv1iLLt/QUEB3r17x+5bUlKC169fw8DAoFySERUVhXfv3sHU1BRKShX/xv748SOAL1UZZSUmJgJApT1OuFwuXr16BW1tbTYBEraPn58fSktLsWDBAraU5+XLl1BTU2OrZsq2z1BSUgKHw2HjLIvfLZdPlBFWFRUV0b9/f+zevRscDgdXrlypcN/qaNCJxqtXrzBixAgMGDBA1qEQQkidFBUVBS0tLYE6f1FoamrCyMgIT548weHDh+Hh4cEW5/O9e/cOAATaO5T9Yv1a2W1lG1wqKSmhcePGyMvLYxtXAsxQ4CtXrgQAdOjQodK4+bHcvn0bpaWlAtvu3buH8PBwWFhYVPp8xMbGoqCgoMLSjJSUFEyaNAnh4eGYNGkShg0bBkCwNEdRUVHgnvmJgoqKCszNzfHs2TP2+eM7evQohg8fjtu3b7Pryj7HPB4PS5YsYRvufi01NRU8Hq9cSU5NNdiqk/Pnz2P16tWwtLRk/yMJIYR8kZKSgvT09HI9P0RlYWGBs2fPQktLC7NmzSq3ncvlAgD279+PLl26YPjw4ZVWObx48QLNmjVjv+SFVQs4OTnh8OHDmDt3LoYOHYr4+HicPHkSM2fOZLvGXrx4EUOGDBEas7GxMUaMGIGzZ8/C09MTTk5OUFNTw7Nnz3DkyBE0bdqUTVoqwq9uUlZWZksHioqKkJKSgsePH+PatWtQUlJCQECAwERw/Ia3X9+PsMHSZs+ejZ9++gmenp6YOHEidHR0cPfuXZw6dQp2dnbo06dPpc+xqqoqDh06hI8fP2LAgAFQU1PDu3fvEBISAg0NDYFeNpLQYBONvLw8/PPPP7h27RolGoQQIkRV42dUhd8WwtvbG02aNCm3/YcffmBLPOzs7DB8+HC8fPkSHA6nXKKRn5+PuLg4gXYJwqoFfHx8oKCggAsXLuDu3bvo2rUrNmzYgE6dOuHChQv477//EBwcXGGiAQBr1qyBtbU1jh49ig0bNqCgoAAGBgZwc3PD1KlToaurW+l9879Tbt68iZs3bwJgkg4dHR0YGhpiwYIFcHJyKlc9U1FDUH19fTRr1oxdZ29vj7/++gvbt29HYGAgOBwOjIyM4OvrC09PT4F5U4Q9x4sXL4a+vj5Onz6NFStWoLi4GDo6OujduzemTZvGju0hKRwej8eT6BmrqaioCH/88Qf27NkDa2trHDx4UOh+mZmZ2LJlC65evYrU1FRoaWmhf//+mD17doV1YZU5dOgQ9uzZg2vXrtX0FgghhPy/oqIijBw5EsrKyjh58mSl7SKIcDk5ObCysoK9vT22bdsm63CqrU78z8fExGDBggWIjY1FZXlPQUEBPD09ERsbCw8PD1hYWCAuLg67d+/GvXv3cOLECTRt2rQWIyeEECLMypUrERcXh+DgYEoyqunRo0fg8XgVTiUvL2T+v5+VlYUxY8bAyMgIx48fx9ChQyvcd//+/Xj9+jWWLVsmMFCJubk5Zs2ahaCgIPj5+eH69euYMWOG0HN4eXnJZJpcQgip716+fIlXr17hwoULuHbtGn755Re22yYR3YcPHxAREYHdu3dDWVlZYNwQeSTzRKO4uBjOzs5YtGhRpePBA8CpU6egoaEBV1dXgfUODg7Q09PDmTNnsHDhQtjb2wudzIcQQoj0BAcH4+TJkzA0NMSaNWswatQoWYckl+7duwd/f3+0aNECq1evRps2bWQdUo3IPNFo0aKF0G42ZeXk5CAmJgZWVlYCw6gCzAh1lpaWuHTpEhISEtC6dWtphUsIIaQCv/32G3777TdZhyH3Ro8ejdGjR8s6DImReaIhKv4gKRX1XdbX1wfAjLQm6USjoKAAz58/h46ODtu3mRBCCKmvuFwuUlNTYWFhATU1tRqdS24SjdzcXACo8Ib5Y8Dz96vKkCFDkJSUhNLSUpSUlLCTAV24cAGtWrUS2Pf58+eVTl5DCCGE1EfBwcE1ns1VbhINSbt48aLI++ro6ABgnnBxR8cjhBBC5E1ycjI8PDzY77+akJtEgz87YH5+vtDteXl5AvtJEr+6RE9PD4aGhhI/PyGEEFIXSaK5gNzMdWJoaAgOh4Pk5GSh25OSkgAARkZGtRkWIYQQQiohN4mGhoYGzMzMEBUVhcLCQoFtXC4Xjx8/hr6+PgwMDGQUISGEEELKkptEAwBcXFyQn5+PkJAQgfVnzpxBWloaXFxcZBQZIYQQQoSReRuN6OhoREdHC6xLT0/HhQsX2Mf9+/eHuro63NzcEBoairVr1yIpKQkWFhaIjo7G3r17YWpqismTJ9d2+IQQQgiphMwTjbCwMGzZskVgXXR0NGbPns0+vnr1KgwNDaGsrIw9e/YgMDAQly5dQnBwMLS1teHq6gpvb2+2iyshhBBC6oY6M3trXZaQkAAHBwc24SGEEELqM0l+78lVGw1CCCGEyBdKNAghhBAiNZRoEEIIIURqKNEghBBCiNRQokEIIYQQqZF591ZSux49ks11e/aUzXXlCb+V9969e9G7d+8q9z9x4gSWLFmCqKioWoiOODnJ5rqhobK5rjwxMzPDihUr4OrqWuW+9+/fx4QJE3Dz5k2aJLOWUKJB6rT09HTs3r0bV69eRXJyMhQUFGBsbAxnZ2e4ublBSUnwJfy///0Pu3btQnh4OLKzs9GkSRP06NEDU6ZMgaWlJYAvHzTKysrgcDgAmBl6raysMGHCBFhYWIgcHz85+PpcWlpa6N69O8aNG4devXpJ6JkgRHSenp54+PAh+/5QVlaGvr4+nJycMH36dPa1yuVysXv3bpw7dw4JCQkoLCyErq4uhgwZgp9//hlqamrs+R48eIBt27ZhwIABQq/34MEDHDhwADY2NiLFaGZmBiUlJSgoKIDH46Fx48awsLCAi4sLhgwZIqFngtQFVHVC6qykpCSMGTMG0dHR2LRpEyIiInDv3j14eXnh4MGDmD59OkpKStj9L1++DDc3NxgZGeHEiRN4+vQpQkJCoKurC3d3d1y/fl3g/OfOnUNkZCQePXqEv/76CwYGBnB3d8fx48fFjnXnzp2IjIzE48ePcfjwYXTv3h0zZ84sNxgdIbVl+PDhiIyMRGRkJB48eABfX18EBQVh//797D5r167FoUOHsGTJEty9exdPnjzB+vXrceHCBfj5+QmcT0dHR+h7Iz4+HrGxsdWKMSAgAJGRkXj27BmOHz+OIUOGICAgAIsWLarW+UjdRIkGqbP8/f3RuHFjBAUFoWPHjlBQUICKigr69++PAwcO4OnTpzh06BAAIDc3F0uWLMHYsWMxZ84ctGzZEhwOB4aGhli8eDFmzJiBtLQ0oddRUVGBsbEx5syZAx8fHyxfvpydDVhcSkpKaN26NSZNmoSNGzciMDAQT58+ZbeHhYVhzJgx6N69O3r16oWlS5ciJydH6Lk8PT3h7++PlStXwsbGBt26dcPChQvLTSoYEREBZ2dndOnSBY6Ojnj8+DG7LSMjA76+vujfvz+6du2K0aNH4+bNm+z2+Ph4zJgxAzY2NujevTtGjRqFK1euVOveSd2lpKSEfv36oW3btoiPj2fX//vvvxg0aBCsra2hoqICJSUldO/eHYGBgRg5cqTAOfr164ebN2+Wex+dPHkSdnZ2NYpPQUEBBgYGcHFxwZ49e3Dq1CmcP3+e3X7v3j24u7vDysoK1tbWmDt3LlJTU4Wea+HChZg6dSp27twJW1tbdOnSBdOnT0dWVpbAfjExMXBzc4OlpSXs7Oxw7do1dltBQQFWrFiBAQMGwNLSEkOHDsWpU6fY7RkZGZg/fz569+6Nbt26wdHREUeOHKnRc1CfUaJB6qTMzEzcunULkydPhqKiYrnt/OLd06dPAwDu3LmDzMzMCue78fLyEmnSPTc3N6ioqODixYs1uwEA9vb2MDExQej/V7L/999/8PX1xcyZM/Hw4UP8888/eP78OVauXFnhOc6ePYu2bdvi1q1bCAkJwc2bNxEUFMRu5/F4OHr0KPbs2YO7d+9CT08Pv/32G7vdy8sLWVlZOH78OMLDw+Hi4oKZM2eyXzYBAQHQ0tLCjRs3EB4ejkmTJuGXX35BZmZmje+f1B0FBQUICwtDfHw8hg8fzq7v0KEDLl68iJs3b4LL5bLrO3fuDHt7e4Fz6Ovrw8rKin3PAczr79SpUxgxYoTEYu3YsSNsbW1x5swZAMyUFNOnT8fIkSNx9+5dnD9/HtnZ2Zg/f36F54iIiEB2djauXr2K8+fP4927dwLvCwA4cOAA1q9fj/DwcHzzzTdYvHgxSktLAQDLli3D06dPsX//fkRERGDevHlYvHgxwsPDAQCbNm1CRkYGLly4gIiICCxduhSrV68uN28XYVCiQeqk9+/fg8fjwdjYuMJ92rdvzxbZvnv3DhoaGjAwMKjRdZWUlNCmTRvExcXV6Dx8xsbG7LmCg4MxaNAgDBw4EIqKimjTpg28vb0RGhqKgoICoccbGBjAw8MDKioqMDc3h5OTEy5dusRuLy0txZQpU9C8eXNoamrCwcEBb968AQC8fPkSDx8+hK+vL1q0aAEVFRV4eHjAzMyMLQL//PkzFBUV2V+zzs7OiIiIgJaWlkTun8jOuXPn0KVLF3Tp0gXdunWDj48PZs6cie7du7P7+Pv7o3Pnzpg2bRq+/fZbTJs2DTt37qzw9e/i4iJQfXL//n0oKCiI3C5DVMbGxnj//j0A4MiRI+jYsSPc3NygrKwMHR0d+Pj44P79++w+ZSkqKmLOnDlQVVVF69at4e7ujsuXL+PrGTfc3d1haGgIVVVVODo6Ij09HWlpacjMzERoaChmz56N1q1bQ0lJCYMGDcKAAQPYUovPnz9DQUEBqqqqUFBQQJ8+fRAREQETExOJPg/1BTUGJXUS/wOB/wtDGC6Xy+7H4XCgrKwskWtzudxyjUyrq6SkhC2RiYmJQVxcnECiADD3mJKSIrTkpn379gKPW7dujeTk5HLr+NTU1FBUVMReD0C5InAej8d+IM6ZMwcLFixA37590atXL/Tr1w9Dhw6FiopKdW6X1CHDhw/H+vXrATCvw+joaCxbtgyRkZHYvHkzAKB58+bYsWMHPnz4gPDwcERERODIkSPYuHEjZsyYgTlz5gicc9CgQfj111/x5MkTdOvWDSdPnsTo0aPZxqWSUvZ98/TpU3Tp0kVgH0VFRSQkJKBNmzbljm/Tpo3Ae7h169YoKChARkaGwDo+fqPXwsJCJCUlobS0FDNmzBC4Lx6Ph65duwIApk+fjlmzZsHW1hY2NjawtbXFiBEjoKmpKYG7r38o0SB1Utu2bcHhcPD69Wv2zV3W27dv2RKP9u3bIysrC+/fvxf6wSOq/Px8xMTEYNy4cdU+Bx+Px8OLFy/YYmU1NTW4u7tjyZIlQvdPSEgot65sosXj8cp9qCsoCC+YVFVVBQDcvn0bTZs2FbpP7969cePGDdy/fx///fcfNmzYgO3bt+Po0aP0oVmPKCkpwdzcHD4+PvDw8EBsbCzatWvHbtfX18fIkSPZpHTLli0IDAzEmDFjBN5PqqqqGDlyJI4fP44OHTrgypUrOHv2rMTjjYqKQocOHQAw7xs7Ozts27ZN5OOFvW8AwfdKVe+bI0eOoFOnTkL3MTc3x6VLlxAREYHbt29j79692LJlC/755x+0atVK5DgbCqo6IXVS06ZNYWtri127drG/0L/24cMHnD9/nv1g7NOnD7S1tREYGCj0fGvWrCnXil6Yffv2QVlZGYMHD67ZDQAIDQ1FcnIyRo0aBYBJnl68eCGwT3Z2dqXtId69eyfw+P379yJXD7Vt2xYAyo2zER8fz37wpqenQ0VFBX379oWvry/Onz+P5ORk/PfffyJdg8in/Px8JCYmIiAgAB8/fiy3nd+48+sSAD4XFxeEhYUhLCwM3bp1g76+vkRje/jwIR4+fIgxY8YAYF7Hr169EkgeCgsLkZKSUuE5EhMTBXqkxcfHQ0NDQ6QqwdatW0NRUbHc+yYpKYk9Z3Z2NkpLS9mGqaGhoVBTUytXWkkYlGiQOmvp0qXIzs7G1KlT8fz5c5SWlqKoqAi3bt3Cjz/+CFtbW4wfPx4A86tn9erVCAsLg6+vLxITE8Hj8ZCYmIgVK1YgJCSE/cIXJiUlBZs2bUJQUBBWrFgBbW1tAMCzZ8/g6OgoVi+U9PR07Nu3D0uXLoWPjw9b6jJhwgQ8evQIwcHBKCgoQGpqKhYsWIC5c+dWeK7ExESEhISgqKgIL168wNmzZ+Ho6ChSHMbGxrC1tcWaNWsQFxcHLpeLy5cvY/jw4Xj06BHy8vIwePBg7NmzB/n5+SgtLcWzZ89QVFTEJimk/oiPj8fmzZthaWkJc3Nz6Ojo4O7du5g7dy4iIyNRVFQELpeLN2/eYNOmTTAxMUHnzp3Lncfc3Bxt27ZFUFAQvvvuO6HXSklJKdcDqio5OTk4ceIEZs6cCU9PT9ja2gJgGminpqbijz/+QE5ODrKysrB8+XL88MMPFVatFhcXIzAwEAUFBYiPj8fff/8t8vumUaNGcHFxwdatWxEVFQUul4vw8HCMHj0a58+fB4/Hg6urK9avX4/Pnz8DAN68eYOsrKxyVZ2EQVUnDYw8jdDJHw8jMDAQM2bMQEZGBkpLS2FmZgY3NzdMmDBBoBrBzs4OR48exY4dOzB27Fh8/vwZzZs3h42NDY4dO1auYenw4cPB4XDYwYKsra0RHBzMDuwFML/8YmNjBVrkCzNt2jT2XOrq6ujatSuCgoLQp08fdp/u3buzVRO///47mjRpAltbW/j6+lZ4Xnt7e0RHR6Nv374oKirCsGHDMG3aNJGfw3Xr1mHVqlVwdXVFcXExjIyMsGbNGlhZWQEAtm/fjvXr1yMwMBAcDgdt2rTBmjVrYGpqKvI1Ggp5G6Hz3LlzbO8pDocDbW1t2Nvbw9vbm+0qfvjwYWzbtg3z589HamoqSkpKoKurCzs7O6xbt67Ctkqurq7YuHEjBg4cKHR7cXExYmNjK2zkzBcQEIBff/0VADOoWKdOnRAQEIBhw4ax+xgaGmLHjh3YtGkT9u3bBw0NDfTs2RN//fVXhdUf5ubmUFdXx8CBA5GdnQ1bW1ssXry4yueMz8/PD0pKSpgyZQpyc3NhYGCAn3/+mS1B3bp1K1auXAk7OzuUlpZCT08P3t7e6N+/v8jXaEg4vK+b4RKh+KM/Xr16FYaGhrIOp8EKDg7G6tWr8e+//7IlDrVh2rRpWL9+PZo0aVJr1wSYcTR0dXXZBn2EyJOAgAB8//336NixY61ed+HChYiLi8Pff/9dq9etbyT5vUdVJ0RujBgxAk2aNMHy5cuRk5MjUAcrLWlpacjPz6/1JIMQeVZUVIRnz55RyRgBQIkGkSNNmzbF9u3b8f79e/Tq1atc1ztpaN68OQ4ePCj16xBSn6ioqODEiRNCu2yThofaaBC5YmlpiZMnT8o6jFpBCQ4h4vv9999lHQIpg0o0CCGEECI1lGgQQgghRGoo0SCEEEKI1FCiQQghhBCpoUSDEEIIIVJDiUYDwuFwZP5HSH2RlJTEvq6FzT/z+vVrdruw8SQePXrEbu8pZMjeGzdu0HuH1AuUaBBSRyQkJMDMzEzkCc1OnDhR4eyShDQkZmZmOHr0qEj73r9/H2ZmZkhOTpZyVISPxtEgdVp6ejp2796Nq1evIjk5GQoKCjA2NoazszPc3NzKzcXwv//9D7t27UJ4eDiys7PRpEkT9OjRA1OmTGHnMLl//z4mTJgAZWVl9peijo4OrKysMGHCBFhYWIgcH3+Y3q/PpaWlhe7du2PcuHHo1auXhJ4J+ZaQkIBdu3bh1q1b+PTpE5SVlWFubg5XV1c4Ozuz+3Xp0oVd5nK54HK5UFFRYdf99NNPmDlzZq3GDgCNGzdml/kTadUWIyMjsSe58/T0xMOHD9n3h7KyMvT19eHk5ITp06ezr1Uul4vdu3fj3LlzSEhIQGFhIXR1dTFkyBD8/PPPUFNTY8/34MEDbNu2DQMGDBB6vQcPHuDAgQOwsbERKUYzMzMoKSlBQUGBnW/IwsICLi4uGDJkiFj3W1+lp6dj165duHbtmsDn38iRIzFu3Dj2/3fIkCHsxI+lpaUoKSkReN84OztjxYoVMrkHAACPVCk+Pp5namrKi4+Pl3UoNQKA/ZOH6yYmJvL69+/PmzZtGi8qKorH5XJ5hYWFvBs3bvAGDx7M+/HHH3nFxcXs/pcuXeJZWFjwNm3axEtJSeGVlpby4uPjeStWrOB17tyZd+3aNR6Px+Pdu3ePZ2pqynv37h2Px+PxCgsLedHR0bxNmzbxunTpwjt27JjIMfJfG3fu3OHxeDxecXEx7/3797w9e/bwunXrxgsMDKz2uapy/PhxXseOHUU+v6w8ffqU17NnT56vry8vJiaGV1paysvNzeWdO3eO16dPH56vr6/Q4zZv3szr27dvLUcrnKzeO9U1fvx43vz589nHxcXFvJs3b/K6dOnC27t3L7t+1apVvL59+/IePHjAKyws5BUXF/MiIiJ49vb2vDlz5gicr0+fPryZM2eWu9b79+95ffr04ZmamvLu3bsncoympqa8I0eO8Hg8Ho/L5fISExN5R48e5X377bc8Pz8/se7363NVhf/+//Dhg1jXqG3x8fG8vn37Vvj5N3HiRF5RUVG5444fP84zNTUV+Gys7vUl9b1HVSekzvL390fjxo0RFBSEjh07sjNO9u/fHwcOHMDTp09x6NAhAEBubi6WLFmCsWPHYs6cOWjZsiU4HA4MDQ2xePFizJgxA2lpaUKvo6KiAmNjY8yZMwc+Pj5Yvny5WNPCf01JSQmtW7fGpEmTsHHjRgQGBuLp06fs9rCwMIwZMwbdu3dHr169sHTpUuTk5Ag9l6enJ/z9/bFy5UrY2NigW7duWLhwIQoLCwX2i4iIgLOzM7p06VJuau6MjAz4+vqif//+6Nq1K0aPHo2bN2+y2+Pj4zFjxgzY2Nige/fuGDVqFK5cuVLlffKreapSWlqKhQsXokePHvj999/Rrl07cDgcaGhoYNiwYdi1axfOnDnDzjJKpENJSQn9+vVD27ZtER8fz67/999/MWjQIFhbW0NFRQVKSkro3r07AgMD2ZlK+fr164ebN2+Wex+dPHkSdnZ2NYpPQUEBBgYGcHFxwZ49e3Dq1CmcP3+e3X7v3j24u7vDysoK1tbWmDt3LlJTU4Wea+HChZg6dSp27twJW1tbdOnSBdOnT0dWVpbAfjExMXBzc4OlpSXs7Oxw7do1dltBQQFWrFiBAQMGwNLSEkOHDsWpU6fY7RkZGZg/fz569+6Nbt26wdHREUeOHBHpXs3MzJCQkFDlfv7+/mjatGmFn3+RkZFyM3owJRqkTsrMzMStW7cwefJkofMl8It3T58+DQC4c+cOMjMzMXnyZKHn8/LygouLS5XXdXNzg4qKikS++Ozt7WFiYoLQ/59f/L///oOvry9mzpyJhw8f4p9//sHz58+xcuXKCs9x9uxZtG3bFrdu3UJISAhu3ryJoKAgdjuPx8PRo0exZ88e3L17F3p6evjtt9/Y7V5eXsjKysLx48cRHh4OFxcXzJw5k/2yCQgIgJaWFm7cuIHw8HBMmjQJv/zyCzIzM2t8/wDw4sULvH37FlOnThW63dzcHN988w3OnDkjketJS3Z2NvsnjwoKChAWFob4+HgMHz6cXd+hQwdcvHgRN2/eBJfLZdd37twZ9vb2AufQ19eHlZUV+54DmNffqVOnMGLECInF2rFjR9ja2rKviejoaEyfPh0jR47E3bt3cf78eWRnZ2P+/PkVniMiIgLZ2dm4evUqzp8/j3fv3gm8LwDgwIEDWL9+PcLDw/HNN99g8eLFKC0tBQAsW7YMT58+xf79+xEREYF58+Zh8eLFCA8PBwBs2rQJGRkZuHDhAiIiIrB06VKsXr0a0dHREnkO0tPTcefOnSo//+r6+4aPEg1SJ71//x48Hg/GxsYV7tO+fXvExsYCAN69ewcNDQ2hrf/FoaSkhDZt2iAuLq5G5+EzNjZmzxUcHIxBgwZh4MCBUFRURJs2beDt7Y3Q0FAUFBQIPd7AwAAeHh5QUVGBubk5nJyccOnSJXZ7aWkppkyZgubNm0NTUxMODg548+YNAODly5d4+PAhfH190aJFC6ioqMDDwwNmZmY4fvw4AKa9gaKiIvtr1tnZGREREdDS0pLI/fPvXdT/x7qqcePG7J+8OHfuHLp06YIuXbqgW7du8PHxwcyZM9G9e3d2H39/f3Tu3BnTpk3Dt99+i2nTpmHnzp0Vvv5dXFzY1w7AtHdSUFAQuV2GqIyNjfH+/XsAwJEjR9CxY0e4ublBWVkZOjo68PHxwf3799l9ylJUVMScOXOgqqqK1q1bw93dHZcvXwaPx2P3cXd3h6GhIVRVVeHo6Ij09HSkpaUhMzMToaGhmD17Nlq3bg0lJSUMGjQIAwYMYEstPn/+DAUFBaiqqkJBQQF9+vRBREQETExMJHL/8fHxYn3+1XXUGJTUSfwPBP4vDGG4XC67H4fDgbKyskSuzeVyyzUyra6SkhL2F0lMTAzi4uIEEgWAuceUlBShv1zat28v8Lh169blWsu3bt2aXVZTU0NRURF7PQDlisB5PB77gThnzhwsWLAAffv2Ra9evdCvXz8MHTpUoCEZX3h4OH788Uf2HIBg483IyMiKnoYq/x+p+6Zw/NIwAHBychLr2OHDh2P9+vUAmNdhdHQ0li1bhsjISGzevBkAMzvxjh078OHDB4SHhyMiIgJHjhzBxo0bMWPGjHIzJA8aNAi//vornjx5gm7duuHkyZMYPXq0xP//yr5vnj59KvBaA5hkIiEhAW3atCl3fJs2bQTew61bt0ZBQQEyMjIE1vHxG70WFhYiKSkJpaWlmDFjhsB98Xg8dO3aFQAwffp0zJo1C7a2trCxsYGtrS1GjBgBTU3NcrEkJibC0dFRYJ2joyN77gsXLqBVq1YC2/nb6sv7hhINUie1bdsWHA4Hr1+/Zt/cZb19+5bN+Nu3b4+srCy8f/9e6AePqPLz8xETE4Nx48ZV+xx8PB4PL168YIuV1dTU4O7ujiVLlgjdX1i9bdkPGh6PV+7DRUFBeMGkqqoqAOD27dto2rSp0H169+6NGzdu4P79+/jvv/+wYcMGbN++HUePHi33oWltbc0mE/zeNpUlF8CXROnNmzdo0aKF0H2+/n8kgr5OEr/+NS4uJSUlmJubw8fHBx4eHoiNjUW7du3Y7fr6+hg5ciR7vS1btiAwMBBjxowReD+pqqpi5MiROH78ODp06IArV67g7Nmz1Y6rIlFRUejQoQMA5n1jZ2eHbdu2iXy8sPcNIPheqep9c+TIkQq7j5ubm+PSpUuIiIjA7du3sXfvXmzZsgX//PNPuaShVatWAu8TMzMzXLhwAYaGhhXG37ZtWygoKFT5+Vf2h0hdRVUnpE5q2rQpbG1tsWvXLvYX+tc+fPiA8+fPsx+Mffr0gba2NgIDA4Web82aNfDz86vyuvv27YOysjIGDx5csxsA82s0OTkZo0aNAsB8eLx48UJgn+zs7ErbQ7x7907g8fv370WuHuJ3iYyKihJYzy+WBZi6YBUVFfTt2xe+vr44f/48kpOTRR7Loyrm5uYwNzcXaFfytWfPniE8PJx9jkjtyM/PR2JiIgICAvDx48dy2/mNO78uAeBzcXFBWFgYwsLC0K1bN+jr60s0tocPH+Lhw4cYM2YMAOZ1/OrVK4HkobCwECkpKRWeIzExESUlJezj+Ph4aGhoiFQl2Lp1aygqKpZ73yQlJbHnzM7ORmlpKdswNTQ0FGpqauVKK6urSZMmsLOzq/Lzb/To0RK5nrRRokHqrKVLlyI7OxtTp07F8+fPUVpaiqKiIty6dQs//vgjbG1tMX78eADMr57Vq1cjLCwMvr6+SExMBI/HQ2JiIlasWIGQkJBKv8xSUlKwadMmBAUFYcWKFdDW1gbAfBE6OjqK1QslPT0d+/btw9KlS+Hj48P+Wp8wYQIePXqE4OBgFBQUIDU1FQsWLMDcuXMrPFdiYiJCQkJQVFSEFy9e4OzZs+WKYStibGwMW1tbrFmzBnFxceByubh8+TKGDx+OR48eIS8vD4MHD8aePXuQn5+P0tJSPHv2DEVFRWKP21CZNWvW4NWrV5g7dy7evn0LHo+HvLw8hIWFYdasWRg3bpzQsRnqElmN0DlixAj2r6bi4+OxefNmWFpawtzcHDo6Orh79y7mzp2LyMhIFBUVgcvl4s2bN9i0aRNMTEzQuXPncucxNzdH27ZtERQUhO+++07otVJSUsr1gKpKTk4OTpw4gZkzZ8LT0xO2trYAmAbaqamp+OOPP5CTk4OsrCwsX74cP/zwQ4VVC8XFxQgMDERBQQHi4+Px999/i/y+adSoEVxcXLB161ZERUWBy+UiPDwco0ePxvnz58Hj8eDq6or169ezY6q8efMGWVlZEi1h8Pf3R1FRUYWff3369IG7u7vEridNVHXSQMlD3Z6RkRFOnDiBwMBAzJgxAxkZGSgtLYWZmRnc3NwwYcIEgfuws7PD0aNHsWPHDowdOxafP39G8+bNYWNjg2PHjpUrnh8+fDg4HA47WJC1tTWCg4PZgb0A5pdfbGysQIt8YaZNm8aeS11dHV27dkVQUBD69OnD7tO9e3e2auL3339HkyZNYGtrC19f3wrPa29vj+joaPTt2xdFRUUYNmwYpk2bJvJzuG7dOqxatQqurq4oLi6GkZER1qxZAysrKwDA9u3bsX79egQGBoLD4aBNmzZYs2aN0CGzv2ZoaIhXr16JFIO5uTlOnjyJHTt2YOrUqfj06RNUVFTQsWNH+Pn5YdiwYSLfT13A4XDg7++PgIAAgfXz58/Hxo0bAQDr168v1yti2rRp+OuvvwAAO3bsKPf/6O7ujr///hsA03DY3d1doI2GuM6dO8f2nuJwONDW1oa9vT28vb3ZrpKHDx/Gtm3bMH/+fKSmpqKkpAS6urqws7PDunXrKmyr5Orqio0bN2LgwIFCtxcXFyM2NrbCRs58AQEB+PXXXwEwg4p16tQJAQEBAq8JQ0ND7NixA5s2bcK+ffugoaGBnj174q+//qqw+sPc3Bzq6uoYOHAgsrOzYWtri8WLF1f5nPH5+flBSUkJU6ZMQW5uLgwMDPDzzz+zJahbt27FypUrYWdnh9LSUujp6cHb2xv9+/ev8tyivm/09PRw/Phx/PXXX5g/fz47YFeHDh0wYcIEfP/99xXef51T45E4GoD6OGCXrP5q4tChQ7zOnTvz0tLSJPSMiGbq1Km8rKysWr0mj1d+0CUiO/7+/gKvY39//3L7zJs3j92+fv36ctunTp3Kbt+xY0e57ePGjWO3BwcHS+M2apW/vz8vKiqq1q/r6+vLc3Nzq/Xr1jc0YBdpkEaMGIEmTZpg+fLlyMnJEaiDlZa0tDTk5+ejSZMmUr8WqbsCAgLg7+8v6zDkRlFREZ49e1ZlyRhpGDg8Xg2aMjcQ/Bb2V69erbSlMJG+Z8+ewd/fH9HR0ejfvz+2bNki65CkxtPTE7q6umwXRUJI1RYuXIi4uDi2GopUjyS/96iNBpErlpaWOHnypKzDqBXyMrwwIXXJ77//LusQSBlUdUIIIYQQqaFEgxBCCCFSQ4kGIYQQQqSGEg1CCCGESA0lGoQQQgiRGko0CCGEECI1lGgQQgghRGoa7DgaiYmJ+P333xEeHg4AsLGxwaJFi6CrqyvjyKQs/ZFsrqvdUzbXJURSbjjJ5rp21Z/vhJC6oMGWaMyYMQOqqqq4cuUKzp07h8zMTCxbtkzWYZEy0tPTsW7dOjg6OqJbt27o0aMHXF1dcejQIaFDkP/vf//D3LlzYWtrC0tLS9ja2uLnn3/Gs2fP2H3u378PMzMzWFhYoEuXLujSpQsGDBgAHx8fPH/+vFpx3rt3Dz/99BN69erFXnfKlCm4efNmte+9upKTk3Hs2LEKt0+ePBlmZma1GBGpbZ6enujYsSP7+u7RoweGDx+O7du34+vBoLlcLnbu3AlnZ2f07NkTFhYWcHBwwNq1awUmRPP09ISZmRmuXbtW4fXMzMxw//59seL83//+h/nz56Nv376wtLREr169MH78eJw5c6Z6N14Dubm52Lt3b4Xbly1bBjMzMyQkJNRiVPVDg0w0srOzYWFhgV9++QWamppo3rw5xo4dy5ZukLohKSkJY8aMQXR0NDZt2oSIiAjcu3cPXl5eOHjwIKZPny6QbFy+fBlubm7srK9Pnz5FSEgIdHV14e7ujuvXrwuc/9y5c4iMjMSjR4/w119/wcDAAO7u7jh+/LhYcfJj6dWrF86dO4enT5/i+PHj6NmzJ2bNmoUjR45I5PkQ1eXLlyu8h6NHj+Lp06e1Gg+RjeHDhyMyMhKRkZF48OABfH19ERQUhP3797P7rF27FocOHcKSJUtw9+5dPHnyBOvXr8eFCxfg5+cncD4dHR2hr6v4+HjExsaKHd/ly5cxbtw4GBkZ4dixY3j27BnOnTuHYcOGwd/fH5s2bRL/pmvg/v37FSYad+7cwfnz52s1nvqkziQaRUVFWLt2LczNzeHp6VnhfpmZmVixYgXs7e1hYWHBTv/78eNHka/VpEkTrF69WqCa5MOHD/W/2kTO+Pv7o3HjxggKCkLHjh3Zqa379++PAwcO4OnTpzh06BAA5tfIkiVLMHbsWMyZMwctW7YEh8OBoaEhFi9ejBkzZiAtLU3odVRUVGBsbIw5c+bAx8cHy5cvR1JSkkgxJiYmYs2aNfjll18wYcIEaGtrg8PhQFdXFz/99BPWrFmD5s2bs/tzuVxs2bIFQ4YMQdeuXeHg4IBdu3YJnPPUqVNwcnJC9+7dYWtri2XLlrG/LhMSEmBmZoZbt25h0qRJ6N69O+zt7dkvgA0bNmDVqlV4/PgxunTpIlCS8+HDB6xbtw4zZswQ/T+B1AtKSkro168f2rZti/j4eHb9v//+i0GDBsHa2hoqKipQUlJC9+7dERgYyE6JztevXz/cvHmz3Pvo5MmTsLOzEyse/vvVw8MDP//8M/vZq62tDXd3d2zevBnt2rUTOCY4OBhOTk7o1q0b+vXrh3Xr1gn80Lh58yZcXFzQs2dPfPvtt5g7dy7S09PZ7WZmZggNDcXPP/+Mnj17wtbWFtu3bwcA/P333/Dy8kJKSgq6dOmCsLAw9ricnBwsWbIEs2bNEuseyRd1ItGIiYmBm5sb/v77b1Q2x1tBQQE8PT0REhKCwYMHY/Xq1XBzc8P58+cxbtw4ZGVlVfv627Ztw8yZM6t7C0TCMjMzcevWLUyePBmKiorltuvq6mLIkCE4ffo0AOYXR2ZmJiZPniz0fF5eXnBxcanyum5ublBRUcHFixdFivPSpUtQVlbG999/L3T78OHD4eDgwD7esmULTp06hc2bNyMiIgJr1qzBtm3bcOrUKQBAZGQkfH19MW/ePDx+/BiHDx/GtWvXsHPnToHz/vnnn/D19cXDhw/h7OyMgIAAZGRkYP78+XB2dkb37t0RGRkJS0tL9pglS5bAxcUFXbp0EeneSP1RUFCAsLAwxMfHY/jw4ez6Dh064OLFi7h58ya4XC67vnPnzrC3txc4h76+PqysrNj3HADweDycOnUKI0aMECse/vt14sSJQrf37dsXo0aNYh8fO3YMmzdvhr+/PyIiIrBz506cP3+eTRQ+fvyIWbNmYfTo0QgPD0doaCiio6OxZs0agfNu2bIFEyZMwIMHD/Dzzz9j06ZNeP36NcaNG4effvoJurq6iIyMxNChQ9lj1qxZA0tLSwwaNEiseyRfyDzRyMrKwpgxY8Dlcqssst6/fz9ev34NPz8/+Pn5wcnJCV5eXli3bh0SEhIQFBQEALh+/TrMzMyE/gUGBgqcMzIyEuPHj8ekSZPg5CSjxl6knPfv34PH48HY2LjCfdq3b88W2b579w4aGhowMDCo0XWVlJTQpk0bxMXFibT/u3fvYGhoCGVl5Sr3LS0txeHDhzF16lSYmZlBUVERVlZWcHV1ZatXLCwscPfuXfZDvk2bNujZs2e56o7Ro0fD3NwcioqKGDFiBIqKiiotvj5y5AiSkpIwe/Zske6LyL9z586xbTS6desGHx8fzJw5E927d2f38ff3R+fOnTFt2jR8++23mDZtGnbu3Fnh69/FxUXgc/r+/ftQUFCAjY2NWLHx36+iliIfOnQI33//PaysrKCgoABzc3P8+OOPOHr0KACgZcuWuHXrFtzc3KCgoAAdHR307du33PvGwcEBVlZWUFRUZD/vX716VeF1b9++jStXrsDf31+s+yOCZN7rpLi4GM7Ozli0aBFUVVUr3ffUqVPQ0NCAq6urwHoHBwfo6enhzJkzWLhwIezt7St98fDdunULc+bMwfz58+Hu7l6j+yCSxS/ZKi0trXAfLpfL7sfhcET6shcFl8uFkpJobw0Oh1OuUeqzZ8/g4eEhcL7Lly9DVVUVmZmZ+O2337BixQp2O4/Hg46ODgDmfg8cOICzZ8/i48eP4PF4KCkpgZWVlcA1jIyM2GU1NTUAEGi897WkpCSsW7cOO3furPI9RuqP4cOHY/369QCAkpISREdHY9myZYiMjMTmzZsBAM2bN8eOHTvw4cMHhIeHIyIiAkeOHMHGjRsxY8YMzJkzR+CcgwYNwq+//oonT56gW7duOHnyJEaPHg0OhyNWbBwOBzweDzwejz3206dPAqUoXC4X+/btwzfffIOYmBi8efNGoA0F/71fVFQEFRUVnD59mk2ouVwuuFwu9PT0BK779ftGXV0dQMXvG36VydKlS6GtrY28vDyx7pF8IfMSjRYtWmD58uVVfgDm5OQgJiYGnTp1goqKisA2DocDS0tLpKeni9wi+OnTp5g7dy7WrFlDSUYd1LZtW3A4HLx+/brCfd6+fcuWeLRv3x5ZWVl4//59ja6bn5+PmJgYmJiYiLR/hw4dkJCQgPz8fHadpaUl2wgvLCyMTYj4CcGmTZvY7ZGRkXj+/DnbUHXbtm04cOAA/P398ejRo3LFuHwKCqK/dflVJl//kiUNi5KSEszNzeHj44OLFy+WK/3S19fHyJEjERAQgCtXrsDLywvbtm0r935SVVXFyJEjcfz4ceTm5uLKlSsYM2aM2PF06NAB+fn5Au1FWrRoIfC++PqHhJqaGn755Zdy75vnz59DRUUFJ0+exNq1a/HTTz/h/v37iIyMFFqNKs775vfff4elpSWGDRsm9v0RQTJPNESVmJgIAOUyVD59fX0AEHjhVqSkpASLFi2Ct7c3Bg4cKLkgicQ0bdoUtra22LVrF4qKispt//DhA86fP882WOvTpw+0tbXLVY3xrVmzplwremH27dsHZWVlDB48WKQ4hwwZAgUFhXINOvm+bnOkqamJFi1aICoqSmCflJQU9h4fP36Mb775Bn379oWysjJKS0ur3eUWYN43d+7cwbFjx2BjYwMbGxu2LZKNjQ3OnTtX7XMT+ZWfn4/ExEQEBAQIbUjPb9yZkZFRbpuLiwvCwsIQFhaGbt26sZ+94ujduzdatmzJVneXVbYks23btuXeN2lpacjNzQXAvG+MjY3h5OTE/mitae+qo0eP4s6dO+z7hp9QjRkzBn/99VeNzt3QyE2iwX9B8X8VlsUvBuPvV5knT54gOjoa69evZ+sw+X/8hIbI3tKlS5GdnY2pU6fi+fPnKC0tRVFREW7duoUff/wRtra2GD9+PADmdbF69WqEhYXB19cXiYmJ4PF4SExMxIoVKxASEiLQuKyslJQUbNq0CUFBQVixYgW0tbUBMNUgjo6OFfZC4ZfI7dixA4GBgUhJSQHAlMBdvHgRM2fOhLGxMbS0tAAAP/zwA4KDg3H37l1wuVy8fPkS7u7u2L17NwCmTUZMTAwyMjLw6dMnLF++HI0bN8bHjx+FjhsijLq6OlJSUpCZmYlmzZrh5s2bCA0NxenTp3H69Gm22ub06dMYMGCASOck8i8+Ph6bN2+GpaUlzM3NoaOjg7t372Lu3LmIjIxEUVERuFwu3rx5g02bNsHExASdO3cudx5zc3O0bdsWQUFB+O6774ReKyUlBY6Ojnj8+LHQ7SoqKli7di3CwsKwfPly9gdifn4+bt26BU9PT7Rs2RKtWrUCwLxvzp8/j7CwMBQXFyM+Ph7Tpk3D6tWrATDvm+TkZCQmJiIrKwtbtmxBXl4eMjMzRa7yUFdXR3Z2NlJSUpCXl4ebN2/i3Llz7PuG3yB7586dGDdunEjnJAyZt9GQBSsrK5HacNRLcjRCJ388jMDAQMyYMQMZGRkoLS2FmZkZ3NzcMGHCBIG6YTs7Oxw9ehQ7duzA2LFj8fnzZzRv3hw2NjY4duxYuYalw4cPZ+uKGzduDGtrawQHBwv01MjPz0dsbKxAi/yyRo0ahXbt2mH37t0YPXo0Pn/+DA0NDTZOFxcX9lfW5MmTkZ+fDz8/P6SlpaFly5YYPXo0pk+fDoAZSO7t27ewt7dH8+bN4e3tDWdnZ8ycORNOTk4i/ZJydnbG5cuX0b9/f2zYsKFcqR0/iaqodJBUQM5G6Dx37hzbe4rD4UBbWxv29vbw9vZmu4ofPnwY27Ztw/z585GamoqSkhLo6urCzs4O69atq7CtkqurKzZu3FhhiXBxcTFiY2MrbP8AAL169cKJEyewc+dOjB8/HhkZGVBTU0P79u0xYMAAjBs3Do0bNwbAvFfT0tKwadMm+Pj4QFtbG4MGDcKCBQsAAOPGjcPjx48xYsQIaGpq4ocffsCGDRvwww8/wN7eXqSB8wYPHox//vkHDg4OWLhwIfsjho+f6Ldo0QKamppVno98hVfHmJqa8saPH19u/atXr3impqa8uXPnCj1uxYoVPFNTU95///0n8Zji4+N5pqamvPj4eImfm4ju0KFDvM6dO/PS0tJq9bpTp07lZWVl1eo1CZF3/v7+vKioKFmHQapJkt97clN1YmhoCA6Hg+TkZKHb+UXbX7cqJvXLiBEj0KRJEyxfvhw5OTkiVyXURFpaGvLz89GkSROpX4uQ+qKoqAjPnj2DqamprEMhdYDcJBr8ouioqCgUFhYKbONyuXj8+DH09fVrPI4CqbuaNm2K7du34/379+jVq1e5rnfS0Lx5cxw8eFDq1yGkPlFRUcGJEyeEDrZHGh65STQAprVzfn4+QkJCBNafOXMGaWlpIo38SOSbpaUlTp48icjISGzZskXW4RBCCKmCzBuDRkdHIzo6WmBdeno6Lly4wD7u378/1NXV4ebmhtDQUKxduxZJSUmwsLBAdHQ09u7dC1NT0wqHnyaEEEKIbMg80QgLCyv3yzQ6OlpgqOSrV6+ywzzv2bMHgYGBuHTpEoKDg6GtrQ1XV1d4e3uzXVwJIYQQUjdweLxKZjEjAJgZMx0cHNiEhxBCCKnPJPm9J1dtNAghhBAiXyjRIIQQQojUiJxobN26FR8+fJBmLIQQQgipZ0RONAIDA+Hg4IBJkybhzJkz5cayIIQQQggpS+REY+PGjRgwYAAiIiLg4+ODPn36YOnSpRVOmkMIIYQQInL31mHDhmHYsGHIy8vD1atXERYWhtOnT+PYsWNo06YNvvvuOzg7O0NXV1ea8RJCCCFEjojdGFRDQwNOTk4ICgrCf//9h9WrV6NNmzbYvHkzBgwYgClTpuD69evSiJUQQgghcqZGA3ZpamrCwcEBJSUlyMzMRGRkJG7fvo07d+7A2NgYy5cvR8+e8jMtOSGEEEIkq1qJRklJCa5fv46TJ0/i1q1bKCkpQbNmzTBp0iSMGTMGr169QmBgICZMmIDNmzfDwcFB0nETQgghRA6IlWhERUXhxIkTOHfuHDIzM8HhcGBrawtXV1fY29tDSYk5XYcOHWBnZwcPDw9s2LCBEg1CCCGkgRI50Rg5ciTevHkDHo8HQ0NDeHp64rvvvquw8aempibGjx+PgIAAScVKCCGEEDkjcqLx7t07DBs2DC4uLujVq5dIx5iZmdHU7YQQQkgDJnKiMW/ePAwZMgT6+voV7nPnzh28efMGEydOBABYWlrC0tKyxkESQmrZDafqHWcXKtk4CCFyT+TurWvWrMHz588r3YffCJQQQgghBKiiROPBgwd48OABAIDH4+H8+fN49eqV0H2LiooQGhoKBQWap42QusTOzg43b94EAFy/fh12dnYC23v27ImIiAgAwMOHD+WqS7pTNQteACCUCl8IqRWVJhrJyck4ceIEkpKSwOFwEBYWVunJOBwOfvrpJ4kGSAghhBD5VWmiMXLkSIwcORJpaWno06cPvL29YW1tLXRfBQUFGBoaQk9PTyqBEkIIIUT+iNQYtHnz5vDy8oKjoyOMjY2lHRMhRIJu3LhR6fZHjx5J8GLUiJQQIkjkXideXl7SjIMQUkeYmpoC+UkAgNcH7WUcDSFE3lWYaDg4OCAgIAB9+/ZlH4uCw+HgypUrkomOEFJtoV+1dnQSo9XkmzdvpBEOIaSBqjDR4PF4lT4W9ThCiGyMHDmSXab3JSFEVipMNK5du1bpY0JI/fTq1Svg/nRZh0EIqSdqNE08IaTuGjFiRLWOMzU1BZI0JRwNIaShEivR+PTpE/7991+kpqaiuLhY6D4cDgezZs2SSHCEkOoLDQ390gukur1BCCGkhkRONO7cuQMvLy8UFBRUWt9LiQYhRFw1GeGTEFK3iZxobNiwAQDg7e2NLl26QFVVVWpBEUIIIaR+EDnRePPmDaZPn46ZM2dKMx5CiIwZuFxml5OODZJhJISQ+kDkRKNRo0YwNDSUZiyEEAk6fPgwEJUIAHAf2Erk4z6kFUorJEJIAyRyomFjY4NHjx5h1KhRUgyHkIZjw4YNWLBgAQBg3rx5bPUkX0BAAJYvXw4A8Pf3R0BAgFjn9/DwYJfFSTRkYWm/6jXS+O1fGrqckLpO5DndFy1ahAcPHmDv3r0V9jghhMi/xKMD2T9CCKkpkUs0/Pz80KxZM6xduxYbNmyArq6u0AahHA4H586dk2iQhBDxjRs3Dki5KfZxBi3UpBANIaSh4vBEHJvY3Nxc5JO+fPmy2gHVRQkJCXBwcMDVq1epnQqRL3IyfsaD8OodV5Oqk1CqdSGkQpL83hO5RKO+JQ+EyJt3797Bzs4OcXFxWL9+PebPny+wfdq0afjrr78AADt27MC0adNkEabcqMnYHZSkECI6GoJcHqQ/Em0/7Z7SjYNIxNcJwM6dO0U6hp9U9OzZE3FxcVKJ62vvkvNgN+cu4lLykX3OEY01BD8qOPZn2WXe9S9DnVe3ZIIQUn9VmmikpaVVejCHw0HTpk2hqKgo0aDqOw6HU63jeGkPKz5nc6vqnZNm9ax1/FIHQPREY+PGjdIKp5zGw8IAADn53Fq7JiGk/qo00ejTp0+VX4ocDgedO3fGxIkTMXz4cIkGRwRVN5kg8k+UhHDnzp0iJy6VqW6C8TmvBLkFzHIjNSosJYQwKv00sLa2rvIEmZmZ+N///ocFCxbgf//7H3x8fCQWHCH10Y4dO2QdQqW+rgoRZ58mwy+wy/fXVm/mWEJI/VNponHw4EGRTvLx40cEBARg79696Nu3L3r16iWR4BqCyqpDgOqVYkjjnERyqJEmIaQhkUj5ZsuWLbF582Y4OzsjODiYEg0JqippqKnqthcRBbX/aFg01RVRSs06CCFlSKwiVUlJCUOHDsU///wjqVMSOSdOEkNJifz7fH4o9TohhJQj0RZbhoaGyMjIkOQpCSF8cjL4FgCkZhVgxMorAIAWjVVxbqngLLDvU3Pguu4GAKB1i0Y45mNf2yESQmqJRBONzMxMaGhoSPKURAp4aQ/x4oXo+3fsKPq+Db39R2hoKEaOHAkAGDFiBELLjOx0+PBheHh4oFevXmjbti0zwyqROzTYFyGik2iice/ePXTo0EGSpyRiECd5kNZ5o24/FDkxachJyd27d3H37l1KNAgh9Z7EEo3du3fj5s2bYk9lTaomrQRCWuQtXiJ5Ok3VKu3i2kZHk7rAEtJAVJpoiJI05OTk4MmTJ0hMTISVlRXGjh0rqdhIA/JIxFHWe9axUdbt7OzY5Rs3bsDJyanShq3u7u5wd3evhcgIIaRuqDTRCAkJEekk6urq8PDwwPz586XaXZKQuubmTfGnYSeCXiRksssdDbVkFgchRDoqTTQOHDhQ6cEcDgeNGzdGu3btoKqqKtHASMNiZSVegvrwoWjdYetaCQgpb+Lm2+wyVacQUv9Ummh88803tRVHrXvy5AnWr1+PqKgoqKurw8bGBn5+ftDR0ZF1aESOXL9+XdYhEEJIndYgZz7KysrCjz/+iNmzZ2Pv3r3Izs7GnDlz4O/vj6CgIFmHR0QgbgmItAYE+7qNBqkes1ZNq33s0n7V72f627/Uz5SQ2tAgE42ioiIsXrwY3333HQCgefPmGDRoEPbt2yfbwBqYqNviDa/eybb63WFFbWwKUHVLbTswu6+sQyCESFGdSTSKiorwxx9/YM+ePbC2tq5wQrfMzExs2bIFV69eRWpqKrS0tNC/f3/Mnj0bLVu2FOlaOjo6bJLB4/EQExODkydP0jT3hBCpo8G+SENTJxKNmJgYLFiwALGxsZUWcRcUFMDT0xOxsbHw8PCAhYUF4uLisHv3bty7dw8nTpxA06aiF8O+fPkS3333HUpLS+Hq6oo5c+ZI4G6ItNSkBETcqhaA5l+pbY/efsLMHfcAAD3aa2PbjN4C229FpWDBPmYyFduOLbFhUv1tQ0ZIfSLzRCMrKwtjxoyBkZERjh8/jqFDh1a47/79+/H69WssW7YMHh4e7Hpzc3PMmjULQUFB8PPzw/Xr1zFjxgyh5/Dy8oK3tzd73PPnzxETE4OAgADMmzcPf/75p2RvsAo0uFXdJUp1y/jxPcEfdf+ROPUzhBDSQMg80SguLoazszMWLVpUZRfZU6dOQUNDA66urgLrHRwcoKenhzNnzmDhwoWwt7fHq1evRLo+h8OBsbEx5s2bBzc3N6SmplLPEyKyly8jZB0CIYTUaWIlGgUFBTh06BBu3LiBd+/eIS8vD40aNYKxsTGGDBkCV1dXKCmJl7u0aNECy5cvr3K/nJwcxMTEwMrKCioqKgLbOBwOLC0tcenSJSQkJKB169aVnissLAx//fUXTpw4wa5TUFAAALHjJ3WXuFUtgGB1i0aBeCUU1OC0Znoat6h0HI2+nXRpnA1C5JDI36oZGRnw8PBATEwMAEBLSwtNmjRBfn4+7t27h/v37+PMmTPYu3cv1NTUJB5oYmIiAEBPT0/odn19fQBAfHx8lYlGjx49EBcXh61bt2Ly5MnIzc1FYGAgevTogWbNmkk2cFIvvHsfh2HuTANiI8M2CAthktSjuw6iQFWM6W2JxFx8nMguD+neSoaREEIqI3KisXXrVrx//x6//PILxowZI/CF/OnTJ4SEhCAoKAg7d+7Ezz//LPFAc3NzAaDCJEZdXV1gv8ro6upiz549WL16NXbs2AFNTU18++23WLlypeQCJnKvou60cQnvhW4TdbRSIhnL/n7MLpdNNE7ei8PvJyIBAM7ftMEiF0uB7cE33+LsOaaB8IgR9P9GiDSJnGhcv34dbm5umDx5crltLVq0gJeXFz5+/Ijz589LJdGQtK5du4o8lwshkladdqNU2yJcFyPxSyE/ZORLIRLpq0nXWIC6xxLZEDnR+PjxIywsLCrdp1u3bjh16lRNYxJKU1MTAJCfL/wDIi8vT2A/QmpbdbrQUilIzek3Uxf7mKP/vZN8IIQQoURONNTU1JCenl7pPp8/fy7XUFNSDA0NweFwkJycLHR7UlISAMDIyEgq1ycNw41TYexyyxZV9z6qyWilQDXH97gu/QaRD8Klfokaq6xh6OhvjTD624o/C9z7tse9xFFSiIoQUpbIiUanTp1w4sQJjB07VmipwefPn3Hs2LEqSz2qS0NDA2ZmZoiKikJhYaFAV1gul4vHjx9DX18fBgYGUrk+aRhESS5kLTNL9H21qj+NSL0226kTsv/dIOswCGkQRE40Jk+ejOnTp8PR0RHDhw+HiYkJNDQ0kJeXh1evXiEsLAwZGRlYtGiR1IJ1cXHBihUrEBISgh9++IFdf+bMGaSlpbEDcRFSW2rahZYQQuo7kRONfv36YdWqVVi9ejX2798PDudLkS+Px4O2tjbWr1+PXr16iRVAdHQ0oqOjBdalp6fjwoUL7OP+/ftDXV0dbm5uCA0Nxdq1a5GUlAQLCwtER0dj7969MDU1FdpQlZC6Jur2Q+Spide0szpVLIQQUheINTrV6NGjMWzYMDx48AAxMTHIy8uDhoYGTExMYG1tXa32GWFhYdiyZYvAuujoaMyePZt9fPXqVRgaGkJZWRl79uxBYGAgLl26hODgYGhra8PV1RXe3t5sF1dC6jpxBwOrLnGqWQghRBo4PBFnjgoPD4eJiUmlA1o9e/YMiYmJlc5XIo8SEhLg4ODAJjw19XVpUHWK3on0WA3qxy4/vPyvDCMRVNPqloxT4jUgff26Rper8/669Ao349wBAGZmAbINphZR91YiKkl+7ymIuuOECRMQHl55U/SIiAgEBATUKCBCZCkvP4/9I/XXritv8ObNcrx5U/X0B4SQmqm06uTly5d4+fIlAKYdxoMHD9jxKsoqLCzE8ePHUVRUJPkoCSGEECKXKk00Hj16hNWrV6OkpAQcDgfBwcEV7suvgfnuu+8kGyEhtSj80k1ZhyBU1O2HMPw4V6xjmoy5JaVo5N+UgR3YqhNCiHRVmmh4eHjA1dUVL1++xNixY+Hh4VHhOBkKCgowNDRET5qWksixRhqNZB2CXCsurt5xysqSjaMqUwebIfnfgNq9KCENVJW9TlRUVGBpaYnRo0djzJgx6Ny5c23ERQiRoeomDIQQUpbI3VtXr14tzTgIIVLUbNTZah13e6X0hzsnhNRvYo2jQQhpWGwX198EJS/vHe7etUN+fhw6dlwPY+P5AtufPZuG9+//AgB06bIDRkbTZBEmIXKPEg1CvvL1eBU0xkn99WdoFN4lBqJp057Iz4+TdTiE1GuUaBBST2Wf6Fut4xpCb5XDt2IAbKzWsRERX3qr9OhxWEIREVJ/UaJBiAyI21W1NmWf6ItnL8Sf9rW61SxA7fdWub92BH77t/JhMi0td8LScme59UlJf7PLlGgQUjVKNAj5ClWXMDi8mnU7qc9tOwgh4qkw0ahquPHKWFtbV/tYQgipDUv7OVXruG8ad8fJlwskHA0h9VeFiYanp6fA5F/iePHiRbUDIoSQumxI91Z48JlGFSVEVBUmGrNmzSqXaISHh+Px48fo1asX2rdvDzU1NeTl5eH169cIDw9Hnz590K9fvwrOSAipz+6sGFKt4/osuSjhSGrPgwdO+PiRqSaytj4DXV3BUpL//rNDejozrP23315HixZ2tR0iITJXYaLh7e0t8PjGjRs4ceIEzp49izZt2pTbPyYmBj/++CO+//57yUdJSC3Jzctll0UdjrwuN+wkhBBZE7kx6NatWzF27FihSQYAtG/fHmPHjsXWrVsxYMAAiQVISG2yHtyfXaaGoXVfTYZKr+35VQhpqERONN68eQN398rrJfX09PD27dsaB0WIOFx+HI+o1y8BAEd3HURn844yjog0FN98U3kX2d69b9ROIITUYSInGqqqqggPD8fo0aMr3OfJkydQVVWVSGCElPUg4ksJwzc9rCrZs/o01DWkcl5CAODWrS+zW/ft+0iGkRBSe0RONPr374+TJ0+Cx+Nh2LBhaNu2LdTU1FBYWIj4+HhcuHABx44dg6OjozTjJQ3YxJ9nsMuiVGsMdRuDxOQklJSU4Pzh42jbxkhge39nR6SmfQIA3DgVhpYtdPDw8r+SDVqCIqPET4K6dMqTQiSkurKyImR6fafq9ehlhVZegEOIUCInGgsXLsTbt29x8uRJnDp1qtx2Ho8HY2Nj+Pr6SjI+Qqp0bM8hoevjEt7XciSEEELKEjnR0NbWxrFjx3Dt2jXcu3cP8fHxKCgogJqaGgwMDGBtbY1BgwZBSYkGGyXSYd2th6xDIFJWk2HMq+v+2tobjdTWlhoYk4ZHrKyAw+HAwcEBDg4O0oqHkArt31J+3onKVFW9cvP0hZqEIxeqU93S0NSk54q4tLR6Vr0TIfWM2MUP6enpiIiIwIcPHzBw4EDo6+sDAHJzc9GokWjjDhBCCCGkYRAr0di+fTu2bt2KkpISAIC5uTn09fXB5XIxZMgQfP/99+UG+iKEkMpUd0RRAOBxqjcYhiyqaAhpqERONEJDQ/HHH3+ga9euGDRoENavX89uy8vLQ+fOnREUFIRWrVphzJgxUgmWEEK+VtNZZiV1jtpw9iwzJUTTpj3KdY399OkG7t2zBwBoa/cvN35HSkoowsNHso9HjOBJN1hCvqIg6o6HDx9Gjx49EBISUm6Y8caNG2PHjh2wtrbG33//LfEgCQGA67f/Zf8IkUfXr5uyf9WVl/dOcgERUgtELtF4/fo15s2bV+mMriNGjMDq1aslEhghZc1aOI9dpuHBiaRUb1I35nNQ3JKB3Nw31biWoOLi9Gof27ixBdTV29Y4BkLEIXKiUVxcXGVjTyUlJfB4VCRHCGkY+NUZtaGyrrEtWthVmvTo6jrVmeoSGjSs4RE50Wjbti3u37+PUaNGCd3O4/Fw+vTpCiddI6Sm7Hr3lXUIhMhMXegaW9MkgTRMIicazs7O2LhxI0xNTTFkCNNKnMfjsd1d9+/fj4cPH+KXX36RWrCkYQtau0nWIZB6oiY9XapX1UJIwyVyojFp0iQ8ffoUa9aswdq1awEAEydOZKtKeDweBg0ahIkTJ0olUEIIqQvurBiChRerN9jbrVu1V9VCSF0hcqKhoKCAzZs34+rVqzh//jxiYmLYQbpMTEwwdOhQDBgwQJqxEkJInaDIzZR1CGJLTDzMLrdq5S7DSEhDI/bIoDQEOSGE1FxWlnj7S7I0hBINUptEHkcjKSkJ+fn5le7z9u1bbNmypcZBESLM2UsX2D9CCCHyQeQSjQEDBkBfXx+bNm1Ct27dhO4THR2NrVu3wsvLS1LxEcLy+XUJuzxisKMMIyGEECIqsapOUlNTMX78ePj4+GDChAnSiokQQkgl+vYVf0wMaohKZEWsRGPx4sW4dOkSVq9ejYiICKxcuZJmbCW1ZvjA6ndJJESSVg5zq9Zxff77siyPDUoJqQ6xEo3mzZtjz5492Lx5M7Zv345Xr14hMDAQJiYm0oqPENa6gJWyDoEQQoiYRG4MysfhcDB79mzs2LED6enpcHV1xZkzZ6QRGyGEEELknNjdW/n69euHkydPYs6cOfD19cWjR49gZWUlydgIIaTeuvFfM1mHQEitqHaiAQAGBgYIDg7G77//juDgYJw/f15ScRFCCCGkHhC56sTAwADq6url1isrK2Pp0qXYuHEjSkpKJBocaTjWBG5CJ1srdLK1wt6/D5Xb7r9mJTrZWmHTjq04cvqEDCIkpO5Q5GaK/UeIrIhconHt2rVKtw8bNgzdu3dHfHx8jYMipCJ/HdwLABjrPEbGkRBSM3a9M8Tan6paiLyqUdVJWfr6+tDX15fkKQmpdYYf58o6BELK6WUVJbFznT1bszE1RowQfxwP0nBVmGg4ODggICAAffv2ZR+LgsPh4MqVK5KJjjQYvt5z4etd8Rf8ct/FWO67uBYjIkTyLi11wK+X9lbrWFUV+hFH5FOFiQZ/+veKHot6HCEV2bJ7B7vsNXm6DCMhpHY0UlWCkmJj9vHX1SFlq1JKuJ9x+34bAICigib6fkvV0kQ+VZholG2TUVUbDULEFbT3L3aZEg1CpMuudwa4ilrVPp6GMCfVJfaAXfXRqlWrYGZmJuswCCFEauITtyAhYQMSEjbIOhTSwFRYonHq1Klqn3TUqFHVPra2vXjxAqdPn5Z1GA3SzElTZR0CITJVWc8TJcXGYvdMqczbuKXssqHh/BqdKyur+sc2bVqjSxM5VGGisXDhQnA44hWV8Xg8cDgcuUk0SktL4e/vj0mTJmHTpk2yDqfBoeoSQmSDxtUgtanCRGP16tW1GQeKiorwxx9/YM+ePbC2tsbBgweF7peZmYktW7bg6tWrSE1NhZaWFvr374/Zs2ejZcuWYl0zJCQEampqGDFiBCUahJB6zVB/lqxDIA1UhYnG6NGjxT5ZVlYWsrOzxT4uJiYGCxYsQGxsbKW9VgoKCuDp6YnY2Fh4eHjAwsICcXFx2L17N+7du4cTJ06gqYjlcp8+fcLWrVsrTGgIIaQ+MWm3QtYhkAZKogN2Xbt2DevXr8edO3dEPiYrKwtjxoyBkZERjh8/jqFDh1a47/79+/H69WssW7YMHh4e7Hpzc3PMmjULQUFB8PPzw/Xr1zFjxgyh5/Dy8oK3tzdWr14NV1dXtG/fHgkJCaLfJJErNPgWIZIniR4offvSUAgNhViJRkpKCk6ePInExMRy85oUFhbi7t27KC4uFiuA4uJiODs7Y9GiRVBVVa1031OnTkFDQwOurq4C6x0cHKCnp4czZ85g4cKFsLe3x6tXryo8z927dxEZGYlVq1aJFSuRrDWBX6qrKhusixBCiPwSOdGIjY3F999/z1aNcDgcgWoODocDRUVFzJolXj1gixYtsHz58ir3y8nJQUxMDKysrKCioiKwjcPhwNLSEpcuXUJCQgJat25d6bnOnDmDlJQU9OvXD8CXQcZsbGywbNkyDB8+XKx7INWz/59gdpkSDUIIqZ9ETjT+/PNPcDgcLF++HAYGBpg6dSr8/f3RunVrhIeHIzQ0FCtWrEDv3r2lEmhiYiIAQE9PT+h2/hwr8fHxVSYaCxcuxOzZs9nHycnJ+P7773H69GmR23gQQog8iX3/O7vcrs1CsY+XRFdbmhiuYRI50Xj27Bl++OEHfP/99/j8+TMAwMTEBNbW1rC1tcWwYcMwadIkbN26Fd27d5d4oLm5uQAANTU1odv5U9jz96tM06ZNBRIKfjVQRUkMkY5fZs2RdQiENBhxCWvY5bKJRnTsEiR82AoAMDb6Da1beQlsf/V2Dj6k7GcfS3J8D1L/iZxofPz4EUZGRgAABQVmQNGioiJ2u5mZGdzc3PDnn39i3759ko1SygwNDStt00GkY9K48bIOgZAGR1dnnKxDIA2MyIlG48aNkZqaCgBo1KgRVFVV8f79e/Tp04fdp23btlLrLqqpqQkAyM/PF7o9Ly9PYD9CCCHlqam2qfaxbVrNrdHxX6tuzxX+OJI0gaf8EDnRsLa2xoEDB2BhYQErKyu0a9cOwcHBGDZsGFsN8e+//1bZc6S6DA0NweFwkJycLHR7UlISALClLoQQQr6orLrDpN2KSsfZMDP+A2bGf0ghqup79Khmx/fsKZk4SNVETjSmTJmCCRMmYPPmzThw4ABcXV3x22+/YfDgwTA3N8eHDx8QHx+PkSNHSiVQDQ0NmJmZISoqCoWFhQIJDZfLxePHj6Gvrw8DAwOpXJ8QQkjdMbeGHdX+/VcycZCqiZxoWFpaIiQkBK9fvwYAeHh4IDU1FQcPHsT9+/ehqKiIIUOGYNGiRVIL1sXFBStWrEBISAh++OEHdv2ZM2eQlpYGb29vqV2bSJ7/mpXs8nLfxTKMhBBSG2raiJR6rcgnsQbsMjc3h7m5Oft4zpw58Pb2RkZGBrS0tKCkJP5Ao9HR0YiOjhZYl56ejgsXLrCP+/fvD3V1dbi5uSE0NBRr165FUlISLCwsEB0djb1798LU1BSTJ08W+/pEdo6GnmSXKdEghIijpqOTijlnaKWovUjlajwEuaKiIlq0aFHt48PCwrBlyxaBddHR0QLjXFy9ehWGhoZQVlbGnj17EBgYiEuXLiE4OBja2tpwdXWFt7c328WVVE2cobkTWtKEc4Q0dFGvp6Cg8D2yP4ejY4ed0NURHKE58oUb0jIuAgAszA+jhbbgdBKPn49AVvaX6Smoi2zDIVaiERkZiRs3biA1NbXCocY5HI5YQ3t7e3uLVeWhqakJPz8/+Pn5iXwMqZsCfpFeNRshRLI+fjou6xCInBI50QgODsaKFSuqLCISN9EgDddY5zGyDoEQUsvMTLZWu4usJEtBNLW1anT82bMSrHup50RONA4ePAhdXV0sXLgQ5ubmUFZWlmZchBBC6pCqvuS7dAypdHt3i7OSDIfIEZETjQ8fPsDHxweOjo7SjIcQQggh9YiCqDu2bt0aHEk20yWEEEJIvSdyica0adOwc+dOjBgxAk2aNJFmTKSB+CXgS5fWdQErK9mTEFJf5Be8x5P/jUBhYTx6Wl5HY81uAtvvR1ghv+AtAOCb7uHQUDcR2P5feEcUFSfXuL1GVlaNDidiEDnRGDlyJD59+gR7e3vY29ujVatWQttpcDgczJo1S6JBkvrp3JWL7PLfM/NEOoa62hIivx4+tQMAcLk51To+v+C95IKRIEmV9tfX8ThETjQuXLiAjRs3oqSkBGfPVtyohxINQgip2MphbtU+dvH5yhtc1nU5uU9rdPyT/41AUbHw+a7EpcjNlMh5SNVETjT+/PNPNGrUCNOmTaNeJ0Qi1i5bAe1s8Wb7FWegMUJI3SJKdYdNj4cVbissjJdkOKSWiJxoJCYmYt68eZg4caIUwyENyYjBjjD8eLHqHQkhBEAvqyhZh8Cy651R47E4gIYxHofIiYa+vr7UpoAnhBBCqqKqoi/rEEg1iJxo/PTTT9izZw+cnJygqakpzZgIIYSQOo96rohG5ERDUVER7du3h4ODAwYOHAgDAwOoqKgI3Xfq1KkSC5AQQghpCB49qtnxPXtKJg5JEznR+OWXX8DhcMDj8XD8eMWT63A4HEo0iEhm+syFWtH/AABHFnWWcTSEECIe6rkiGpETjVWrVtHIoDIiak8LeRtj4sZ/t2QdAiFEjty615pd7vtt3emBkpF1G0//5wQAaNqkT7l5XT6lh+H5S3cAQPNmQ6qcF6a+ETnRGDOGZtokhBAiO9zS6g30RWRL5ERj7NixmDlzJuzs7KQYDmlItv6+ES2ydsk6DEIIqRM0CmrYSAN1s5GGyIlGamoqUlNTpRkLaWDsbfvB8ONJWYdBCJETtjZ1cwjyZk1tKx2MrIX20BrPzSLPRE405syZgy1btqBDhw7o1q2bFEMihBBCylNSbCzrEFj8eVsAwKrrDYmcs5OtlUTOU9fmTBE50Xj48CGMjIzg7u4OHR0dtGrVCo0aNSq3H4fDwc6dOyUaJCGEEFKX1HTeloZE5ETj6NGj7HJKSgpSUlKE7kc9U2SH5gEhhBBS14icaFy9elWacZAG6AevaVAtfgsAOP+bpYyjIYQQ0fW0vC6R89j1zkDQ2rc1Po+kql2kQeREo1WrVtKMgzRA4U8iZB0CIUSO3PivGbss68aVjTW7yfT68kTkRIPv2bNnuHHjBt69e4e8vDw0atQIxsbGGDx4MExMTKQRI5Fj/3v5Aq5TPAEAnUzNcWzPIRlHRAghdUdS8j4cOc306BzrXD/HqxI50eByufDz80NoaKjQFq2BgYGYNGkSfHx8JBogqb/2bd4Oncytsg6DEEJk5nXMXASsY5YbfKKxf/9+nDlzBnZ2dhgzZgzatWsHNTU15OXl4e3bt/jnn3+wd+9emJiY0CiiRCTf9LCC4UctWYdBCJETsq4uIdUjcqJx+vRpODg4YOvW8r9AzczMMGzYMEyePBkhISGUaDRgQ92+/N+HhZxAZ/OOiLr9UIYREUKI5N2P+NL40qZH9T/j9HV/gO032ZIIqc4SOdGIi4uDh4dHpfs4ODhg3bp1NQ6KyK+4hLo5cl9dExmlIesQCCE1kF9Q854iAGBm/AeW+0rmXHWVgjg7l5SUVLpdSUmpzo1IRgghhBDZEblEo127drh58ybc3d0r3OfatWto27atJOIicur84eOyDoEQUk+VcD+zy7Iejvyb7uEyvb48ETnRGD16NFatWoVp06Zh7NixMDU1hbq6OvLy8vD69WuEhITgv//+w5IlS6QZL6nj2rYxknUIhJB66vb9NuyyrBuGaqjTcA6iEjnR8PT0xMuXL3HixAncunWr3HYejwd3d/cq23EQQgghhBGfuAV7/04DAEwaN17G0UiHyIkGh8PBqlWr4ObmhsuXLyMmJgZ5eXnQ0NCAiYkJhgwZgk6dOkkzVkIIIQ2YooKmrEOQuLdxS7Hu/ztzNvhEg8/S0hKWljQvBSGEkNrV99t4WYdAqkHsRIOQyvR3dmSXb56+IMNICCFEev4L78gu97Z+Ue3zGOrPgkO/TAlEVHdVmmhMmDBB7BNyOBzs37+/2gGRukec6edT0z5JMRJCCKkbioqTJXIek3Yr4Otdv8fRqDTRSEhIEPlEpaWlSE5OBofDqXFQhBBCCKkfKk00rl27JtJJnj17hoCAACQnJ8PCwkIigRH59GrXN/jQYrmswyCE1EOFRR/YZVUVfRlGAvSyipLp9eVJjdpofP78GevXr8exY8fQqFEjLFu2DOPGjZNUbEQO6WurgttCR9ZhEELqobsPv/RslPU4GrJOdORJtRONU6dOYd26dUhLS8OoUaPg4+MDbW1tScZGCCGE1Gux73/Hlt3pAACvydNlHI10iJ1oREdHY/ny5Xj48CFMTEzw559/wsrKquoDCSGEkBpQUdaTdQgSF5ewBkF7meUGn2gUFBQgMDAQ+/fvh4qKCn755Rf88MMPUFRUlGZ8hBBCCICadSMlsiNSonHlyhWsXLkSHz58wODBg7F48WLo6upKOzYihwzc/0Mppx8A4OHlf2UcDSGESMete63Z5ZoMJGZk6Ivhg9IlEVKdVWX31hUrVuDmzZto06YNdu3aBVtb29qKjcihnAIugDxZh0EIIVLFLc2RyHnatVkIr8kNeByN4cOHo6ioCDY2Npg6dSpUVFQQHl711LjW1tYSC5AQQggh8qvSRKOwsBAAcO/ePdy/f7/Kk/F4PHA4HLx4QfVoDVVicC8k6fwu6zAIIfVQXn40uyzradptbd7L9PrypNJEY/Xq1bUVB6knGqsroZFGI1mHQQiphx48/lJaLutxNJQUG8v0+vKk0kRj9OjRtRUHIYQQ0uBExy7BmsBMAICvt+jzSsmTBjt7q5mZGZSVlQXmZhkzZgyWL6fhswkhpC5SVzOWdQgSl/BhK/b/wyxTolEP7d69GzY2NrIOgxBCiAhsejyUdQikGhp0okEkr8mYWwCYkWKjbtOHAiGkfrrxXzN2uSbtRYyNfsOYEWmSCKnOUpB1AHxFRUVYu3YtzM3N4enpWeF+mZmZWLFiBezt7WFhYQFbW1ssXrwYHz9+FPua+/fvh4ODA3r27AkfHx9kZ2fX5BYIIYQQsbRu5YVJ48Zj0rjxsg5FaupEohETEwM3Nzf8/fff4PF4Fe5XUFAAT09PhISEYPDgwVi9ejXc3Nxw/vx5jBs3DllZWSJfs2vXrrCyssK5c+dw4sQJvHr1CsuWLZPE7RBCCCHk/8m86iQrKwtjxoyBkZERjh8/jqFDh1a47/79+/H69WssW7YMHh4e7Hpzc3PMmjULQUFB8PPzw/Xr1zFjxgyh5/Dy8oK3tzeOHDnCrjMyMsK8efMwffp0FBQUQE1NTXI32MBkn+iLhJabZB0GIaQe+pzzhF1urNlNZnEAsu9eK09knmgUFxfD2dkZixYtgqqqaqX7njp1ChoaGnB1dRVY7+DgAD09PZw5cwYLFy6Evb09Xr16JVYchoaG4PF4SE1NRevWras+gBBCSK169MyeXaYvevkh86qTFi1aYPny5VUmGTk5OYiJiUGnTp2goqIisI3D4cDS0hLp6elISEio8ppRUVH4/XfB0Svfvn0LZWVl6OnVv2mICSGE1E2v3s6B/5qV8F+zUtahSI3MSzRElZiYCAAVJgL6+voAgPj4+CpLJJo3b45//vkH2tramDhxIhITE/Hnn39i7NixUFZWlmzghBBCJEKzUVdZhyBxH1L242gos7zcd7Fsg5ESuUk0cnNzAaDC9hPq6uoC+1VGV1cXO3fuxIYNG7Bt2zaoqKhg9OjRmDu3fg6WUps+55cgN4/5P6ChyAkhkmTV9YasQ2CVcD+zyzQceeXkJtGQNGtra4SEhMg6jHqnlcddAP0B0DgahJD66/b9NuxyTdqLmLbfBPfvUiURUp0lN4mGpqYmACA/P1/o9ry8PIH9CCGEkLrOQG8ixjq/lXUYUiU3iYahoSE4HA6Sk5OFbk9KSgLAdFUlsqOppohSTuUNewkhRN4pKtCPWlHJTaKhoaEBMzMzREVFobCwUKCXCpfLxePHj6Gvrw8DAwMZRkmSDvemcTQIIVKRkXWbXW7W1FaGkQB9v42X6fXlicy7t4rDxcUF+fn55dpWnDlzBmlpaXBxcZFRZIQQQqTt6f+c2D8iP2ReohEdHY3o6GiBdenp6bhw4QL7uH///lBXV4ebmxtCQ0Oxdu1aJCUlwcLCAtHR0di7dy9MTU0xefLk2g6fEEIIqbao11PwS0AOAGBdQP0cS0PmiUZYWBi2bNkisC46OhqzZ89mH1+9ehWGhoZQVlbGnj17EBgYiEuXLiE4OBja2tpwdXWFt7c328WVEEJI/dO0SR9ZhyBxHz8dx7krzDIlGlLi7e0Nb29vkffX1NSEn58f/Pz8pBgVqa4P6YX4qMB01WrZQkfG0RBC6pPuFmdlHQKrsOgDu6yqoi/DSOo+mScapH4xm/IAADMxHo2jQQipr+4+7MQu12QcjY4ddmLSuI+SCKnOokSDEEIIkRFdHVeMGEzjaBAiMr1mKuAqNJF1GIQQIlUqyjQBp6go0SAS9Xq3DY2jQQiRik/pYexyC+2hMowE6G39QqbXlyeUaBBCCJELz1+6s8s1aRdBahclGoQQQoiMRL5ww0wfZq6uoLX1szSYEg1CCCFyoXmzIbIOQeLSMi7ixn+yjkK6KNEgEvUmKQ8pBXEAgLZtaII7QojkdOkYUvVOtSQv/8uI1hrqJjKMpO6jRINIVE+vRwC+A0DjaBBC6q8Hj63Z5Zq0F7EwP4yfJgqflby+oESDEEIIkZEW2kNhb0vjaBAiMmN9dZQo0tDjhJD6TV3NWNYhyA1KNIhEPd5qReNoEEKkIiX1KLusq+Mqw0gAmx5UNSwqSjQIIYTIhRdvprHLsk40iOgo0SCEEEJk5PHzEfjBKx8AsH/LThlHIx2UaBBCCJELLVt8J+sQJC4r+w7Cn8g6CumiRINI1OO3n/ExnZkDoLN5RxlHQwipTzqZ7pJ1CKzPOU/Y5caa3WQWhzygRINIVP9fngDwBEDjaBBC6q9Hz+zZ5ZqMo9G1cyjmTk+UREh1FiUahBBCiIw0a2qLb3rQOBqEiKxbe00UKRnKOgxCCJEqzUZdZR2C3KBEg0jUv+u70zgahBCpSErexy4b6E2UWRwAYNX1hkyvL08o0SCEECIXXsfMZZdlnWgQ0VGiQQghhMjIw6d2cPmxEABwbM8h2QYjJZRoEEIIkQv6uj/IOgSJy8l9iqjXso5CuijRIBJ163kmUrWYbq3f9LCScTSEkPrEzPgPWYfAysi6zS43a2orw0jqPko0iEQNXxYJYAYAGkeDEFJ/Pf2fE7tck3E0elpex8KfEyQRUp1FiQYhhBAiI401u6GzeWNZhyFVlGgQibLt3BSFysayDoMQQqSqaZM+sg5BblCiQSTq/G+WNI4GIUQq4hO3sMutW3nJMBKgu8VZmV5fnlCiQQghRC68jVvKLss60SCio0SDEEIIkZH7EVYY6lYMAAgLOSHjaKSDEg1CCCFywVB/lqxDkLj8greIq9+dTijRIJIVFp6GT03/BQDY2/aTcTSEkPrEpN0KWYfA+pQexi630B4qw0jqPko0iER9vzoKwDwANI4GIaT+ev7SnV2uyTga33QPR8Av7yURUp1FiQYhhBAiIxrqJmjbhiPrMKSKEg0iUY5W2ihQ6SzrMAghRKqaNxsi6xDkBiUaRKKOLOpM42gQQqQi9v3v7HK7NgtlGAnQpWOITK8vTyjRIIQQIhfiEtawy7JONIjoKNEghBBCZOS/8I7o71wCALh5+oKMo5EOSjQIIYTIBSNDX1mHIHFFxclITZN1FNJFiQaRqCP/fkR6EyYrHzHYUcbREELqk7pUXZKSepRd1tVxlWEkdR8lGkSipvzxCsASAJRoEELqrxdvprHLNUk0ellFYfWSOEmEVGdRokEIIYTIiKqKPlq2yJN1GFJFiQaRKJe+OshX7SHrMAghRKpatvhO1iHIDUo0iETtmWuOhJYrZR0GIaQeio5dwi7Let6TTqa7ZHp9eUKJBiGEELmQ8GEruyzrRIOIjhINInGGH+fKOgRCCJELt+61htWgUgDAw8v/yjga6aBEgxBCiFwwNvpN1iFIHLc0B3n5so5Cuhp0orF7927s378f2dnZ6NSpE3799VeYmJjIOqxqaTLmFgDgzxkmmDRYX2Dbj5te4titVADArjlmGNuvpdTi2HvpA7tcNg5CCKmJ1q28ZB0CKyl5H7tsoDdRZnHIgwabaISEhOCff/7Brl270KpVK+zYsQPbt2/H+vXrZR1ajbz/WCjT68/eHs0uU6JBCKmvXsd8qSKuSaJha/Mem36LlUBEdVeDTTT++usvzJs3D6ampgCAefPmyTgiydhwIh7+49vKOgxCCCEiUFJsjEYajWQdhlTVmUSjqKgIf/zxB/bs2QNra2scPHhQ6H6ZmZnYsmULrl69itTUVGhpaaF///6YPXs2WrYUrUogJSUFCQkJyMvLg5OTE5KTk9GjRw8sX74cenp6krytWjNxUMVx75lrjj1zzWUeByGE1Bf6uj/IOgS5UScSjZiYGCxYsACxsbHg8XgV7ldQUABPT0/ExsbCw8MDFhYWiIuLw+7du3Hv3j2cOHECTZs2rfJ6ycnJAICzZ89i586dUFZWho+PD+bNm4fDhw9L7L5q0+afOoi879hV/2OXjyzqLLM4akNklIbYx3TpVL9H6SNEXr16O4ddNjP+Q2Zx1IXryxOZJxpZWVkYM2YMjIyMcPz4cQwdOrTCfffv34/Xr19j2bJl8PDwYNebm5tj1qxZCAoKgp+fH65fv44ZM2YIPYeXlxf69u0LAJg8eTL09Zl2BPPmzcN3332H5ORkuS3VENWFh+myDoEQQsT2IWU/u0xf9PJD5olGcXExnJ2dsWjRIqiqqla676lTp6ChoQFXV8EJbBwcHKCnp4czZ85g4cKFsLe3x6tXryo8T0JCAgBAS0uLXdeqVSsAwMePH+t9okEIIaRuuPFfM3SyZZajbj+UbTBSIvNEo0WLFli+fHmV++Xk5CAmJgZWVlZQUVER2MbhcGBpaYlLly4hISEBrVu3rvRcenp6aNy4MaKiomBpaQngS/JhYGBQzTuRH//4dZJ1CIQQIjbT9ptkHQKpBpknGqJKTEwEgApLG/hVIPHx8VUmGkpKShg3bhy2b98Oa2trtGjRAn/88Qfs7OzQokWLcvvn5zOjqTx58oRt31ETampq7PL/Xr+scv/UjKIq9zl+6yO7/F3fyhvF6rVozC4/jq363OIQJ47aEJPCEfuYEnXJPifCVCcuQgqLnsg6BJlqrt2NXZb1c/Ex9Ti73FKn+hOsift9IMp5Hj6seckI/7uO//1XE3KTaOTm5gIQfDK/pq6uLrBfVX7++Wfk5+fD3d0dhYWFsLOzQ0BAgNB979+/DwCYP3++mFEL16ZNG3Z58e+/SuScgDK7dCEqQ0LnlOc4CKmPpss6APL/FL+q6U/LvlTt80jq++Dr83zdhrGm7t+/jw4datbIX24SDUlTVlbGkiVLsGTJkir3/eabbwAAwcHB1H6DEBng8QBOmUKgixeBo0e/PB4yBCjTfIsQUk3Jycnw8PBgv/9qQm4SDU1NTQAVF+Pk5eUJ7CdJGhpMF0k9PT0YGhpK/PyEEOF4PCApCUhNBRQVgebNAR0dQEUFUFcHlL8UoKFlS4DenoRIFv/7ryYUJBBHrTA0NASHw6mwjURSUhIAwMjIqDbDIoRISUkJ8OYNkJwMlJYCxcXM8vPnQEwMkJ0tuH8VndYkZsWKFXB3d6+dixFSD8hNoqGhoQEzMzNERUWhsFBwPg8ul4vHjx9DX1+/QfQaIaS+y8sDXrwAPn8uv43HAzIygLg4ICeHSUCA2ks0OBwOFBTk5qOTEJmTq3eLi4sL8vPzERISIrD+zJkzSEtLg4uLi4wiI4RISloa8PIlUFRF55+iIoDLZZKS7GwgN5cpBZG2xYsX49ChQ9K/ECH1hMzbaERHRyM6OlpgXXp6Oi5cuMA+7t+/P9TV1eHm5obQ0FCsXbsWSUlJsLCwQHR0NPbu3QtTU1M4OjrC09MTDx48wNWrV0VuTzF58mTcvn270kG+CCHSxeMB8fFMe4yyGjcGmjUDPn4ECgqYdV8nIjweU/rx7Bmgrc2015BA1TIhRAJknmiEhYVhy5YtAuuio6Mxe/Zs9jE/aVBWVsaePXsQGBiIS5cuITg4GNra2nB1dUWXLl0wYcIEdnhxUR09ehRPnz6VyL0QQqqnuJhpd5GTU34bv5Enh8M0BM3OZhIOfpUJn4oKk3CkpTF/mprMsVpa5XusVNfbt28xcuRI/Pnnnxg4cKBkTkpIPSfzRMPb2xve3t4i76+pqQk/Pz/4+fkJrD969CiCg4Px4cMHnDp1SqRzffjwAevWrcOMGTOwbt06ccImhEhIbi7w9m35xEFBATAyYkoovtakyZc/FZUvJRtlBgxGTg7zp6IC6OoySUdNlZSUwNTUFNplgyKEVEjmiYak8Oc/+fDhg8jHLFmyBC4uLujSpYu0wiKEVOLTJ+D9e6Yk4msqKoCxceXVH1wu08VVTY1JNirq2V5UxFTJFBXVvPurmZkZTp48WbOTENLAyFVjUEk6cuQIkpKSBKpoCCG1g8djeo3ExZVPMpo0ATp2rLqNBb+tBofD9DixsGCSk8aNhe+fkgJIYDRlQoiY6k2JhjiSkpKwbt067Ny5s8oZYwkhklVUxLTHEDZbgJ4eYGAgWpuKMr3coabGtMfQ0mISio8fgfR0ZgwOvqQkJhmpLjc3N5SUlODYsWPVPwkhDUyDTDT4VSbdu3eXdSiENCg5OUx7jLLdUBUUgLZtmZ4louKXaPB9/ZtBXZ1p39G0KXM9vsxMJsFp1EjcyBk9evQAl8ut3sGENFANLtFITEzEnTt3EBkZiRMnTgBgGngBgI2NDZYtW4bhw4fLMkRC6qWPH4GEhPJVJaqqTCnD/8+LKDJhJRplaWkxScXXpSeJiYCpqXjX4vPx8anegYQ0YA0i0Xj27Bl8fHywZ88e6Onp4ebNmwLbHz9+jDlz5uD06dNo2rSpjKIkpH4qLWUafKalld/WtCnQrh0zj4k4eLzyiUZFtaCtWgGvX395/Pkz00W2SRPxrllaWgoOhwOOpPrKEtJA1JtEY8iQIUhKSgLv/38uOTo6gsPhwNnZGU5OToiNjUVxcTEUFRXLzcDK76pGM7MSIllFRUzVxf/PeShAX59pj1EdxcWCJSNKShUnK40bM0nF13OjJCaKn2jcuHEDo0aNwuXLl2FjYyN+0IQ0UPUm0bh48WKl2ysb9dPGxoZGBSVEwj5/Zhp9lm2PoajItMfQ0qr+uUUtzeBr1Uow0cjLY+ZLEadNSMuWLTFx4kS0adNG9IMIIfUn0SCE1B0pKUx7jLLU1Jj2GMLaU4hDlPYZX9PQYJKKjIwv65KSxBs11MLCAps3bxYrTkJIAx5Hg5CGLC+PaTNR9gu7pkpLgdhY4UmGlhZgbl7zJAMQv0QDKF9NU1AgvN1IRfKE1f/UcT/99BMmTJiAoqpmqCNEiqhEg5AGgD/pWGYm8/f1cN/NmjFDdFe3yydfYSHTHkPYoFgGBkybDEkp27VVlORFTQ1o0YIZjZTvwwdmiHNRZn3v3bs3TE1NceTIEfGClZFPnz7hr7/+ApfLRU5ODv755x8oKyvLOqxqKS5mqr5UVJgRYGXdHvfFixfIz89Hjx49ZBuInKBEg5B6issFsrKYxCI7m3ksTEYG89e4MTNglriNJAHm/DEx5a+hqMj0KpF0Z67qlGgATLKTlvalIWlRETNbrK5u1cdOnz4dLVq0EC9QGTp9+jS4XC6mT5+OHTt2wNPTE4cOHYKSknx97GdmMqVk/IHXFBWZ12jTpsy/tZ07paSkoF+/fvj06RMcHR3h7++Pb7/9tnaDkDM1esUlJSWhefPmlY6uGRsbi4yMDMr8CKkFRUXMB3NWFlOCUXbMisp8/sz8qaszCUezZqL9ckxOZnpxlKWuzrTHkMbgu5UN1lUZFRVmcrWUlC/rkpOZko6qutj+9NNP4gUpY8eOHUP79u2xbds2dOjQAQsWLICysjL27dsHRXH7E8uIsLY+XO6X5Bhg2t80bcr81bRUrio8Hg8zZszA58+fsWjRIuzcuRO9evWCo6MjAgICqDdSBWrURsPBwaHcmBRl3bx5E15eXjW5DCH1XmkpUyqQns58gPKThexs5ss/J4cZdCovj6maKChgftUXFjLrPnwAXrwAIiOZCcSysytPMhQUKv5yzs9nfkE+f84MsvX1EN5f43KZUgxhSUazZkx7DGmN8F/dEg2ASaK+/p4tKRFMPITJzc1FVlaW6BeRsYyMDFy5cgUuLi7gcDiYP38+Vq5ciUOHDmHq1Kkoreg/tY7gz4UjrK1PWfzX/8uXwNOnwLt3zHtIGgO4Hjp0CKdOncLKlSuxcuVKxMbGYs2aNXj48CG+/fZbDBs2DPfv35f8heWc2CUaSUlJSPz/TxYej4c3b96gWQV9xAoLC3H58mW5bERFSG0oKGDaDKSlle8GKmnKysyvPi0tpppEQYFJSJKTmWSmLP6spx8+ADo6TEkAv9S9sBCIji5fsgAwM6SKUhVRE9Ut0QCYe9DVZXqd8KWkCN5fWceOHcPEiRPx5s0bmJiYiB9wLTtz5gxKSkrg4uLCrlu0aBGKioqwfPlyKCsrY/v27XVy8LGSEiaBLfua5HCYBLGy90lJCfNeSktj9tfU/FLaUdNGyPHx8fD29oatrS3mzJkDANDU1ISPjw9mzpyJoKAgrFu3Dt9++y2GDh0Kf39/KuH4f2InGqdPn8aff/7JjpC3ZcuWSvfn8Xiwt7evdoCE1DelpUyJRWoqU1IhTV9PNCasWLlJE+YvN5f5sv26+ydfSQmTbKSkAM2bM+eJjy//i1FJiWmPUZ02HuISt3trWS1bMqU1/C+t0lLmHlu3Fr6/lZUV1q5dKzdjaBw7dgxt2rSBlZWVwHp/f38UFRVh9erVUFFRwebNm6tMNni82mt8WVDAJLBl/38VFYH27b+8VrOymL/KfsPyG0B//syUjKiofEm0xX2N8ng8TJ48GcXFxUKrnr5OOLZu3SqQcAQEBOCbb74R74L1jNiJxk8//QQnJyc8e/YM8+bNg6OjY4UZvoKCAgwNDeHo6FjjQAmRd7VVesH/FaelJfoXcKNGzAd5YSGTUHz6VL7qpbSUSY5SU8sfr6HBHF9bkyHXpOoEYL649PWZhImP3yhURaX8/p07d0bnzp3FD1QGsrKycOnSJXh5eZVLIjgcDlauXImioiJs2LABysrK2LBhQ7n98vO/tIMoKGBeUwYGTEmYtFTUoFhVFTAx+fJabtSI+TMw+NIbhV/NWFl1Cb/hb2oq895o3170BGrHjh24fPkygoKCYFzJ9L+amprw9fXFrFmz2ITDxsYGw4YNg7+/f4NNOKrVGNTQ0BCGhob4+++/MX78+HJZMyGEIU7phZLSlw9yHk/0P4BpeKmlxSQYNWmFr6oKtGnDfAmnpjK/+quq69bWZmZKFaWLqKTUtEQDYBqApqQwX0AA81wmJTGjlpb15s0b6OvrQ1NTU/wL1bKzZ8+iqKhIoNrkaxwOB+vWrUNRURE2bdoEVVVVrFq1Cvn5HDa5KPv85uQw88U0acJUjYk7AV5VUlOZpK9scqupyTQorqhKS1mZKWVr3pw5NifnS2mHsGo9vsxMpi1Hu3ZVx/b27VssWLAAgwYNwowZM0S6H37CwS/hWL9+fYNOOGrU6+TgwYMCj3NycqCqqiq3fbUJkRRxSi80NZk2EFpatftlXRllZeYXo64ucx8fP375QubjcJgvnZYtaz++mrTR4FNQYO7x3bsv69LSmMaiZRMXGxsbuLm5ISgoSPwL1bKjR4/C0NCw0vYBHA4Hf/75J/LyOPj994PIymqNyZNnVnnu7GwgKor5YjcwEF76Iw4ej6nW+Pix/LbmzZkEVtRSBw6HSdQbN2Zel4WFX0o7Pn8u36g5PZ0p2aqsNozL5WLSpElQUlLC7t27xW7T0rhxYyxcuLDBl3DU+GPt7du3mDVrFqytrfHNN9/gyZMn7Lbly5fj5cuXNb2EyOLj4+Hp6QkzMzMkVNFcOSwsDKNHj0b37t3Rr18//Pbbb8gXNtIQIWLg98T43/+YX8sVJRn8BomdOwNmZqIPGlXbFBWZOC0smF/6/F+yKipAhw6ySTIAyZRoAMzzXvbYsr1oeDweduzYgYkTJ1bvIrXo8+fPuHDhAr777jsoVPCCyslhSg+eP+dg+vRNGDlyGrZt243du3eLfJ20NKZXUkJC9asBuVxmgDdhSUarVszrrSZtQ1RVmQTexATo2pX5t2xCmprKtM2pyJ9//olbt25h8+bNaF1RAx4R8BOOd+/eYdWqVbh37x5sbGwwYsQIPH78uNrnlRc1KtGIiYnB999/j/z8fHTo0EFgYrL09HScOnUKoaGhOHz4MExNTWscbGUuX74Mf39/9O3bt8p9//33X/zyyy9Yv349Bg4ciNjYWEyZMgWKiopYtGiRVOMk9VdpKdOQrbIqkrpYeiEKDudLEXVJCZOAyLLDgiRKNADmHlq1Yr7w+DIzmQaH/MazHA4Hrq6u1btALTt37hwKCwsFqk34VQr8btNfjwqroKCAxYsXo6SkBNu2BUFZWRkTJkwAh8NUkzRrxiSVSUnlX9c83pf2PHp6TNIp6mu6qIh5r5T9baegwFRn1GTCPWEUFJhqRXV1phvs189BUhKT+OvoCB7z4sULLFq0CM7OzvD09JRIHI0bN4afnx+8vLywZcsWrF+/Hr1798arV6/kpqFxddToo27r1q1QUlLCiRMncODAAXaKdoCZev3MmTNQU1OrleLGzMxMBAcHw9nZucp9s7Ky4OXlBUdHRygpKaFDhw4YPHgw7t27J/U4Sf1UWZIhL6UXolJSkv0Q0DVtDPo1YT1yvi7VSExMxPPnz8GVxsAMEnbs2DHo6+ujV6/eyM4G3r8Hnj1j2lekpgp+wfIpKirC398fgwcPxubNy3Hx4na2BKB5c6YqwsyMeSysbQaXyzxfz58Lb0RcVm4uM+ZL2SRDWZm5jqSTjK/xS+LKjlf2/r1gj6vi4mJMmDABmpqa2LFjh8S7AfMTjoiICACo9z9wa/Rxd//+fYwfPx5mZmZC/yNat24NDw8PPHjwoCaXEYmrqyvaidKyB4CTk1O5Rj3x8fHQl+RkDKTB4CcZZfv9q6szv866dGHqjCUxmRhhSDLRAJhSja99/vxlWvk9e/agS5cuKJT0DHQSlpOTi3PnbmPgwCl4/lwBb94wyUVlVRsKCkypRYcOijh9+leMGdMNixf/hB07yv84bNoU6NiRqdIQ1jajuJgZZCsqiik5ESY9HXj1qnxMGhrMuTU0RL7dalNXZ5Kmssl+bOyX//Pff/8dDx8+xPbt26ErxUFhjIyMMHfuXAQHByM8PFxq15G1GlWdZGZmVlncY2hoiGz+/14ddfLkSdy+fRuHDx+WdShEzpSWMsXuZZMMDQ3A1LTqYa1J9VRnUrXK8BsRfv3/mJjIVB+4u7ujc+fO0KiNb0Ex8UeUzcgAjh37DwUFrdCnz8gqkwv+2CpNm379hauMv//+Gy4uLpg1axZUVFQwZcoUgWP5VWjNmn1p31C2oKeggHlPaGoyCRy/o05SkvD2EM2aMclLbZbyaWoy3Vvfvv1SAsPj8ScFfIpff/0V48aNq7DnjiQtXLgQu3btwvz583Hz5s1qlZ6UljKlqcrKku8RJAk1SjSaNWuG+K87ogsRFRUFbW3tmlxGqnbt2oWtW7fijz/+gKWlpazDIXKEx2MafpbNozU0hBfPEsmRdIkGwHwpft12PS+P+QI3NjaudOyE2lZayvSkyMhg/uX3prh8+Rq0tJqhe/fu5Y5RVPySXDRpUvGXuoqKCo4ePYpRo0Zh2rRpUFZWxg8//FBuPwUFpjqwRQtmZFlhQ9Xn5DClF1paTIIibDA4fX2m94osNG3KJDixsV/WFRQUYcKEX9G8easqB6OUlCZNmmD58uWYOXMmTp06hdGjR4t1fFoakxTzq8XatWOqZ+uSGuWQtra2CAkJEdrDg8vl4tixYzh06BD69OlTk8tIRWlpKRYvXoz9+/dj//79GDhwoKxDInKE/+un7PQX6upMkiFnE2TKHUn1Ovlao0bl2wckJQEXLlxkp12QFS6X+UJ5+5aZzyMmhvni5n+5FxQU4NatWxgwYAA7aqWSEpMI8HtdtG0rWiNkVVVVnDhxAg4ODpg0aVKlJb2KikyC1rkzcy1hMjPLJxkcDhOPrJIMPm1twdFgd+7cgbdv32HZsoPQ1Ky9b+upU6eiY8eO8PHxQVHZfuQVyMlh2rq8eyfY9iY9XTox1kSNPg69vLxw/fp1jBo1Cj179gSHw8HOnTuxa9cuREZGIiMjA1paWnVyUrVly5bh6dOnOHbsmFTr4Ej9wy/JEJZkmJpSklEbpFGiATBfml+3L0hPz8HQoY5YtWoV/Pz8RDoHl8v0qigqYr4A+MlA2RJxYSXkZdfxSy+qmiTv3r27KCjIx5AhA9heTY0bV7/Rrrq6Ok6fPo3hw4fD09MTOjo6GDRoUIX7q6gwY160bMkkZxW10QCY94ex8ZcqFVlr2ZJpM3Lx4jPs338Ao0aNgo1NX7x5wzROrY33s5KSEtatW4cRI0Zg27ZtmD17doX7FhUxJRgVJRTSnsG2Omr0FLZq1QrHjh3D77//jhs3boDH4+HWrVvMiZWUMHDgQPj4+KBV2ZZWtezZs2fw8fHBnj17YGBggMuXL+PSpUsIDQ2lJIOIhcdjilrLfpCqqVFJRm2SdBuNr8/TvDlTegAAKiqqOHjwLvr0YRqK83hM8sBPJIT91XbnFBUV4Pbtk9DW/oSJE3vWeBAtPg0NDYSGhsLGxgYeHh54+vRplQ3m1dWZJCInhxljIzdXcLuamvDxLGStadNc/PqrN3R1dTFnzlwAX+ZdMTWtnfYjw4YNg4ODA3799VdMmDCh3GSlpaVMd+LkZOEzKvOrs+pin4YafywaGhpiy5YtKCgowLt375Cbm4tGjRqhXbt2UK3FV9OQIUOQlJTEdrF1dHQEh8OBs7MznJycEBsbi+L/L18KDg7G58+fhVaXXLhwQeaJEamb+ElG2WJgNTXmw4gGxK090irRAJji/PR05v9bSUkZHTt+i/x8IDKSSTKq6r5ZG1RVmVKLZs0AJaVCXL26D2PHjoWKimRfhJqamjhy5Aisra3h4eGBy5cvl5tQTPhxgLk5k5B/+MB0ZW3WjBmFsy62XfLz88O7d//i0KH/BIaZz81lqqtMTKTfpZvD4WDDhg3o3r07VqxYgQ0bNrDbMjKYxK2iWpVmzZiebZJKMiVNYr+/1NTUYG5uLqnTie3ixYuVbv96MLF9+/ZJORpS3/B4TF1o2SRDVZWSDFmQ1IBdwqioMIM3ffwIvHwZgbS0FPTu7Vit3gAPHz5E27ZGaNFCp+qdq8CfibdZM8FuoGfPXsbnz5+l1kOic+fO2Lp1K3788Uf89ttvCAgIEPlYfgPU2pwBVlzXrl1DYGAgfv75Z7i798KbN4K9j7KzmR8Y7dpJ/x66du2KiRMnIjAwELNmzYKeXnskJJTv1canrs60MZHmZHeSUKNEQ9RWuYqKimjWrBl69Ogh9RFCCZE0Ho8ZH6BsnSglGbLB5QqOw8DhSP7/QF+fGXxq377fERl5F+fOVd677mscDpOsPHv2EDNmuKJHj244ceKEQKJStlREWCnJ/7V332FN3V0cwL8hDBFQcIGIIkNQWUUFnFgXuKqWWkVtrbjrqG3VuusoFa21tq7WOuvWYsW9iqviFrUWqKKiAioyZIhAYnLeP/KSEgiQkAQQzud5eCT3/nLvCd7cnPxm/rb8BKO4YYuhoaEwNzdH165dVY5RXSNGjMCZM2ewaNEi+Pr6qn2uyppkZGZmIigoCM2aNUNISAgEAlnTz717ikvQv3wpaxYtj8k7g4ODsXt3KD77bBkWLvxZaRl9fVl/orp1K+/ftiCNEw3FN89/75b87YW39e7dG999951K1W+MVQZPnvzXZp/P0FCWZGirqlIqleKHH36AhYUFRo4cqfWZCKsSZc0m2v5z6evLqqK//noTnjyJLbLP0LD4HwMD4PXr1wgIGIoaNZ4jMjIM16//gQ8++EC7QQIQiUQ4cOAA+vfvD0Md1psLBAKsXbsW165dw7Bhw3Dr1q0q0b/tyy+/REJCAi5cuCCfJ0UolPW3untXseYsOVn2f6vLPhCy5jprDB/+HdatW4+AgNvw8PCQ7xcIZJ1XGzasnE1QxdEo0di8eTP27duHU6dOwd/fHx4eHjAzM8OrV69w69YtnDp1Cr1790b79u3x8uVLHDhwAEePHoWzszPGjh2rrdfAqikiWaeo/H8L/q5sW/7vqi69TiRrEy1cbantJCMvLw9BQUHYtWsXAGDv3r3YuHEjbGxstHOCKkaX/TMKql8fMDExRcuWnjAw+C+RUKVj4Lx58xAbG4tTp05hypQpmDlzJvr166f1la1Pnz6N9PT0cplYKr+/hre3Nz766CMcP378rf7CeOTIEWzcuBEzZ85Eu3btFPbp68uSDVXXRdGGzEzZYne5ucBHH32Cffv2Y8WKFdi8eTMEAgFq1ZI1k7yNMwxrlGikpaXh+vXrOHz4cJGV7YYOHYpHjx4hKCgI/v7+6NOnDz766CMEBQXh4MGDnGiwMsvNlfW+fvmy/Hv4GxjIkgxtfbhlZGQgICAAp0+fRkhICMzNzTFt2jS4urpi5cqV+Pjjjytd7UZMTAzs7OxQo4LueLoacVLY4sWL4eDggMGDB6v1vIsXL2LFihX49NNP0b17d3z33Xfo27cv1q1bp/Wh/qGhoahVq1aJQ0+1yc3NDatWrcKYMWMQEhKCuXPnlst5tS01NRWjR4+Gm5tbsX1O8tdFuXtX8T7z5ImsNkFbk2Ll5so6ehYcLm9sbIxPP52Ab75ZhLNnj2PMmF6oXVs756sIGg3aWbduHYYNG1bs8rlNmzbFkCFDsGbNGvm23r17lzqbKGPKZGXJhptFRcnaz9/2JOPp06fw9fXF+fPnsXXrVsycORPjx4/H7du34ebmhk8++QQBAQFISkrSzgm1YNmyZWjZsiXs7OywdOlSZBSeTKQcqFOjkZubi+HDh+Onn35S6xxSqRR79uzB+fPn1XpeTk4ORo4ciSZNmmDp0qUAZPe8Ll26YOHChVr9e4nFYuzfvx/9+vUr1xF+o0aNwtChQzF//nycO3eu3M6rTZMmTUJKSgq2bt1a4t8ufwK+wrVYjx4VnRGYSNZ3SCSSJQ/Z2bJ7VkaGrH9XSoqsg/GzZ7J5MOLjZfPxREcXnZMHAPr374sWLWph9eoJMDLKLVrgLaJRovHo0SOYldLd1cLCAvfu3VPYpsu2RFa1EMn6R8TEyDpoVcDnGoD/kgxtfXuOiYlBu3bt8PDhQxw5ckRhGWoHBwecPXsW33//PY4dOwZXV1eEhoZq58RlRESYM2cOvvrqK/Tv3x9ubm6YOXMmmjRpgpkzZ+L58+flFouqicabN28wZMgQbNu2DZ9//jn279+v8jn09PRw69YtfP/992rFNn/+fNy9excbN26U3xsFAgGWLVuGlJQUefKhDWfPnkVaWlq5NJsUJBAI8Msvv8DR0RFDhw5FcnJyuZ5fU3v37sXu3bsxf/58vPPOO6WWNzGRrYtSsGKRSPal5++/gVu3gMhI2c/t27Jh0FFRsmaXe/dk5eLiZB3K4+NlzS/507a/fKm8I3C9eoCHhxArV87B48ePsGrVKq29/oqgUaJRr149HDt2DG+KWcGHiHDq1CkY/7+7tEQiwcGDB4utAWEsn0QiezPeuSP79lCwB3hBenr/dc4zMpJ9AzExkY3jr1VLtp6BhYWsd3a9erKOVJaWgJXVf+ssWFvLenDb2MjaQJs0kc1y2LSpbEibg4NsimVtJRkRERHo0KED8vLycO7cOfj5+RUpIxQKMXXqVERGRsLW1hYffvghhg0bhrQKmF9YKpXis88+w+LFizFmzBjs27cPJ0+exPXr19GzZ08sW7YMTZs2xbhx43D//n2dx6PK9ONEhHHjxiEsLAzLli2Dt7c3hg8fjqioKJXPIxAI5PcuVVy5cgXLly/HuHHj0K1bN4V9rVu3xtChQ7FixQqt1eiGhobC1NRU6fWja2ZmZti7dy9SU1Px8ccfQ6psBqlK6Pnz55gwYQK8vLwwc+ZMlZ+Xvy5KQfmTt0kk2ptbxdRUtoqtra3svta9e3f07t0b3377LVJSUrRzkopAGvjxxx/J2dmZ/P396YcffqDQ0FA6cuQI/fHHH7Rq1Srq168fNW/enL7++msiIvr000+pefPmtHv3bk1OW+7i4+PJycmJ4uPjKzqUKi83l+jJE6LISKLr14v/iYkhSksjkkorOmL1/PHHH1SjRg1q1qwZPXjwQKXniEQiWrRoEenr61PDhg3pyJEjOo7yP2KxmD7++GMCQNOmTSOpkj94bGwsjRs3jgwNDUlPT48GDRpEN27c0FlMV68S9e3738+CBUXLTJ8+nQDQ/PnziYgoISGBrKysyMHBgVJTU0s8/uvXr8nNzY3++OMPlWPKycmhFi1aUOPGjSkjI0Npmbi4ODI0NKRPPvlE5eMWRywWU/369SkwMFDjY2ni559/JgAUEhJSoXGUJCUlhbZt20aDBw+m2rVrk5GREUVHR5fpWElJJd+Xyvrz99+y+5kyUVFRJBQKadKkSRr8FdSnzc89jRINsVhMixYtIhcXF3J2dqbmzZvLf5ydncnZ2ZkmT55M2dnZRET09ddf0w8//KBx0OWNEw3dy84mun+/9Dfk/ftEWVkVHW3ZrFmzhgQCAfn4+FBycrLaz4+MjCRXV1cCQKNHjy72A01bcnNzacCAAQSAgoODlSYZBT19+pRmzJhBtWrVIgDUo0cP+vPPP0t9nrrOn1dMNJYsUdy/dOlSAkCTJk1SOHdERAQZGBiQn58ficXiYo//+PFj6tmzJ50+fVrlmGbOnEkA6MSJEyWWmzZtGgkEArp165bKx1bm9OnTBIBCQ0M1Oo6mpFIpDRo0iIRCIf31118VGks+qVRKUVFRtGTJEurYsSPp6ekRALK0tKSRI0fShQsXNDr+s2fK7003bhDdvClLGv75hyg6mujuXaLYWKIHD4gePZJ9iUpIkB0jKYkoOZkoPZ1IIin5nOPHjyd9fX36999/NYpdHZUm0ciXmppKJ06coE2bNtGaNWto/fr1dPDgwSrzwcyJhm6lpsrepMUlF5GRsjdobm5FR1o2UqmUZs2aRQDovffekyfeZZGbm0szZswgPT09srW1VevDUB2vXr2i7t27EwBauXKlWs9NT0+nJUuWkJWVFQGgNm3aUGhoKL1580YrsZ06pZhorFjx377169cTABoyZAhJlNy98/dPmzZNK7EQEV29epX09PRo9OjRpZZNS0sjCwsL8vPz0+icEyZMoJo1a2p0LWlLRkYGOTg4kI2NDaWkpFRIDLm5uXTy5EmaPHky2dnZEQACQJ6enjRv3jy6cuWK0uuhrMRiopwcorw82e+6rllNSkoiMzMz6tevn25PVEClSTRiY2Pp5cuXGgdR2XGioTvJycUnGLdvyzJ/LX0+VQiRSETDhw8nADRmzJgSv0mrIyIighwdHQkAffbZZ1p9H6alpVG7du1IT0+PtmzZUubj5OTk0Lp168jBwYEAULNmzWjr1q0a13AcOqSYaKxdK9seGhpKenp61KtXL8rLyyv2+RMmTCAAtH379iL78vLyKEuNKrPc3Fxq2bIl2djYUHp6ukrP+eGHH1Sq/SjOmzdvyMrKigYOHFim5+vCjRs3yNDQkPr06aPVD/SSJCUl0ebNmykgIIBMTU0JANWoUYP69OlDv/zyS5W7Xy9evJgA0JkzZ8rlfJUm0fDw8KDDhw9rHIS2PHnyhD766COV/jgXLlygwYMHU+vWrendd9+lefPm0evXr5WW5URDN4pr74yOJkpJefv6XxSWmZlJ/v7+BIAWLVqk9SaEV69e0aRJkwgAmZqa0tSpUykhIUGjYz5//pw8PDzI0NCQ9u3bp5U437x5Q3v37iVPT08CQL1799YoztBQxURj40aiU6dOkaGhIbVv377Ub/kikYh8fX2pRo0adP36dYV9YWFhVKNGDbp586ZKscyZM4cA0LFjx1SOPzc3l+zs7Mjd3b1MtTznz58nAJWur9uqVasIAC1btkwnx5dKpXTr1i0KDg4mHx8fEggEBICsra1p7NixdOjQoUpRw6Mrr1+/piZNmlCrVq3KJZmrNIlGYGCgzi4qdZ08eZLatWtHX331Val/nLi4OHJ1daWtW7fS69ev6cmTJ/T+++/TzJkzlZbnREM1UqmUoqOjacWKFdSzZ09q3rw5LVu2jDIzM4uUVdbOeeOGLMGoCp49e0atWrUioVBIGzZs0Om5bt68SUOGDCE9PT0yMDCgESNGlKmz2+PHj8nJyYlq1qxJJ0+e1HqcEomEVq5cScbGxmRubl7m2o3t2xUTjUWL7pOJiQm5ublRWnE96gpJSkqixo0bU+PGjSkpKUm+PSoqiqZOnUoikajUY1y/fp2EQiEFBQWp/Rp2795NAGjz5s1qP/ezzz6jGjVqKH1fVSSpVEoBAQGkr69Ply5d0tpx7969S7Nnz6YmTZrIm0S8vLxo4cKFFBkZqfUEvjLbvn07AaDffvtN5+eqNIlGfHw8jRkzhubNm0cRERH0+PFjSklJUfqja3v37qWHDx9SREREqX+cJUuWFGnrOnXqFLVs2VJpj/S3JdGIjo6mn376qVyz+pcvX1JoaCiNGTNG4Ubg7OxMHTp0IABkYWFBX3/9tfw6SExUnmRUlVa427dvk52dHdWsWbNca/wePnxIkyZNImNjYwJA/fr1o4iICJWee/fuXWrcuDHVrl1b5eeUVWxsrPza6N+/Pz179kyt52/c+F+S8e67mWRq+gnZ29vT06dP1TrOjRs3yNjYmDp16lRiU4syeXl55OrqStbW1mVqtpJKpeTt7U2NGjVS6/0qkUioUaNGNGDAALXPWR5evnxJdnZ21KRJk1JH95QkIyOD1q9fL79O9PT0qGfPnrRhwwa1r5eqRCKRUJs2bdS+bspCm597AqKyjwB2dXUFIJsfoyQCgQDR0dFlPY1aLl68iKCgIISHhxe7VsSQIUPg6OiIb775Rr7txYsX6NSpEzZs2IBOnToplE9ISEC3bt3w7bffol69ehrHuHLlSjx8+BBff/016mhpHttbt25h7ty5yM7OhqWlJSZPnowOHTpo5dgFSSQS3L17F9euXcO1a9cQExMDqVQKExMTtGrVCl5eXvDy8oKVlRUA2cRUO3fuxIULF1CjRg107Tocvr4DYW5uIT+mnh5gbZ0DE5PSp/okIjx69AgXL17E1atXUbt2bfj5+cHHx0fr60ioi4hw6NAhrF69GmZmZvjmm2/QsmXLco8jPT0dYWFh2L9/PzIzM+Hq6oohQ4agbdu20FOyUMf9+/fx1VdfgYiwbNkyODo66jxGiUSCffv2YcOGDTA2NsaUKVNUXhH0wAFrXLlSB3l5efjnn39gaLgZ69b1g7W1tdpxhIeHIzg4GP369ZMvetZEhSU6N23ahG3btmHx4sVF1slQ1d9//40pU6Zg1KhR+Oijj1R6TlRUFCZNmoTZs2eX27Tj6vr3338xefJkeHt7Izg4WOUp9KVSKW7evIkTJ07g/PnzyMvLQ5MmTdCzZ0/06NFDK/feqqAs101ZpKSkYM6cOSV+lqpKo7VOPD09NTp5RUlLS0PtQhPHW1jIPvhSCy/TqQPt2rXD0aNHMXHiRISEhKBp4Zlg1HTmzBmEhITA2toa06dPx5YtWzB37ly0a9cOkydPRkMNlxuUSqW4fv06jh8/jhs3biAzMxMCgQDOzs4YNmwYvLy80KJFC+jrF72cWrRogW+++QZxcXHYtOlPHDsWiRMnbqFdu7bw9+8JS8v6sLHJgbFx8UmGSCTC7du3cenSJVy6dEk+C2WzZs0QHx+Pv/76C7Vq1ULXrl3h5+eH5s2bl/v6IK9evcL333+Pc+fOwcvLC7NmzZJfU+XN3NwcI0aMwODBg3Hs2DHs3bsXc+bMga2tLQIDA9GtWzd5UhYVFYWZM2eiZs2aWLZsmUofstogFAoxaNAgtG3bFiEhIfjmm29w/vx5fP755zA3Ny/xuWKxHsTiN4iOjoFEIsHYsZ/A2lr1ibUK6tatG2JjY7Fnzx7cv38fT548wf79+5Vey/liY2Oxc+dO+Pn5lTnJAAB3d3d06NABu3btQp8+fVS6Xs6dOwcDAwONzqtrzZs3x7hx47BmzRrs27ev1JlLnz59iuPHj+PkyZNISkqCiYkJ/P394e/vjxYtWlS6tX4qmru7Ozp16oSdO3eid+/eWvuyqlMa14lUMqo0nfj5+RXpWyISicjJyYnCwsKKlNdF08m1a9fI0tKSateuTeHh4WU+zooVKwgAdezYUV5VKRKJaNmyZWRiYkI1atSg4OBgyi3D2NCsrCxavXo1OTs7EwCqX78+jRgxgnbt2qXWPBBSKVFcnKyJ5MCBBBo4cDEZGLQlgcCTBg78hG7fvl3kOUlJSbRp0yaFHuXGxsb03nvv0a+//kqJiYlEJJvL5ciRIzR48GAyMjIiANS8eXNavHgxPXnyRO3XXBZXr14le3t7EgqFtGTJknLrda8qkUhE27dvJ3d3dwJANjY2tHz5cjpw4ACZmJhQs2bN6PHjxxUWn1gsppCQEDI0NKT69euX2gl1/vwcql37HOnpHaH27VNJ0+kb3rx5Q/7+/qSvr08rCo6VVSIvL488PDyoYcOGKvcHKUlMTAwJhUKaOHFiqWWlUik1adKE+vbtq/F5dU0qlVL//v3JwMCArl69WmR/VlYWbd68mXx9fQkACQQC8vPzo127dhXbKZ/95969e6Svr09jx47V2TkqTR8NVZw7d44mTJig69PIqZJoDB06lObMmaOwLSkpiZycnJS2T+uqj8ajR4/IxcWF9PX11R5GKJFIaOrUqQSAAgICKCcnp0iZ+Ph4GjhwIAEgJycnOnXqlErHfvDgAX355ZdUu3Zt+TwI27ZtU7sdm0iWZDx4ULRPxqlTL+iLL2bLk4g+ffrQwYMHKTg4mNq2bSvvUd6oUSMaP348HT58uNQb0MuXL2n9+vXUqVMn+c2ra9eutGXLFrWGLKr+2qS0YsUKMjAwoMaNG+u8b4OmpFIpHTt2jLp06SLvS+Pu7k7Pnz+v6NCIiOjOnTvUqlUrAkBDhw5V2safk5NDTZtuJoHgEHl7P6e+fWUzhWoqLS2NHB0dydLSssT3+YIFCwgAHTx4UPOT/l/+ZEx3794tsdzVq1cJgEZDjstTWloa2draUtOmTenly5ckkUjo7NmzNGLECDIxMZEPef72228rff+3ymjKlCmkp6dHd+7c0cnxK12i8fz5c7px4wZdvXpV4ef8+fM0ZswYcnd318ZpVKJKorFs2bIi3wqOHj1Krq6uSnty67IzaHp6unxipHnz5qnUgzo3N5eGDBlCAGjixImlDpE7fvy4fM6FwYMHy2sDCpJKpRQeHk79+/cngUBA+vr6FBgYSBcvXixzr26JRDYrnrL5MfLzorS0NFq0aBHVrVtXoUf5okWLNOpR/uDBA1q4cKF8DoeaNWvSxx9/TCdPntTKXBapqanUr18/eadLTTq+VYQrV67Q/PnztfKtXJtEIhEtXLiQ9PX1ycrKig4dOiTfJxaL/z9T6WLy9EyQdwhVUiGmtvXr19O6devI1NSUvLy8lCbut27dIn19fRo2bJjmJyzg+fPnZGpqSgEBASWW++qrr0hfX7/S/Z+V5PLly6Svr0/e3t5kb29PAMjMzIxGjx5NERER1WrEiLalpqaSubk59ezZUyfHrzSJRm5uLk2ePFlh6vHCP87OzjR06FCNA1WVskTj9u3b5O/vL/+AjY+PJw8PD9q8eTPl5OTQgwcPqFevXrRw4UKlx9T1qBORSEQjR44kADRs2LASmznS09Opa9euBICWLFmi8hs1JyeHFi5cSEZGRmRmZkY//PADicViys7OpvXr15ObmxsBoHr16tGcOXM0no9BIiG6d0/5nP7KXt6rV6/oyJEjao8cKI1UKqULFy7Q2LFj5TU09evXp7Fjx9LJkydVGsZY2IULF6hx48ZkYGBAP/74I98sdeDmzZvypp5PPvmE0tLSKCgoiABQly43FYa3ajors1QqJXt7exo+fDiFhYURABo+fLjC/6tIJCJPT0+ytLTUySi6RYsWEYBip8eWSqVkZ2ensw8VXVqxYgUJBALq1q0bbdu2rUrPdVHeli9frtHkbyWpNInGzz//TM7OzjR48GCaMWMGOTs70/jx4+nLL7+kLl26kLe3N23fvl0n1daF+fn5kaurK7m4uJCTkxO5uLiQq6srzZkzhy5fvkxOTk706NEjefmrV6/SwIEDydXVldq3b0+LFy8utmmgPIa3SqVS+cxvvr6+Sr8hJyYmkru7O+nr69PWrVvLdJ779+9Tz5495X0Z6tSpQwDIw8ODNm3apJX20exs2aJnhZOMO3dkU/ZWlJycHAoNDaXAwEB5k02dOnVo5MiRdPTo0VKbhiQSCYWEhJBQKCR7e3u6du1aOUVePeXl5dHcuXNJKBTK/78WLFhAn36qOI9Ggbe1Rud68eIFEf3XPFKwv8Y333xDAGj//v2an0yJV69eUcOGDaldu3ZKE9cbN24QANq4caNOzq9rr169qugQqqTc3Fyyt7cnNzc3rU3xn6/SJBp9+vSRrxmQkZFBzs7O8o4/b968oZUrV9L7779f6SaWUVd5zqOxa9cuMjQ0JCcnJ4qNjZVvj46OpiZNmpCpqanGkylJpVLat28feXh4UEBAAJ07d04r38ozMmSLCCmb7TMqiqgMlQc68/r1awoLC6OPPvpIvghY7dq1afjw4XTw4MEiVedJSUnk5+dHAGjQoEEqTzfNNHft2jXy8vKir776iqRSKY0cqZhoaHtaBYlEQgMGDCChUEh//vkn/f3332RgYEBDhgzR7okK2bBhAwGg33//vci+WbNmkVAorLC1RFjl9fvvvxMAWr9+vVaPW2kSjXfeeUee4WdlZZGzszNdvnxZoczkyZOLbZJ4W5T3hF1//fUX1a1bl+rVq0cRERH0119/kYWFBVlaWlJkZGS5xKAqqVQ2m2dUVPFrlkRHyxYeqqxyc3Pp8OHDNGLECDI3N5e3Iw8dOpT++OMPOn78OFlZWZGRkRH98ssv3FRSwYYNU0w0NJnoTSQSkb+/Px09elRhe2ZmJrm4uFCdOnXIzc2NGjRoUKYVd9Xx5s0bcnV1JQcHB4XaNalUSo6OjtSjRw+dnp+9naRSKXXo0IGsrKy02nqgzc+9orP3qD88FgBQs2ZN6OnpIS0tTWF/165dcfr0aU1PU6107NgRly5dgoWFBbp27YoePXqgfv36uHTpUqWZu0QiAZKSgH/+AR49AnJylJczMwOcnIASpiWocEZGRujTpw82b96MpKQkHDt2DIMGDcLx48cREBCAnj17onbt2rh69SrGjRvH4/orWF6e4mMjo7If6+nTp0hKSoJIJFLYbmZmhrCwMEilUty5cwdr167V+YRRQqEQ3333HR48eIBffvlFvv3OnTu4f/9+qfNRsOpJIBBg+fLleP78ObZv317R4SilUaLh4OCAEydOyA6kp4cGDRogPDxcoUxycjIyMzM1OU211KxZM1y6dAkdO3aEj48PIiIiYGdnV9FhQSwGEhOBO3eAhASg0P1ZzswMcHSUJRlCYfnGqAlDQ0P07NkTGzZswPPnz3Hq1CmsXr0a169fh7u7e0WHV+0RaTfRsLW1xc2bN9GvX78i+xwdHXHixAmsXbtWPmuorvXs2RPdunXDokWLkJ6eDgAIDQ2Fnp4eBgwYUC4xsLePj48P9u/fj549e1Z0KEpp9D1z4MCBWLBgAcaOHYtff/0Vvr6++P3336Gvr4+2bdsiMTERmzdvRrNmzbQVb7VSt25d/Pnnn+V2PiJAKpX9SCRFf8/MBFJTZeWKY2EBWFoCJiblFrbOGBgYoHv37ujevXtFh8L+TyxWvP4MDGRT2JeFVCoFEUEoFBZbS+Xt7Q1vb++ynaAMBAIBli1bhtatW2Pp0qUICQlBaGgoOnfujAYNGpRbHOztU5kTUbUTjbCwMPj4+KBhw4YIDAxEYmIikpOTAQBTpkzBlStXEBYWhgMHDoCIULNmTUyfPl3rgTPNvHwJvHghq5HITyik0rIdS08PqFtXlmBo8u2SsdIUrs2oUaPsxzpz5gyGDh2KEydO4J133tEoLm3y9PTERx99hB9//BHvvvsuYmJiMGnSpIoOi7EyUzvRmDVrFn766Sf5+hlTp06V76tbty4OHjyI8PBwJCQkoF69eujcuTPq1q2rvYiZxhITgf8vF6IRfX2gfn2gQYPK3QeDVR25uYqPNUlsa9euje7du8PJyUmzoHTgm2++wd69e/Hhhx9CIBDg/fffr+iQGCsztT8eqJTFXo2MjNC7d+8yB8R0hwh4/FjW/KEJIyNZ7UXdumWvtmasLLTZP6NNmzbYsWOHZgHpiK2tLaZMmYLvvvsOnTp10nhhRMYqEn8PrSYkEuDhQ1k/i5Lo6ck6bxb8N/93oRCoVQswNwd44AWrCNqq0UhMTISRkVGlXnp81qxZOHz4MMaNG1fRoTCmkTIlGjy87+0iFgP37wOvXytu19cH7O0BY+P/EgrGKjNt1WgsXrwY27dvR3JyMgwNDTUPTAfMzc0RFRVV0WEwprEyJRo///wz9u7dq3J5gUCAX3/9tSynYhrKzQViY4sOQzUyApo1486b7O1SuEajrJ1Bx48fjw4dOlTaJIOxqqRMiUZ0dLRa5bkGpGJkZ8tqMt68Udxes6YsyeAOnOxto60aDTc3N7i5uWkeEGOsVGX6qFmwYAE6deqk7Vg0lpOTg6VLl+L8+fPIyMiAo6MjPvvsM3To0EFp+QMHDmDTpk148uQJTExM4OPjg+nTp8PKyqqcI9e+9HQgLq7okNXatWXNJdxMwt5G2hjeeuDAAdjY2KB169baCYoxVqIyJRp16tRBo0aNtB2LxhYtWoTo6Ghs3LgR1tbW2L9/P8aPH48DBw7A3t5eoeylS5cwc+ZMLF++HD169EBaWhqmTZuGadOmVdppXFWVnAw8eVJ0e716QJMm3JGTvb00rdEgIkyZMgWtWrXCH3/8ob3AGGPFqjLfazMyMnDo0CFMnjwZdnZ2MDIyQmBgIBwcHLB79+4i5e/cuQMLCwv07t0bBgYGsLS0RO/evXHnzp0KiF57nj5VnmQ0bAjY2nKSwd5umvbREAgEiIyMxPfff6+9oBhjJaoyiUZUVBTEYnGRdld3d3fcvn27SPl3330X2dnZOHjwIEQiEVJTU3H06NFKO1d8aYhki5s9e1Z0n60tYG1d7iExpnXa6KNRp06dIjWcjDHdUTvRsLa2hrGxsS5i0Uj+qrHm5uYK2y0sLJCqZIYqJycnLF++HPPnz4e7uzvat28PAPj66691Hqu2SSSykSWFX6aeHuDgIGsyYawq0GQeDalUilGjRuH8+fPaDYoxViK1Eo2lS5diwYIF8PLy0lU8OqFs1Mv169cxffp0BAcH4+bNmwgPD4eRkREmTJhQARGWXV4e8O+/QFaW4nZ9fdnKqYXyLsbeaprUaMTHx+P48eOIj4/XblCMsRKp1Rl0586d2LJlC/T19fHOO++gffv2aNeuHdzd3aFXwcMY8tdTSU9Ph6WlpXz7y5cvlc7+t2PHDrRp0wZ9+vQBANjY2OCLL75AQEAAYmNj34oVZ7OygAcPZDUaBRkZyZZo12TBKcYqI01Gndja2iI+Ph6Swm8YxphOqZVoXLt2DTdu3EBERAQuXbqEVatWYeXKlTA1NYWXlxc6dOiAtm3bwsHBQVfxFsvV1RWGhoa4desW/P395dsjIyPRpUuXIuUlEgmkhcZ+5t+ACm+vjFJSZJ0+Cy89Y2oqay7hOTJYVVTWphMigkAggJ6eXoV/KWKsulHr48jQ0BDt2rVDu3btAMhqDy5duoSLFy/i0qVLOH36NAQCARo0aCAv1759e9SvX18nwRdkZmaGDz74AKtWrYKTkxOsrKywc+dOJCYmIjAwEElJSfjkk08QEhICT09P+Pv746uvvsKJEyfQpUsXZGRkYPXq1XBycoKjo6PO4y0rIiAhQbbEe2F16/LIEla1lbVG4+rVqxg1ahR27twJd3d37QfGGCuWRt97zc3N0atXL/Tq1QuArA00IiICFy9exJkzZxAWFgaBQABHR0ccOnRIKwGXZPbs2fjuu+8wdOhQZGdno0WLFtiwYQMaNWqEhIQExMXFIScnBwDQp08fZGdnY82aNZg5cyb09PTQsWNHrFu3DkKhUOexloVEIpuEKyOj6L5GjYAqMM8YYyUqax8NsViMBg0aoEmTJtoPijFWIgGVtu57GRER/vnnH1y4cAFXrlzBli1bdHGacpGQkIBu3bohPDwcNjY2FRJDXp6sP8b/8yQ5PT3Azo47fbLq4auvgJiY/x5/9x3QokXFxcNYVaXNzz2dNVYKBAK4ubnB3t4eV65c0dVpqoVXr2QjSwonGQYGgLMzJxms+ihLH4309HTkFa4KYYyVG+4VVcmlpgL37ilfGK1FC9m/jFUXZWk6CQkJQaNGjZBbOEthjJULHptQiSUlyTp+FmZhATRtygujseqnLDUaffv2hZWVFWrweG/GKoTaicbw4cPVKv/y5Ut1T8EgayZJTCy6vWFDnk6cVV9lGXXSqVOnSrnaNGPVhdqJxrVr1wDIOnuqStnMnKx4RMDjx4pzZAgEslqMOnUqLCzGKpy6TSfJycnIyMiAvb09z5/BWAVRO9EICgrCnj17sG/fPjRt2rTU8sePH8cXX3xRltiqreRkIDtbcZu9PXf6ZNXbmzeKfZUEgtInptu1axemTJmCp0+fomHDhroNkDGmlNop/hdffAFbW1t8+eWXEIvFpZbn2gz1iERFm0wsLDjJYExZs0lpt5e+fftiy5YtsOJJZhirMGonGgYGBvj+++8RFxeHxYsX6yKmau3xY6DgDOhCIdC4ccXFw1hlUZYRJ/b29vjkk0/4Cw9jFahMo07s7e1x7tw5lWo0HB0dMXHixLKcptpJSwMyMxW32djI5stgrLpTtyOoSCRCeHg42rZtCwsLC90FxhgrUZl7R9WqVUu+YmpJHBwcMGnSpLKeptp48wYovHq1mRmgZOFZxqoldWs0YmJi0Lt3b5w4cUJ3QTHGSsXzaFQSCQlFO7rxsgyM/UfdGg0nJyecPXsWLi4uuguKMVYqTjQqgcxM2QygBVlbq74yJWPVgbqTdRkbG6Nz5866C4gxppIqNbA8JycHCxYsQNeuXdG6dWsMHjwYERERxZZ/9eoV5s2bBx8fH3h6emLIkCGIiooqx4hlHT8fP1bcZmwMWFqWaxiMVXrqNp2Ehobi1q1bOouHMaaaKpVoLFq0CDdv3sTGjRtx8eJFvP/++xg/fjwePnyotHz++PoDBw7g/Pnz8Pb2xg8//ABpwWEfOvb0qWxIa0G2tqUP22OsulGnRoOIMHr0aKxbt063QTHGSlVlmk4yMjJw6NAh/Pjjj7CzswMABAYGYvfu3di9ezdmz56tUP727du4fPkyTp8+Dcv/Vx+U98Rir1/L1jMpqEEDwMSkXMNg7K2gbh+N6OhoSCQS3QXEGFNJlanRiIqKglgshpubm8J2d3d33L59u0j5y5cvw8bGBidPnkTXrl3h7e2NcePG4cmTJ+USLxHw6JHiNkNDoFGjcjk9Y28ddWo0BAIBrK2t0ZgnoWGswlWZRCMtLQ0AYF5oCk0LCwukFu5pCeDZs2d4/vw5YmNjERYWhoMHD0IkEmHcuHEqzQ+iqaQk2cJpBdna8oqsjBVHnRqNo0eP4rffftNtQIwxlVSLjzVlswISESQSCebMmYNatWrBysoKs2fPxsOHD5XWgGhTXh7w7Jnitjp1gFq1dHpaxt5q6nQG3bx5M5YuXarbgBhjKqkyfTTyJw9LT0+X97kAZMvU11My61WDBg1Qo0YNGBW4WzX5/8QVz58/12mshacZ19fnacYZK406NRp79uxBcnKybgNijKmkytRouLq6wtDQsMhwtsjISLRp06ZIeWdnZ2RlZeFRgY4Sj/8/ztTGxkZncaakAFlZitsaNy59FUrGqjt1+mjo6ekpfOFgjFWcKpNomJmZ4YMPPsCqVasQFxeHnJwcbNy4EYmJiQgMDERSUhJ69uyJmzdvAgA6d+4MR0dHfP3110hOTkZaWhqWLFkCV1dXeHh46CzOwqNMatWSNZswxkqmatPJrVu3MGvWLCQVfrMxxipElUk0AGD27Nlo27Ythg4dCh8fH5w8eRIbNmxAo0aNIBaL5QkIIFuFdsOGDTA1NYW/vz+6deuGWrVqYd26dTpd6bFgk4meHk8zzpiqVK3RuHXrFr7//nvocc9qxioFARFRRQdR2SUkJKBbt24IDw/XuFnl5UtZHw2BAGjaFKhdWzsxMlbVLVgA3Ljx3+P58wElraIAgLy8PIX+V4wx9Wjzc497BpQzCwvZD2NMPeqMOuEkg7HKg+sWGWNvBVVGneTm5mLQoEE4e/ZsucTEGCsd12ioIH8aY10Pe2WMFe/lS6DgXHovXwIJCYpl4uPjERkZifv378PR0bF8A2SsCsn/vNPGNP7cR0MF169fx7Bhwyo6DMYYY6xc7dixQ+kUEergREMFubm5+Oeff1C/fn0IhcKKDocxxhjTKYlEguTkZLi6uqJGaSsYloITDcYYY4zpDHcGZYwxxpjOcKLBGGOMMZ3hUSeFpKWlYc2aNTh16hRSU1NhZmaG1q1bY8KECXBxcVEom5ubi3Xr1uHo0aNITEyEqakp2rZtiylTpsDOzk6hrFQqxW+//YY//vgDjx49gpGREVq1aoVJkybB3d29SBz79+/H9u3b8eDBAwgEAri6umLcuHHo2LGjTl8/U4+q18uqVauwevXqYo8zfPhwzJkzR/6Yr5eq6+7du9iwYQNu3LiBFy9ewNTUFJ6enhg/frzC8gd8f2H5VLlmKvM9hvtoFJCamoqAgACkp6djyJAhaN68OeLi4rBt2zZIJBLs2rULLVu2BCBbZn7UqFG4ePEiAgIC0LZtW7x48QKbNm2CRCLB77//Ll8NFgDmzJmD0NBQ+Pn5oVu3bsjKysLWrVuRlJSE3377DZ6envKya9euxU8//QQfHx/069cPEokEe/bsQUxMDH788Uf4+/uX+9+GFaXO9ZJ/E5g8ebLSYZdNmzZF8+bN5Y/5eqmabt68iaCgIJiZmWHYsGGwsrLCw4cPsX37dohEImzduhWtWrXi+wuTU/WaqdT3GGJyc+fOJScnJzpx4oTC9lOnTpGTkxN99tln8m2HDh0iJycnWrp0qULZf/75h5ydnWnixInybZGRkUWeT0T0/Plz8vDwoAEDBsi3JSYmkouLCw0aNIjevHkj356VlUW+vr7Uvn17EolEWnm9TDPqXC8rV64kJycnunz5cqnH5eul6nrvvffI3d2d4uPjFbbnXzPjx48nIr6/sP+oes1U5nsM99EooEGDBujbty969OihsN3X1xcCgQB3796VbwsLCwMgq44qyMXFBZ6enjh79iwyMzNLLGtpaYnu3bsjOjoasbGxAIDDhw9DLBZj2LBhCkNpTU1NMWDAAKSkpCAiIkIrr5dpRp3rRR18vVRNUqkU77//PubMmVNk7Yj27dsDAJ49ewaA7y9MRp1rRh3lfc1wolHA5MmTsXz58iKrt7569QpEBFNTU/m2O3fuoGHDhrCysipyHA8PD4jFYkRFRcnLCoVCpe1e+e1rt2/flpcFoFBtVVxZVrHUuV4KE4lEEIlESvfx9VI16enpISgoCIMGDSqy7+HDhwAAZ2dnAHx/YTLqXDOFVaZ7DCcaKti9ezcA4L333gMg+yBJT09XehMAgIYNGwKQrX4HAImJiahTpw4MDAyKLRsfHy8vC8gyy8Ksra0VyrLKqfD1UtCxY8fQp08fuLm5wc3NDe+9957820U+vl6qh8zMTDx//hxHjhzBhAkTYGNjg0mTJvH9hRWruGumoMp4j+FRJ6U4d+4c1q5dCxcXFwwZMgQAkJ2dDQDFzpZWs2ZNhXLZ2dny/zxVygqFQhgaGhYpa2xsrFCWVT7KrpeCzp8/j6CgINja2uLRo0fYtGkTZsyYgRcvXmDs2LEA+HqpLry8vAAAAoEAAQEBmD59OiwsLJCUlASA7y+sqOKumYIq4z2GE40ShIWFYe7cuWjUqBF++eUXpX9oxvKVdL3069cPHh4e8PT0hJmZGQBZX44+ffqgV69eWLNmDQIDA1GrVq2KCp+Vs61btyInJwfR0dHYuXMnLl++jJ9++gkNGjSo6NBYJVXcNePm5lap7zHcdFKMNWvWYMaMGXB2dsbOnTsV3vz5be85OTlKn5uf3ZmYmMj/La7s69evFY5pYmICiUSitG2tcFlWeZR0vQCAra0tfH195TeAfHXr1oW/vz9yc3MRGRkJgK+X6sLHxwfvvvsuJkyYgN27d+PVq1eYNm2a/L7B9xdWWHHXjFQqrdT3GE40lPj222+xcuVKdO3aFdu3b0fdunUV9puYmKBOnTrFLhv/9OlTALJxywDQuHFjpKamKv2Pym//KlgWUL4kfX5ZW1tb9V8U05nSrpfS5Jd/9eoVAL5eqiMbGxu0bdsWjx49QkpKCt9fWKkKXjNPnjwpsWxF32M40ShkzZo12Lp1KwICArB69Wp5G1Rhnp6eeP78ufxNX9D169dRo0YN+WRNnp6ekEqlSnvm3rhxAwDQqlUreVkA8sxTWdnWrVuX4ZUxXVDlehGLxTh69CiOHTum9BhxcXEA/uuExddL1fTgwQN07twZs2bNUro/KysLgGzVTL6/MED1ayYvL69y32NKndmjGrl06ZJ8MhyJRFJi2fDwcHJycqJvv/1WYfuVK1fIycmJZs6cKd8WExNDzs7ONGHCBIWycXFx5OLiQh9//LF8W3JyMrm7u1NAQACJxWL59rS0NPLx8aHu3buXGhsrH+pcL/7+/uTm5kZxcXEK22NjY8nFxYV8fX3lk97w9VI1icVi8vHxoXfeeYeePHmisO/x48fk7u5Obdu2pTdv3vD9hRGRetdMZb7H8BTkBQQEBCA6Ohrz5s0rtvq7c+fO8m+tkydPxsmTJ/HBBx+gbdu2ePr0KTZt2gRjY2OEhoaifv368uctWbIEmzdvRvfu3dGjRw+kp6dj8+bNyM7Oxq5du9CsWTN52W3btiE4OBje3t4YMGAA8vLysGPHDjx+/Bjr169Hu3btdPuHYCpR53q5dOkSRo8ejdq1a2Po0KGwsbFBXFycfBrhNWvWwNfXV/48vl6qpiNHjmDatGkwNzfHsGHDYGNjg4SEBOzYsQNpaWlYvHgxPvjgAwB8f2Eyql4zlfkew4lGAcVNfFJQeHi4fIY2kUiEX3/9FYcOHUJiYiJq1aqFjh074osvvigydIiIsGPHDuzZswePHj2CsbExvL298fnnnyudl/7w4cPYsmULYmNjIRQK8c4772DSpEny6ixW8dS9XqKiovDzzz/j2rVrePXqFczNzeHl5YVx48ahRYsWCs/j66XqunnzJtavX4/IyEhkZWXB1NQUrq6uCAoKUligiu8vLJ+q10xlvcdwosEYY4wxneHOoIwxxhjTGU40GGOMMaYznGgwxhhjTGc40WCMMcaYznCiwRhjjDGd4USDMcYYYzrDiQZjjDHGdIYTDcYYY4zpDCcajLEKNXPmTDg7O8t/du3aVdEhlahDhw4K8SYkJFR0SIxVavoVHQBjjAHA/PnzUadOHbi4uMi3ERH+/PNPHDp0CH///TdSU1MBAPXr14erqysCAgLw7rvvlvmcY8eOxblz57Bjxw60adOm2HJSqRRdu3ZFamoqgoODYWRkhB07duDq1atlPjdj1QUnGoyxSsHX11e+LgwAZGRk4PPPP8fFixfRsmVLBAYGwtraGtnZ2YiNjcWRI0dw4sQJBAYGYv78+dDTU7+CdvDgwTh37hz27dtXYqJx4cIFPHv2DH379kX//v0BAGfPnlX7fIxVR5xoMMYqHSLCF198gYsXL2Lq1KkYM2YMBAKBQpnJkyfj008/xe7du2FnZ4cRI0aofZ53330XlpaWOH78OObMmQNTU1Ol5UJDQwHIEhPGmHq4jwZj1dzly5fh7OyMpUuXIioqChMmTIC3tzdat26NiRMnIjk5GQBw//59TJ06Fe3atUPr1q0xbtw4PH36VCcxnTlzBhEREejVqxfGjh1bJMkAAAsLC/z444+YMWOG0uaTtLQ0BAcHo2vXrnB1dYWPjw8+/fRT3L59W15GKBRi4MCBeP36NY4ePao0lpcvX+L06dOwt7eHt7e31l4jY9UFJxqMVXPR0dEAgLi4OAwdOhRCoRAffPABrK2t8eeff2L27Nk4ffo0PvzwQ2RnZ2PAgAFo2rQpzp49ixkzZugkpgMHDgAARo4cWWI5KysrjBw5Ek2bNlXYnpGRgcDAQISFhaFXr14IDg7GqFGj8O+//2LYsGG4dOmSvOyHH34IPT09ea2FsljEYjEGDRqk2YtirJriphPGqrmoqCgAwJ07d7Bnzx40b94cADBx4kR06dIFERERiImJwaZNm+Dp6QkAEIlE8PPzw7Vr15CXlwcjIyOtxvT333/D2NgYrq6uZXr+2rVrER8fj927d8PDw0O+vX///ujTpw9CQkJw8OBBAEDDhg3h6+uLs2fPIjY2Fs2aNVM41r59+2BoaIgBAwaU+fUwVp1xjQZj1Vx+jUZISIg8yQAAU1NTNGrUCBKJBNOnT5cnGQBgaGgIOzs7EBFev34t375ixQoEBQUV+1hVKSkpqFu3rtIOntnZ2cjMzCzyI5VK5WWOHj0KBwcH2NnZKZQxNjZGmzZtcPfuXWRkZMjL5/e9KFyr8ffff+PevXvw8/ODhYWF2q+DMcY1GoxVa69fv8ajR4/QuHFj+Pr6Ftn/9OlTmJubo3fv3kr3mZiYKHwAjxo1SiE5KPxYVUKhUCFxKGj06NGIjIwssj08PBw2NjbIysrCixcv8OLFC3h5eRV7jmfPnqF27doAgM6dO8PKygoHDhzA1KlTYWhoCOC/xCMwMFDt18AYk+FEg7Fq7N9//4VUKkX79u2L7EtISEBGRgb8/PxgYGCgsC87OxuPHz9G69atFbbXqlWrxMeqatCgAZ4+fQqRSCT/0M83b948ZGVlyR//+uuvuHDhgkJsANC8eXPMnj272HM0atRI/nt+p9DVq1fj9OnT6NmzJ3Jzc3H06FHY29uXmLAwxkrGTSeMVWP5zSYFJ8nKl993Q9m+mJgYEBFatmwp35aZmQlnZ2fcuXNH6WN1tG7dGmKxGFeuXCmyr2XLlvDx8ZH/1K9fX2G/iYkJAEAsFiuUK/xjZmam8LyBAwdCKBTKazFOnDiBrKwsHtLKmIY40WCsGstPJpR1uixpn7IEJSYmBvr6+nByclL6WB0DBw4EAPzyyy8gIrWea2ZmBktLSzx+/Fg+k2hBaWlpSp+X3yn04sWLSEtLw8GDB2FkZMSdQBnTECcajFVj0dHRMDAwKDLSAii5RiM/0ShYoxETEwM7Ozv5CJTCj9XRunVr9O/fH9evX8f8+fMhEomUljt69ChOnjwJPT09headXr164c2bN9i6datC+YyMDAwYMACjR49WerxBgwZBIpFg7969uHz5Mvz8/GBubq52/Iyx/3AfDcaqKZFIhAcPHsDJyalIPwgA+Oeff9CoUSOloy2ioqJQo0YNODg4yLf9+++/CqNWCj9WV3BwMIgIe/bsQUREBPr27Qs7OztIJBIkJCTgzz//xL1792BtbY2VK1fC0tJS/txPP/0U4eHhWLduHVJTU+Hl5YXU1FTs3r0bqampGD58uNJz5ncKXbt2Ld68ecOdQBnTAk40GKum7t27B7FYrLRpJDExEenp6UpnwhSJRHj48CFcXFwgFArl2//991+89957xT5Wl6GhIZYtW4aBAwciNDQUhw8fRkpKCvT09FC3bl20bNkSo0aNQu/evYskSubm5ti7dy/WrFmDM2fOICwsDMbGxvDw8EBwcHCxM3wKhUJ8+OGHWLVqFRwcHEpc/4QxphpONBirplxdXXH37l2l+xo1alTsPkNDQ3mzSj6xWIz79++jRYsWSh9rIr/zprrq1KmDefPmYd68eWo9b9KkSZg0aZLa52OMKcd9NBhjGnvw4AHEYrG8qaTwY8ZY9cU1GowxjcXExKBBgwaoU6eO0seqOH/+POrUqQMXFxc0btxYV6Fq7MyZM8jLy0NiYmJFh8LYW4ETDcaYxu7cuaOwpkjhx6pYuHAhAGDBggUYMmSIVuPTprlz5yIlJaWiw2DsrSEgdQepM8bY/+Xm5uL+/fsYP348PvvsM/Tr10/hMa94yhjjPhqMsTLbtGkTgoKC0KVLF7z//vtFHjPGGNdoMMYYY0xnuEaDMcYYYzrDiQZjjDHGdIYTDcYYY4zpDCcajDHGGNMZTjQYY4wxpjOcaDDGGGNMZzjRYIwxxpjOcKLBGGOMMZ3hRIMxxhhjOvM/6ko/VND1YhUAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 576x576 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "mybins = np.logspace(np.log(2000)/np.log(10),np.log(3500)/np.log(10),30)\n",
    "\n",
    "resmass = 2800\n",
    "\n",
    "fig = plt.figure(figsize=(8, 8))\n",
    "gs = gridspec.GridSpec(2, 1, height_ratios=[3,1]) \n",
    "plt.subplots_adjust(hspace = .001)\n",
    "\n",
    "ax0 = plt.subplot(gs[0])\n",
    "ax0.yaxis.set_ticks_position('both')\n",
    "ax0.xaxis.set_ticks_position('both')\n",
    "ax0.tick_params(direction=\"in\",which=\"both\")\n",
    "plt.xticks(fontsize=0)\n",
    "plt.yticks(fontsize=20)\n",
    "ax0.minorticks_on()\n",
    "\n",
    "nT,b,_=plt.hist(mjjs_delphes_secondhalf,weights=weights_delphes_secondhalf,bins=mybins,alpha=0.2,density=True,color=\"blue\",label=\"QCD, Delphes\")\n",
    "nT0,_,_=plt.hist(my_mjj_calib(mjjs_delphes_secondhalf),weights=weights_delphes_secondhalf,bins=mybins,density=True,color=\"black\",lw=3,histtype=\"step\",label=\"QCD, Delphes + OT\")\n",
    "nT1,_,_=plt.hist(mjjs_secondhalf,weights=weights_secondhalf,bins=mybins,density=True,alpha=0.2,color=\"orange\",label=\"QCD, Geant4\")\n",
    "\n",
    "###\n",
    "\n",
    "ww,bins = np.histogram(mjjsg_secondhalf,bins=mybins)\n",
    "ww = np.interp(mjjsg_secondhalf,0.5*(bins[:-1]+bins[1:]),ww)\n",
    "\n",
    "ww_delphes,bins = np.histogram(mjjsg_delphes_secondhalf,bins=mybins)\n",
    "ww_delphes = np.interp(mjjsg_delphes_secondhalf,0.5*(bins[:-1]+bins[1:]),ww_delphes)\n",
    "\n",
    "nt,_,_=plt.hist(mjjs_delphes_secondhalf,weights=norm.pdf(mjjsg_delphes_secondhalf,resmass,10)/ww_delphes,bins=mybins,alpha=0.7,density=True,color=\"blue\",label=\"BSM, Delphes\")\n",
    "nt0,_,_=plt.hist(my_mjj_calib(mjjs_delphes_secondhalf),weights=norm.pdf(mjjsg_delphes_secondhalf,resmass,10)/ww_delphes,bins=mybins,density=True,histtype=\"step\",ls=\":\",lw=3,color=\"black\",label=\"BSM, Delphes + OT\")\n",
    "nt1,_,_=plt.hist(mjjs_secondhalf,weights=norm.pdf(mjjsg_secondhalf,resmass,10)/ww,bins=mybins,alpha=0.7,density=True,color=\"orange\",label=\"BSM, Geant4\")\n",
    "\n",
    "###\n",
    "\n",
    "plt.ylabel(\"Normalized to Unity\",fontsize=20)\n",
    "plt.yscale(\"log\")\n",
    "plt.ylim([1e-6,1e-1])\n",
    "plt.legend(frameon=False,fontsize=15,ncol=2)\n",
    "_=plt.title(\"Pythia 8 Dijets\",loc=\"right\",fontsize=20, fontstyle=\"italic\")\n",
    "\n",
    "ax1 = plt.subplot(gs[1])\n",
    "ax1.yaxis.set_ticks_position('both')\n",
    "ax1.xaxis.set_ticks_position('both')\n",
    "ax1.tick_params(direction=\"in\",which=\"both\")\n",
    "ax1.minorticks_on()\n",
    "plt.locator_params(axis='x', nbins=6)\n",
    "plt.xlabel(\"$m_{jj}$ [GeV]\",fontsize=20)\n",
    "plt.ylabel(\"1/Target\",fontsize=20, labelpad=20)\n",
    "plt.xticks(fontsize=20)\n",
    "plt.yticks(fontsize=15)\n",
    "\n",
    "plt.axhline(y=1,linewidth=2, color='gray', alpha=0.5)  \n",
    "plt.plot(0.5*(b[1:]+b[:-1]),nT0/(nT1+0),color='black')\n",
    "plt.plot(0.5*(b[1:]+b[:-1]),nT/(nT1+0),color='blue',lw=4,alpha=0.2)\n",
    "\n",
    "bb=0.5*(b[1:]+b[:-1])\n",
    "nt[bb<2800]=0\n",
    "nt0[bb<2800]=0\n",
    "plt.plot(0.5*(b[1:]+b[:-1]),nt0/(nt1+0),color='black',ls=\":\")\n",
    "plt.plot(0.5*(b[1:]+b[:-1]),nt/(nt1+0),color='blue',lw=4,alpha=0.7)\n",
    "\n",
    "plt.ylim(0.51,1.49)\n",
    "plt.savefig(\"OT.pdf\",bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Gaussian examples"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "gauss_truth_prior = np.random.normal(0,1,20000)\n",
    "gauss_reco_prior = np.random.normal(gauss_truth_prior,2,20000)\n",
    "mycalibration = np.polyfit(gauss_reco_prior, gauss_truth_prior, 1)\n",
    "mypoly = np.poly1d(mycalibration)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(6, 6))\n",
    "gs = gridspec.GridSpec(1, 1) \n",
    "plt.subplots_adjust(hspace = .001)\n",
    "\n",
    "ax0 = plt.subplot(gs[0])\n",
    "ax0.yaxis.set_ticks_position('both')\n",
    "ax0.xaxis.set_ticks_position('both')\n",
    "ax0.tick_params(direction=\"in\",which=\"both\")\n",
    "plt.xticks(fontsize=20)\n",
    "plt.yticks(fontsize=20)\n",
    "ax0.minorticks_on()\n",
    "\n",
    "plt.scatter(gauss_reco_prior,gauss_truth_prior,label=\"Data\",alpha=0.1)\n",
    "plt.plot(np.linspace(-5,5,20),mypoly(np.linspace(-5,5,20)),label=\"Fit\",ls=\":\",color=\"black\",lw=3)\n",
    "\n",
    "plt.xlim([-5,5])\n",
    "plt.ylim([-5,5])\n",
    "plt.xlabel(\"Reconstructed $X$\",fontsize=20)\n",
    "plt.ylabel(\"True $X$\",fontsize=20)\n",
    "plt.legend(frameon=False,fontsize=15,ncol=1)\n",
    "_=plt.title(r\"Gaussian Example\",loc=\"right\",fontsize=20, fontstyle=\"italic\")\n",
    "_=plt.text(-0.5,4,r\"$(\\mu,\\sigma,\\epsilon)=(0,1,2)$\",fontsize=20)\n",
    "plt.savefig(\"Gaussian_prior.pdf\",bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "def b(x,sigma,epsilon,mu):\n",
    "    return (sigma**2/(sigma**2+epsilon**2))*x+(epsilon**2/(sigma**2+epsilon**2))*mu"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "meanvals = np.array([np.mean(mypoly(np.random.normal(xx,2,10000))) for xx in np.linspace(-5,5,20)])\n",
    "meanvals_analytic = b(np.linspace(-5,5,20),1,2,0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Start of epoch 0\n",
      "WARNING:tensorflow:Layer g_ifn_1 is casting an input tensor from dtype float64 to the layer's dtype of float32, which is new behavior in TensorFlow 2.  The layer has dtype float32 because it's dtype defaults to floatx.\n",
      "\n",
      "If you intended to run this layer in float32, you can safely ignore this warning. If in doubt, this warning is likely only an issue if you are porting a TensorFlow 1.X model to TensorFlow 2.\n",
      "\n",
      "To change all layers to have dtype float64 by default, call `tf.keras.backend.set_floatx('float64')`. To change just this layer, pass dtype='float64' to the layer constructor. If you are the author of this layer, you can disable autocasting by passing autocast=False to the base Layer constructor.\n",
      "\n",
      "WARNING:tensorflow:Layer dnn_5 is casting an input tensor from dtype float64 to the layer's dtype of float32, which is new behavior in TensorFlow 2.  The layer has dtype float32 because it's dtype defaults to floatx.\n",
      "\n",
      "If you intended to run this layer in float32, you can safely ignore this warning. If in doubt, this warning is likely only an issue if you are porting a TensorFlow 1.X model to TensorFlow 2.\n",
      "\n",
      "To change all layers to have dtype float64 by default, call `tf.keras.backend.set_floatx('float64')`. To change just this layer, pass dtype='float64' to the layer constructor. If you are the author of this layer, you can disable autocasting by passing autocast=False to the base Layer constructor.\n",
      "\n",
      "WARNING:tensorflow:Layer dnn_4 is casting an input tensor from dtype float64 to the layer's dtype of float32, which is new behavior in TensorFlow 2.  The layer has dtype float32 because it's dtype defaults to floatx.\n",
      "\n",
      "If you intended to run this layer in float32, you can safely ignore this warning. If in doubt, this warning is likely only an issue if you are porting a TensorFlow 1.X model to TensorFlow 2.\n",
      "\n",
      "To change all layers to have dtype float64 by default, call `tf.keras.backend.set_floatx('float64')`. To change just this layer, pass dtype='float64' to the layer constructor. If you are the author of this layer, you can disable autocasting by passing autocast=False to the base Layer constructor.\n",
      "\n",
      "WARNING:tensorflow:Layer dnn_6 is casting an input tensor from dtype float64 to the layer's dtype of float32, which is new behavior in TensorFlow 2.  The layer has dtype float32 because it's dtype defaults to floatx.\n",
      "\n",
      "If you intended to run this layer in float32, you can safely ignore this warning. If in doubt, this warning is likely only an issue if you are porting a TensorFlow 1.X model to TensorFlow 2.\n",
      "\n",
      "To change all layers to have dtype float64 by default, call `tf.keras.backend.set_floatx('float64')`. To change just this layer, pass dtype='float64' to the layer constructor. If you are the author of this layer, you can disable autocasting by passing autocast=False to the base Layer constructor.\n",
      "\n",
      "Pre-Train Epoch 0, Batch 0: Loss = 3731.280, a = 0.128 b = 29.705, c = -0.829, Joint = 12.870, Marginal = -0.647, MI = -57.504\n",
      "Pre-Train Epoch 0, Batch 1: Loss = 3970.220, a = 0.153 b = 26.906, c = -0.823, Joint = 18.092, Marginal = 11.156, MI = -52.077\n",
      "Pre-Train Epoch 0, Batch 2: Loss = 4240.471, a = 0.169 b = 40.861, c = -1.473, Joint = 17.448, Marginal = -0.167, MI = -17.098\n",
      "Pre-Train Epoch 0, Batch 3: Loss = 3162.339, a = 0.236 b = 32.692, c = 2.165, Joint = 15.717, Marginal = 8.891, MI = -55.997\n",
      "Pre-Train Epoch 0, Batch 4: Loss = 1934.673, a = 0.256 b = 25.618, c = -1.475, Joint = 8.976, Marginal = -1.366, MI = -30.566\n",
      "Pre-Train Epoch 0, Batch 5: Loss = 1444.271, a = 0.333 b = 28.858, c = 1.034, Joint = 9.174, Marginal = 7.264, MI = -47.097\n",
      "Pre-Train Epoch 0, Batch 6: Loss = 5143.975, a = 0.438 b = 30.008, c = -1.161, Joint = 15.491, Marginal = 1.569, MI = -58.748\n",
      "Pre-Train Epoch 0, Batch 7: Loss = 1141.326, a = 0.378 b = 19.241, c = -1.110, Joint = 5.397, Marginal = 1.825, MI = -76.563\n",
      "Pre-Train Epoch 0, Batch 8: Loss = 3695.695, a = 0.522 b = 29.847, c = -0.266, Joint = 10.270, Marginal = 6.448, MI = -52.777\n",
      "Pre-Train Epoch 0, Batch 9: Loss = 3053.901, a = 0.774 b = 26.499, c = -0.258, Joint = 5.352, Marginal = -0.222, MI = -76.526\n",
      "Pre-Train Epoch 0, Batch 10: Loss = 5684.776, a = 0.928 b = 28.386, c = -1.480, Joint = 19.641, Marginal = 4.613, MI = -45.432\n",
      "Pre-Train Epoch 0, Batch 11: Loss = 3234.787, a = 1.210 b = 30.882, c = -0.566, Joint = 11.758, Marginal = 2.125, MI = -30.791\n",
      "Pre-Train Epoch 0, Batch 12: Loss = 835.050, a = 0.748 b = 19.543, c = -0.783, Joint = 8.202, Marginal = 4.854, MI = -50.219\n",
      "Pre-Train Epoch 0, Batch 13: Loss = 2732.296, a = 1.447 b = 27.867, c = 1.423, Joint = 16.434, Marginal = 8.202, MI = -45.318\n",
      "Pre-Train Epoch 0, Batch 14: Loss = 4590.643, a = 3.169 b = 42.965, c = 0.399, Joint = 25.181, Marginal = 11.772, MI = -46.623\n",
      "Pre-Train Epoch 0, Batch 15: Loss = 1077.184, a = 0.955 b = 18.958, c = -1.485, Joint = 1.654, Marginal = -2.087, MI = -81.541\n",
      "Pre-Train Epoch 0, Batch 16: Loss = 3216.030, a = 1.719 b = 23.515, c = -1.485, Joint = 9.425, Marginal = 0.592, MI = -68.375\n",
      "Pre-Train Epoch 0, Batch 17: Loss = 2513.375, a = 2.732 b = 29.741, c = -0.306, Joint = 12.791, Marginal = 0.904, MI = -48.853\n",
      "Pre-Train Epoch 0, Batch 18: Loss = 2458.186, a = 2.112 b = 25.535, c = 0.065, Joint = 4.555, Marginal = -1.376, MI = -30.854\n",
      "Pre-Train Epoch 0, Batch 19: Loss = 1146.035, a = 2.063 b = 24.861, c = -0.028, Joint = 2.795, Marginal = -2.090, MI = -51.463\n",
      "Pre-Train Epoch 0, Batch 20: Loss = 1533.745, a = 4.325 b = 31.234, c = -0.537, Joint = 8.575, Marginal = -2.256, MI = -67.514\n",
      "Pre-Train Epoch 0, Batch 21: Loss = 1389.990, a = 3.497 b = 24.936, c = -0.095, Joint = 10.656, Marginal = 1.739, MI = -71.477\n",
      "Pre-Train Epoch 0, Batch 22: Loss = 1709.068, a = 3.568 b = 23.068, c = 1.847, Joint = 6.083, Marginal = 1.315, MI = -16.149\n",
      "Pre-Train Epoch 0, Batch 23: Loss = 11696.442, a = 5.694 b = 29.615, c = -0.414, Joint = 16.099, Marginal = 4.932, MI = -60.339\n",
      "Pre-Train Epoch 0, Batch 24: Loss = 14522.170, a = 7.846 b = 33.392, c = -0.855, Joint = 29.889, Marginal = 0.231, MI = -3.859\n",
      "Pre-Train Epoch 0, Batch 25: Loss = 3284.492, a = 8.026 b = 31.487, c = -1.212, Joint = 7.391, Marginal = 7.267, MI = -51.500\n",
      "Pre-Train Epoch 0, Batch 26: Loss = 3593.028, a = 6.043 b = 24.659, c = 1.313, Joint = 16.081, Marginal = -1.436, MI = -24.653\n",
      "Pre-Train Epoch 0, Batch 27: Loss = 1302.474, a = 5.312 b = 23.115, c = -0.699, Joint = 4.216, Marginal = 1.759, MI = -75.906\n",
      "Pre-Train Epoch 0, Batch 28: Loss = 2021.941, a = 5.089 b = 29.275, c = -1.069, Joint = 4.779, Marginal = -1.489, MI = -69.395\n",
      "Pre-Train Epoch 0, Batch 29: Loss = 21205.459, a = 13.516 b = 38.845, c = -1.496, Joint = 32.230, Marginal = 5.762, MI = -43.598\n",
      "Pre-Train Epoch 0, Batch 30: Loss = 3068.367, a = 12.530 b = 39.215, c = 0.543, Joint = 9.935, Marginal = -0.445, MI = -60.592\n",
      "Pre-Train Epoch 0, Batch 31: Loss = 1719.984, a = 6.879 b = 24.220, c = -1.498, Joint = 7.075, Marginal = 0.595, MI = -71.220\n",
      "Pre-Train Epoch 0, Batch 32: Loss = 3132.424, a = 11.789 b = 29.974, c = -0.680, Joint = 9.688, Marginal = 2.693, MI = -13.731\n",
      "Pre-Train Epoch 0, Batch 33: Loss = 1963.083, a = 10.173 b = 27.506, c = -1.500, Joint = 9.917, Marginal = -0.441, MI = -62.827\n",
      "Pre-Train Epoch 0, Batch 34: Loss = 1591.727, a = 11.612 b = 27.229, c = 0.009, Joint = 6.515, Marginal = 5.919, MI = -55.458\n",
      "Pre-Train Epoch 0, Batch 35: Loss = 2711.760, a = 14.620 b = 28.043, c = 0.577, Joint = 8.401, Marginal = -0.674, MI = -37.848\n",
      "Pre-Train Epoch 0, Batch 36: Loss = 1436.098, a = 11.270 b = 24.269, c = -0.238, Joint = 5.910, Marginal = 2.222, MI = -72.801\n",
      "Pre-Train Epoch 0, Batch 37: Loss = 3548.399, a = 19.800 b = 35.695, c = 2.831, Joint = 11.522, Marginal = 6.804, MI = -66.196\n",
      "Pre-Train Epoch 0, Batch 38: Loss = 5849.816, a = 20.332 b = 30.678, c = -1.504, Joint = 9.292, Marginal = 3.977, MI = -43.346\n",
      "Pre-Train Epoch 0, Batch 39: Loss = 1925.323, a = 23.160 b = 36.362, c = -1.504, Joint = 9.739, Marginal = 5.845, MI = -56.952\n",
      "Pre-Train Epoch 0, Batch 40: Loss = 1727.806, a = 14.946 b = 24.075, c = -1.374, Joint = 6.473, Marginal = -2.666, MI = -52.205\n",
      "Pre-Train Epoch 0, Batch 41: Loss = 1834.229, a = 14.381 b = 25.478, c = -1.506, Joint = 6.369, Marginal = 2.117, MI = -52.847\n",
      "Pre-Train Epoch 0, Batch 42: Loss = 1453.215, a = 18.324 b = 28.982, c = 1.773, Joint = 3.113, Marginal = 1.662, MI = -77.289\n",
      "Pre-Train Epoch 0, Batch 43: Loss = 1682.731, a = 17.813 b = 23.796, c = 0.193, Joint = 7.828, Marginal = -1.297, MI = -74.527\n",
      "Pre-Train Epoch 0, Batch 44: Loss = 629.401, a = 16.499 b = 22.226, c = -1.508, Joint = -0.115, Marginal = 0.147, MI = -49.658\n",
      "Pre-Train Epoch 0, Batch 45: Loss = 2591.027, a = 23.910 b = 28.003, c = 1.608, Joint = 10.533, Marginal = 5.883, MI = -70.338\n",
      "Pre-Train Epoch 0, Batch 46: Loss = 1357.053, a = 26.599 b = 26.991, c = -0.325, Joint = 7.515, Marginal = 0.974, MI = -41.633\n",
      "Pre-Train Epoch 0, Batch 47: Loss = 2057.327, a = 27.448 b = 28.375, c = -1.510, Joint = 7.177, Marginal = -5.144, MI = -34.713\n",
      "Pre-Train Epoch 0, Batch 48: Loss = 5977.450, a = 38.704 b = 30.998, c = -1.510, Joint = 17.191, Marginal = 11.295, MI = -65.416\n",
      "Pre-Train Epoch 0, Batch 49: Loss = 472.572, a = 18.193 b = 18.733, c = -1.511, Joint = 3.258, Marginal = -1.377, MI = -27.976\n",
      "Pre-Train Epoch 0, Batch 50: Loss = 2494.106, a = 21.120 b = 22.873, c = -1.512, Joint = 9.071, Marginal = 1.228, MI = -74.196\n",
      "Pre-Train Epoch 0, Batch 51: Loss = 954.457, a = 21.359 b = 20.892, c = -1.512, Joint = 3.529, Marginal = -1.367, MI = -41.036\n",
      "Pre-Train Epoch 0, Batch 52: Loss = 1108.788, a = 35.780 b = 28.422, c = -1.513, Joint = 3.082, Marginal = -2.397, MI = -40.110\n",
      "Pre-Train Epoch 0, Batch 53: Loss = 8690.083, a = 39.127 b = 29.296, c = -1.514, Joint = 15.861, Marginal = -0.727, MI = -46.629\n",
      "Pre-Train Epoch 0, Batch 54: Loss = 919.483, a = 31.590 b = 25.136, c = -1.478, Joint = 3.425, Marginal = -1.056, MI = -50.266\n",
      "Pre-Train Epoch 0, Batch 55: Loss = 3624.267, a = 44.236 b = 29.854, c = 1.104, Joint = 12.577, Marginal = -6.177, MI = -64.061\n",
      "Pre-Train Epoch 0, Batch 56: Loss = 785.886, a = 31.583 b = 24.023, c = -1.516, Joint = -3.901, Marginal = -0.120, MI = -76.729\n",
      "Pre-Train Epoch 0, Batch 57: Loss = 12841.248, a = 51.960 b = 30.727, c = -0.356, Joint = 15.696, Marginal = 16.078, MI = -56.678\n",
      "Pre-Train Epoch 0, Batch 58: Loss = 2017.824, a = 51.064 b = 32.515, c = 0.095, Joint = 7.634, Marginal = -2.657, MI = -22.457\n",
      "Pre-Train Epoch 0, Batch 59: Loss = 5141.310, a = 63.832 b = 33.814, c = -0.481, Joint = 13.471, Marginal = 0.234, MI = -65.314\n",
      "Pre-Train Epoch 0, Batch 60: Loss = 6902.604, a = 51.838 b = 29.351, c = -0.432, Joint = 17.674, Marginal = 13.244, MI = -23.229\n",
      "Pre-Train Epoch 0, Batch 61: Loss = 1296.561, a = 52.683 b = 28.133, c = 0.396, Joint = 0.253, Marginal = -4.986, MI = -77.658\n",
      "Pre-Train Epoch 0, Batch 62: Loss = 1639.723, a = 52.274 b = 25.206, c = -1.358, Joint = 6.036, Marginal = -0.996, MI = -45.334\n",
      "Pre-Train Epoch 0, Batch 63: Loss = 1355.089, a = 61.343 b = 31.362, c = -0.692, Joint = -0.895, Marginal = -2.393, MI = -41.691\n",
      "Pre-Train Epoch 0, Batch 64: Loss = 4056.738, a = 51.562 b = 25.779, c = 0.096, Joint = 11.521, Marginal = 4.223, MI = -64.498\n",
      "Pre-Train Epoch 0, Batch 65: Loss = 4499.939, a = 79.102 b = 34.800, c = -1.522, Joint = 9.100, Marginal = -3.102, MI = -41.274\n",
      "Pre-Train Epoch 0, Batch 66: Loss = 5164.141, a = 84.230 b = 32.187, c = 2.323, Joint = 10.345, Marginal = -0.194, MI = -50.762\n",
      "Pre-Train Epoch 0, Batch 67: Loss = 1675.276, a = 52.407 b = 30.308, c = -1.524, Joint = 1.149, Marginal = -5.807, MI = -79.486\n",
      "Pre-Train Epoch 0, Batch 68: Loss = 1602.103, a = 60.133 b = 27.538, c = -0.564, Joint = 6.341, Marginal = 0.343, MI = -31.080\n",
      "Pre-Train Epoch 0, Batch 69: Loss = 4452.496, a = 76.109 b = 30.144, c = -1.525, Joint = 14.312, Marginal = 2.285, MI = -60.782\n",
      "Pre-Train Epoch 0, Batch 70: Loss = 1486.268, a = 48.004 b = 22.024, c = -1.526, Joint = 6.936, Marginal = -4.671, MI = -28.193\n",
      "Pre-Train Epoch 0, Batch 71: Loss = 1152.060, a = 58.976 b = 24.585, c = 2.568, Joint = -1.017, Marginal = -4.495, MI = -40.840\n",
      "Pre-Train Epoch 0, Batch 72: Loss = 2192.823, a = 58.355 b = 24.821, c = 0.170, Joint = 3.966, Marginal = -7.146, MI = -7.279\n",
      "Pre-Train Epoch 0, Batch 73: Loss = 2681.811, a = 75.948 b = 27.178, c = -1.528, Joint = 7.994, Marginal = -3.298, MI = -28.124\n",
      "Pre-Train Epoch 0, Batch 74: Loss = 463.402, a = 45.341 b = 19.001, c = -1.528, Joint = -2.566, Marginal = -6.418, MI = -17.713\n",
      "Pre-Train Epoch 0, Batch 75: Loss = 1403.562, a = 111.520 b = 30.486, c = -0.384, Joint = -0.072, Marginal = -1.056, MI = -49.779\n",
      "Pre-Train Epoch 0, Batch 76: Loss = 2376.769, a = 124.201 b = 31.799, c = -1.530, Joint = 3.932, Marginal = 10.203, MI = -73.346\n",
      "Pre-Train Epoch 0, Batch 77: Loss = 2313.142, a = 93.367 b = 29.903, c = -1.530, Joint = 0.932, Marginal = -0.371, MI = -72.611\n",
      "Pre-Train Epoch 0, Batch 78: Loss = 1121.899, a = 95.168 b = 26.253, c = -1.531, Joint = 1.728, Marginal = -3.250, MI = -31.163\n",
      "Pre-Train Epoch 0, Batch 79: Loss = 689.912, a = 110.522 b = 29.735, c = 0.605, Joint = -1.282, Marginal = -7.245, MI = -33.284\n",
      "Pre-Train Epoch 0, Batch 80: Loss = 1938.698, a = 111.633 b = 30.115, c = -1.532, Joint = 6.492, Marginal = -0.827, MI = -59.943\n",
      "Pre-Train Epoch 0, Batch 81: Loss = 2126.046, a = 82.515 b = 25.287, c = -1.435, Joint = 3.461, Marginal = -2.959, MI = -35.428\n",
      "Pre-Train Epoch 0, Batch 82: Loss = 1226.533, a = 108.868 b = 27.656, c = -1.409, Joint = 0.205, Marginal = -1.517, MI = -81.620\n",
      "Pre-Train Epoch 0, Batch 83: Loss = 7487.525, a = 98.395 b = 26.169, c = 0.053, Joint = 11.587, Marginal = -1.724, MI = 2.896\n",
      "Pre-Train Epoch 0, Batch 84: Loss = 2516.946, a = 128.912 b = 27.102, c = -0.391, Joint = 7.984, Marginal = 3.891, MI = -39.779\n",
      "Pre-Train Epoch 0, Batch 85: Loss = 2532.988, a = 128.224 b = 28.747, c = -0.749, Joint = 8.107, Marginal = 0.730, MI = -48.333\n",
      "Pre-Train Epoch 0, Batch 86: Loss = 1050.165, a = 132.463 b = 29.314, c = -1.535, Joint = 1.910, Marginal = -4.996, MI = -46.791\n",
      "Pre-Train Epoch 0, Batch 87: Loss = 1467.817, a = 89.054 b = 24.484, c = 0.946, Joint = 3.033, Marginal = -5.365, MI = -33.785\n",
      "Pre-Train Epoch 0, Batch 88: Loss = 502.329, a = 114.312 b = 22.643, c = 0.211, Joint = -1.722, Marginal = -3.786, MI = -85.185\n",
      "Pre-Train Epoch 0, Batch 89: Loss = 2977.049, a = 175.240 b = 33.464, c = -1.536, Joint = 8.648, Marginal = -5.171, MI = -75.257\n",
      "Pre-Train Epoch 0, Batch 90: Loss = 514.806, a = 148.218 b = 28.409, c = -0.735, Joint = -3.870, Marginal = -0.567, MI = -47.204\n",
      "Pre-Train Epoch 0, Batch 91: Loss = 1097.897, a = 94.234 b = 20.759, c = -1.537, Joint = 2.334, Marginal = 1.327, MI = -71.428\n",
      "Pre-Train Epoch 0, Batch 92: Loss = 3468.554, a = 170.100 b = 30.948, c = -0.656, Joint = 9.125, Marginal = -4.221, MI = -47.625\n",
      "Pre-Train Epoch 0, Batch 93: Loss = 11377.982, a = 147.004 b = 30.098, c = 0.858, Joint = 10.746, Marginal = -0.082, MI = -50.308\n",
      "Pre-Train Epoch 0, Batch 94: Loss = 2057.430, a = 135.372 b = 28.243, c = -0.605, Joint = 4.226, Marginal = -2.478, MI = -57.875\n",
      "Pre-Train Epoch 0, Batch 95: Loss = 1240.044, a = 209.393 b = 30.748, c = -1.540, Joint = -1.854, Marginal = -6.913, MI = -63.413\n",
      "Pre-Train Epoch 0, Batch 96: Loss = 765.748, a = 159.598 b = 26.497, c = 1.304, Joint = -1.414, Marginal = -5.163, MI = -23.800\n",
      "Pre-Train Epoch 0, Batch 97: Loss = 646.369, a = 110.457 b = 21.519, c = -0.348, Joint = 0.070, Marginal = -4.488, MI = -68.528\n",
      "Pre-Train Epoch 0, Batch 98: Loss = 1288.971, a = 156.853 b = 24.188, c = -1.385, Joint = 2.202, Marginal = -4.519, MI = -62.672\n",
      "Pre-Train Epoch 0, Batch 99: Loss = 2493.860, a = 162.111 b = 25.918, c = 1.055, Joint = 5.796, Marginal = -2.403, MI = -75.983\n",
      "Pre-Train Epoch 0, Batch 100: Loss = 4941.496, a = 213.320 b = 30.587, c = -0.314, Joint = 13.166, Marginal = -2.983, MI = -52.731\n",
      "Pre-Train Epoch 0, Batch 101: Loss = 483.332, a = 116.540 b = 22.060, c = 1.554, Joint = -5.327, Marginal = -6.866, MI = -19.861\n",
      "Pre-Train Epoch 0, Batch 102: Loss = 1064.654, a = 161.517 b = 23.927, c = -1.544, Joint = -3.010, Marginal = -13.774, MI = -21.596\n",
      "Pre-Train Epoch 0, Batch 103: Loss = 1830.420, a = 172.652 b = 28.913, c = -0.228, Joint = 4.125, Marginal = -5.085, MI = -28.341\n",
      "Pre-Train Epoch 0, Batch 104: Loss = 562.344, a = 134.522 b = 24.622, c = -1.050, Joint = -9.574, Marginal = -10.538, MI = -72.689\n",
      "Pre-Train Epoch 0, Batch 105: Loss = 531.644, a = 133.986 b = 22.838, c = 0.618, Joint = -5.844, Marginal = -4.342, MI = -38.761\n",
      "Pre-Train Epoch 0, Batch 106: Loss = 955.243, a = 162.585 b = 27.226, c = 1.052, Joint = -4.675, Marginal = -6.868, MI = -59.344\n",
      "Pre-Train Epoch 0, Batch 107: Loss = 2437.212, a = 148.610 b = 23.345, c = -0.831, Joint = 1.401, Marginal = -10.242, MI = -27.453\n",
      "Pre-Train Epoch 0, Batch 108: Loss = 2383.618, a = 135.926 b = 25.341, c = -1.546, Joint = 0.761, Marginal = -1.526, MI = -70.410\n",
      "Pre-Train Epoch 0, Batch 109: Loss = 406.287, a = 115.107 b = 22.455, c = 1.240, Joint = -7.585, Marginal = -10.484, MI = -9.136\n",
      "Pre-Train Epoch 0, Batch 110: Loss = 1664.436, a = 109.913 b = 20.482, c = -0.709, Joint = 2.154, Marginal = -8.923, MI = -22.623\n",
      "Pre-Train Epoch 0, Batch 111: Loss = 1369.586, a = 110.349 b = 26.835, c = -1.050, Joint = -9.730, Marginal = -12.565, MI = -29.549\n",
      "Pre-Train Epoch 0, Batch 112: Loss = 859.106, a = 188.135 b = 28.845, c = 1.445, Joint = -0.112, Marginal = -5.614, MI = -52.363\n",
      "Pre-Train Epoch 0, Batch 113: Loss = 635.678, a = 104.878 b = 24.595, c = -0.150, Joint = -6.419, Marginal = -8.374, MI = -77.112\n",
      "Pre-Train Epoch 0, Batch 114: Loss = 1429.508, a = 143.142 b = 24.755, c = -0.352, Joint = 0.580, Marginal = -6.939, MI = -44.878\n",
      "Pre-Train Epoch 0, Batch 115: Loss = 1503.060, a = 182.530 b = 30.280, c = 0.852, Joint = 1.138, Marginal = -0.309, MI = -33.356\n",
      "Pre-Train Epoch 0, Batch 116: Loss = 1950.562, a = 146.703 b = 24.087, c = -1.549, Joint = 1.333, Marginal = -6.175, MI = -55.736\n",
      "Pre-Train Epoch 0, Batch 117: Loss = 1586.026, a = 177.613 b = 28.409, c = 0.505, Joint = 0.872, Marginal = 0.390, MI = -29.304\n",
      "Pre-Train Epoch 0, Batch 118: Loss = 473.936, a = 113.177 b = 19.748, c = -1.549, Joint = -4.096, Marginal = -4.104, MI = -53.143\n",
      "Pre-Train Epoch 0, Batch 119: Loss = 2019.591, a = 192.839 b = 26.318, c = -1.550, Joint = 0.583, Marginal = 0.504, MI = -46.479\n",
      "Pre-Train Epoch 0, Batch 120: Loss = 2031.560, a = 199.874 b = 33.140, c = -1.550, Joint = 6.182, Marginal = -6.015, MI = -44.244\n",
      "Pre-Train Epoch 0, Batch 121: Loss = 1948.197, a = 193.538 b = 31.490, c = 1.167, Joint = 0.252, Marginal = -8.381, MI = -54.594\n",
      "Pre-Train Epoch 0, Batch 122: Loss = 599.107, a = 152.644 b = 24.518, c = -1.346, Joint = -4.949, Marginal = -8.191, MI = -86.089\n",
      "Pre-Train Epoch 0, Batch 123: Loss = 2718.656, a = 241.908 b = 32.707, c = -1.551, Joint = 6.518, Marginal = -4.410, MI = -73.861\n",
      "Pre-Train Epoch 0, Batch 124: Loss = 551.872, a = 116.214 b = 22.372, c = -1.551, Joint = -9.010, Marginal = -8.347, MI = -32.872\n",
      "Pre-Train Epoch 0, Batch 125: Loss = 2450.791, a = 220.132 b = 28.317, c = -1.080, Joint = 5.982, Marginal = 0.146, MI = -22.341\n",
      "Pre-Train Epoch 0, Batch 126: Loss = 420.390, a = 126.434 b = 21.715, c = -0.170, Joint = -5.126, Marginal = -6.553, MI = -35.539\n",
      "Pre-Train Epoch 0, Batch 127: Loss = 1706.654, a = 249.839 b = 31.138, c = -1.552, Joint = -2.583, Marginal = -9.710, MI = -68.661\n",
      "Pre-Train Epoch 0, Batch 128: Loss = 2726.551, a = 233.325 b = 30.271, c = -1.012, Joint = 6.333, Marginal = -4.087, MI = -53.348\n",
      "Pre-Train Epoch 0, Batch 129: Loss = 1643.069, a = 158.840 b = 25.902, c = -1.553, Joint = 2.866, Marginal = -8.206, MI = -23.310\n",
      "Pre-Train Epoch 0, Batch 130: Loss = 979.415, a = 136.267 b = 24.394, c = -1.553, Joint = -3.695, Marginal = -10.358, MI = -34.250\n",
      "Pre-Train Epoch 0, Batch 131: Loss = 662.357, a = 96.034 b = 18.271, c = -0.332, Joint = -1.202, Marginal = -3.040, MI = -82.620\n",
      "Pre-Train Epoch 0, Batch 132: Loss = 871.421, a = 129.770 b = 23.832, c = -1.554, Joint = -2.916, Marginal = -5.066, MI = -46.393\n",
      "Pre-Train Epoch 0, Batch 133: Loss = 829.104, a = 194.957 b = 28.753, c = -0.602, Joint = -2.946, Marginal = -10.249, MI = -34.648\n",
      "Pre-Train Epoch 0, Batch 134: Loss = 1612.365, a = 132.123 b = 22.302, c = -1.555, Joint = 0.263, Marginal = -5.523, MI = -44.571\n",
      "Pre-Train Epoch 0, Batch 135: Loss = 1889.594, a = 162.423 b = 31.325, c = 0.498, Joint = -1.600, Marginal = -10.668, MI = -25.384\n",
      "Pre-Train Epoch 0, Batch 136: Loss = 1203.840, a = 160.481 b = 25.678, c = -1.082, Joint = -1.204, Marginal = -7.421, MI = -35.210\n",
      "Pre-Train Epoch 0, Batch 137: Loss = 4076.259, a = 203.936 b = 32.771, c = -0.810, Joint = 4.618, Marginal = -7.682, MI = -42.240\n",
      "Pre-Train Epoch 0, Batch 138: Loss = 958.846, a = 152.471 b = 25.511, c = -1.556, Joint = -6.840, Marginal = -10.829, MI = -35.814\n",
      "Pre-Train Epoch 0, Batch 139: Loss = 645.749, a = 139.281 b = 24.098, c = -0.518, Joint = -1.806, Marginal = -6.367, MI = -35.460\n",
      "Pre-Train Epoch 0, Batch 140: Loss = 2068.440, a = 105.879 b = 26.853, c = -1.224, Joint = -5.337, Marginal = -10.194, MI = -62.151\n",
      "Pre-Train Epoch 0, Batch 141: Loss = 524.510, a = 121.516 b = 23.785, c = -1.557, Joint = -5.444, Marginal = -9.550, MI = -19.618\n",
      "Pre-Train Epoch 0, Batch 142: Loss = 1012.802, a = 103.586 b = 22.460, c = -1.105, Joint = -1.374, Marginal = -3.356, MI = -75.504\n",
      "Pre-Train Epoch 0, Batch 143: Loss = 1368.909, a = 99.227 b = 22.356, c = -1.558, Joint = -3.430, Marginal = -8.793, MI = -22.051\n",
      "Pre-Train Epoch 0, Batch 144: Loss = 1106.451, a = 218.462 b = 29.893, c = -1.519, Joint = -4.611, Marginal = -5.715, MI = -85.349\n",
      "Pre-Train Epoch 0, Batch 145: Loss = 504.647, a = 92.406 b = 18.420, c = 0.461, Joint = -6.392, Marginal = -8.358, MI = -35.688\n",
      "Pre-Train Epoch 0, Batch 146: Loss = 3255.154, a = 273.575 b = 38.471, c = -0.531, Joint = 9.662, Marginal = 3.093, MI = -60.835\n",
      "Pre-Train Epoch 0, Batch 147: Loss = 3363.836, a = 166.268 b = 27.899, c = -1.559, Joint = 4.136, Marginal = -5.797, MI = -21.119\n",
      "Pre-Train Epoch 0, Batch 148: Loss = 981.240, a = 160.051 b = 27.705, c = 3.185, Joint = 2.220, Marginal = -4.788, MI = -59.557\n",
      "Pre-Train Epoch 0, Batch 149: Loss = 909.114, a = 114.083 b = 21.316, c = -0.922, Joint = -2.836, Marginal = -5.422, MI = -8.202\n",
      "Pre-Train Epoch 0, Batch 150: Loss = 628.723, a = 127.982 b = 22.518, c = -0.473, Joint = -5.464, Marginal = -8.253, MI = -31.929\n",
      "Pre-Train Epoch 0, Batch 151: Loss = 1123.096, a = 221.775 b = 29.252, c = -1.560, Joint = -3.287, Marginal = -8.683, MI = -22.285\n",
      "Pre-Train Epoch 0, Batch 152: Loss = 1414.019, a = 185.456 b = 31.049, c = 0.496, Joint = -3.821, Marginal = -8.045, MI = -21.310\n",
      "Pre-Train Epoch 0, Batch 153: Loss = 378.136, a = 129.159 b = 18.318, c = 1.428, Joint = -2.308, Marginal = -5.182, MI = -73.910\n",
      "Pre-Train Epoch 0, Batch 154: Loss = 626.876, a = 127.324 b = 22.763, c = -0.064, Joint = -3.420, Marginal = -7.436, MI = -61.418\n",
      "Pre-Train Epoch 0, Batch 155: Loss = 596.750, a = 114.213 b = 21.113, c = -0.356, Joint = -6.363, Marginal = -7.681, MI = -61.887\n",
      "Pre-Train Epoch 0, Batch 156: Loss = 1101.727, a = 173.379 b = 29.611, c = -1.561, Joint = -4.509, Marginal = -8.139, MI = -31.896\n",
      "Pre-Train Epoch 0, Batch 157: Loss = 2307.132, a = 217.301 b = 31.144, c = -1.561, Joint = 4.875, Marginal = -6.563, MI = -9.619\n",
      "Pre-Train Epoch 0, Batch 158: Loss = 572.933, a = 126.053 b = 22.601, c = -0.793, Joint = -5.777, Marginal = -9.012, MI = -26.614\n",
      "Pre-Train Epoch 0, Batch 159: Loss = 1246.887, a = 157.297 b = 26.634, c = -0.354, Joint = -1.550, Marginal = -7.502, MI = -69.731\n",
      "Pre-Train Epoch 0, Batch 160: Loss = 1275.960, a = 177.415 b = 26.723, c = -1.562, Joint = -0.341, Marginal = -9.936, MI = -28.687\n",
      "Pre-Train Epoch 0, Batch 161: Loss = 449.893, a = 128.058 b = 21.899, c = 0.721, Joint = -4.616, Marginal = -9.498, MI = -11.656\n",
      "Pre-Train Epoch 0, Batch 162: Loss = 302.187, a = 81.584 b = 20.622, c = -0.551, Joint = -4.418, Marginal = -4.342, MI = -47.919\n",
      "Pre-Train Epoch 0, Batch 163: Loss = 838.521, a = 109.014 b = 20.643, c = -1.563, Joint = 2.840, Marginal = -4.054, MI = -58.949\n",
      "Pre-Train Epoch 0, Batch 164: Loss = 1907.750, a = 171.831 b = 25.100, c = -0.640, Joint = 0.985, Marginal = -2.146, MI = -27.347\n",
      "Pre-Train Epoch 0, Batch 165: Loss = 1346.522, a = 145.588 b = 26.446, c = -1.563, Joint = -3.925, Marginal = -6.398, MI = -66.314\n",
      "Pre-Train Epoch 0, Batch 166: Loss = 402.298, a = 104.354 b = 23.997, c = -0.784, Joint = -7.607, Marginal = -7.270, MI = -61.132\n",
      "Pre-Train Epoch 0, Batch 167: Loss = 476.123, a = 61.767 b = 16.734, c = 1.467, Joint = -5.025, Marginal = -7.467, MI = -24.341\n",
      "Pre-Train Epoch 0, Batch 168: Loss = 948.075, a = 57.395 b = 15.102, c = -0.667, Joint = 1.954, Marginal = -4.953, MI = -11.741\n",
      "Pre-Train Epoch 0, Batch 169: Loss = 746.033, a = 135.966 b = 23.382, c = -0.787, Joint = -1.056, Marginal = -2.101, MI = -82.715\n",
      "Pre-Train Epoch 0, Batch 170: Loss = 394.563, a = 92.993 b = 20.008, c = -1.564, Joint = -3.809, Marginal = -6.354, MI = -76.327\n",
      "Pre-Train Epoch 0, Batch 171: Loss = 811.050, a = 122.874 b = 24.093, c = -1.565, Joint = 0.423, Marginal = -6.646, MI = -40.175\n",
      "Pre-Train Epoch 0, Batch 172: Loss = 1156.064, a = 147.857 b = 27.834, c = -0.857, Joint = -3.897, Marginal = -8.422, MI = -49.394\n",
      "Pre-Train Epoch 0, Batch 173: Loss = 1728.077, a = 157.570 b = 27.303, c = 0.245, Joint = 4.042, Marginal = -8.222, MI = -16.595\n",
      "Pre-Train Epoch 0, Batch 174: Loss = 1020.610, a = 134.670 b = 24.030, c = 0.253, Joint = -1.480, Marginal = -5.358, MI = -49.786\n",
      "Pre-Train Epoch 0, Batch 175: Loss = 373.456, a = 100.866 b = 21.663, c = -1.565, Joint = -3.215, Marginal = -8.449, MI = -27.436\n",
      "Pre-Train Epoch 0, Batch 176: Loss = 1621.431, a = 128.484 b = 24.270, c = -1.566, Joint = 3.720, Marginal = -5.254, MI = -29.727\n",
      "Pre-Train Epoch 0, Batch 177: Loss = 559.520, a = 149.175 b = 25.404, c = 1.131, Joint = -4.922, Marginal = -7.266, MI = -27.660\n",
      "Pre-Train Epoch 0, Batch 178: Loss = 2659.451, a = 122.191 b = 30.213, c = -1.566, Joint = -2.778, Marginal = -3.942, MI = -68.134\n",
      "Pre-Train Epoch 0, Batch 179: Loss = 3599.714, a = 187.613 b = 30.373, c = -0.388, Joint = 8.745, Marginal = -1.743, MI = -61.242\n",
      "Pre-Train Epoch 0, Batch 180: Loss = 587.122, a = 187.170 b = 29.758, c = -1.567, Joint = -2.640, Marginal = -3.689, MI = -55.841\n",
      "Pre-Train Epoch 0, Batch 181: Loss = 1202.549, a = 102.849 b = 22.269, c = -0.780, Joint = 0.860, Marginal = -2.947, MI = -44.140\n",
      "Pre-Train Epoch 0, Batch 182: Loss = 1497.373, a = 129.793 b = 24.092, c = 2.515, Joint = 3.499, Marginal = -6.362, MI = -73.988\n",
      "Pre-Train Epoch 0, Batch 183: Loss = 443.741, a = 70.678 b = 20.007, c = -1.568, Joint = -2.841, Marginal = -6.355, MI = -37.635\n",
      "Pre-Train Epoch 0, Batch 184: Loss = 751.294, a = 143.557 b = 26.659, c = -0.961, Joint = 0.818, Marginal = -8.986, MI = -38.300\n",
      "Pre-Train Epoch 0, Batch 185: Loss = 846.780, a = 92.328 b = 21.750, c = -0.546, Joint = -3.455, Marginal = -5.884, MI = -83.027\n",
      "Pre-Train Epoch 0, Batch 186: Loss = 1311.185, a = 147.180 b = 26.087, c = -0.042, Joint = -0.686, Marginal = -2.457, MI = -56.643\n",
      "Pre-Train Epoch 0, Batch 187: Loss = 1382.337, a = 84.781 b = 19.445, c = 0.222, Joint = -1.328, Marginal = -6.652, MI = -24.062\n",
      "Pre-Train Epoch 0, Batch 188: Loss = 1891.023, a = 102.862 b = 24.694, c = -1.570, Joint = 7.885, Marginal = -2.125, MI = -33.529\n",
      "Pre-Train Epoch 0, Batch 189: Loss = 674.547, a = 107.356 b = 24.174, c = 0.073, Joint = 1.453, Marginal = -3.986, MI = -29.627\n",
      "Pre-Train Epoch 0, Batch 190: Loss = 1032.838, a = 96.977 b = 27.870, c = -1.570, Joint = -3.270, Marginal = -9.097, MI = -81.647\n",
      "Pre-Train Epoch 0, Batch 191: Loss = 1373.671, a = 98.582 b = 25.147, c = -1.571, Joint = -1.748, Marginal = -6.798, MI = -58.316\n",
      "Pre-Train Epoch 0, Batch 192: Loss = 420.725, a = 76.613 b = 19.111, c = -0.653, Joint = -2.980, Marginal = -8.423, MI = -37.437\n",
      "Pre-Train Epoch 0, Batch 193: Loss = 625.386, a = 104.464 b = 24.717, c = 0.393, Joint = -4.050, Marginal = -8.688, MI = -71.495\n",
      "Pre-Train Epoch 0, Batch 194: Loss = 374.463, a = 78.069 b = 18.124, c = -1.571, Joint = -1.077, Marginal = -5.095, MI = -82.399\n",
      "Pre-Train Epoch 0, Batch 195: Loss = 1083.316, a = 101.156 b = 23.149, c = -1.572, Joint = -0.751, Marginal = -9.804, MI = -12.327\n",
      "Pre-Train Epoch 0, Batch 196: Loss = 891.551, a = 88.652 b = 20.916, c = 0.135, Joint = -3.860, Marginal = -8.550, MI = -73.744\n",
      "Pre-Train Epoch 0, Batch 197: Loss = 322.187, a = 50.402 b = 15.843, c = -0.200, Joint = -4.677, Marginal = -7.561, MI = -24.689\n",
      "Pre-Train Epoch 0, Batch 198: Loss = 306.621, a = 72.695 b = 17.891, c = -1.572, Joint = -2.008, Marginal = -3.608, MI = -67.770\n",
      "Pre-Train Epoch 0, Batch 199: Loss = 469.855, a = 91.025 b = 22.338, c = -0.833, Joint = -2.726, Marginal = -9.556, MI = -19.635\n",
      "Pre-Train Epoch 0, Batch 200: Loss = 888.206, a = 107.170 b = 24.438, c = -0.225, Joint = -2.109, Marginal = -7.510, MI = -13.187\n",
      "Pre-Train Epoch 0, Batch 201: Loss = 3242.749, a = 122.235 b = 26.322, c = -0.754, Joint = 5.252, Marginal = -3.487, MI = -54.465\n",
      "Pre-Train Epoch 0, Batch 202: Loss = 1181.587, a = 130.606 b = 22.070, c = -1.573, Joint = -3.067, Marginal = -5.543, MI = -55.839\n",
      "Pre-Train Epoch 0, Batch 203: Loss = 849.581, a = 140.338 b = 27.226, c = -0.226, Joint = -3.632, Marginal = -11.240, MI = -70.988\n",
      "Pre-Train Epoch 0, Batch 204: Loss = 226.493, a = 69.054 b = 16.698, c = -1.573, Joint = -4.521, Marginal = -7.334, MI = -27.941\n",
      "Pre-Train Epoch 0, Batch 205: Loss = 453.508, a = 81.455 b = 21.006, c = 0.410, Joint = -1.782, Marginal = -8.999, MI = -29.829\n",
      "Pre-Train Epoch 0, Batch 206: Loss = 594.006, a = 102.354 b = 22.123, c = -0.707, Joint = -6.942, Marginal = -7.827, MI = -90.512\n",
      "Pre-Train Epoch 0, Batch 207: Loss = 323.077, a = 69.168 b = 18.147, c = 0.260, Joint = -3.569, Marginal = -4.856, MI = -72.380\n",
      "Pre-Train Epoch 0, Batch 208: Loss = 384.158, a = 79.649 b = 18.554, c = 1.739, Joint = -2.622, Marginal = -4.058, MI = -58.581\n",
      "Pre-Train Epoch 0, Batch 209: Loss = 497.821, a = 74.628 b = 20.627, c = -1.574, Joint = 0.232, Marginal = -1.347, MI = -63.551\n",
      "Pre-Train Epoch 0, Batch 210: Loss = 1922.917, a = 130.366 b = 24.207, c = -0.066, Joint = 3.005, Marginal = -2.059, MI = -65.243\n",
      "Pre-Train Epoch 0, Batch 211: Loss = 613.873, a = 59.569 b = 20.118, c = -1.302, Joint = 3.013, Marginal = -3.838, MI = -7.635\n",
      "Pre-Train Epoch 0, Batch 212: Loss = 321.503, a = 76.655 b = 18.746, c = -1.575, Joint = -5.496, Marginal = -6.004, MI = -31.481\n",
      "Pre-Train Epoch 0, Batch 213: Loss = 1129.427, a = 112.658 b = 23.985, c = -1.510, Joint = 0.714, Marginal = -5.155, MI = -44.174\n",
      "Pre-Train Epoch 0, Batch 214: Loss = 910.041, a = 84.453 b = 23.444, c = -1.575, Joint = 1.394, Marginal = -5.756, MI = -81.378\n",
      "Pre-Train Epoch 0, Batch 215: Loss = 987.644, a = 97.524 b = 24.145, c = -1.329, Joint = -4.523, Marginal = -9.169, MI = -22.229\n",
      "Pre-Train Epoch 0, Batch 216: Loss = 2196.658, a = 90.470 b = 19.567, c = 1.037, Joint = 2.201, Marginal = -5.739, MI = -5.483\n",
      "Pre-Train Epoch 0, Batch 217: Loss = 1505.160, a = 99.992 b = 26.429, c = -1.576, Joint = 0.754, Marginal = -2.397, MI = -34.198\n",
      "Pre-Train Epoch 0, Batch 218: Loss = 926.831, a = 116.758 b = 24.243, c = -1.576, Joint = 4.267, Marginal = -1.683, MI = -42.426\n",
      "Pre-Train Epoch 0, Batch 219: Loss = 583.622, a = 100.454 b = 24.174, c = -0.027, Joint = -0.198, Marginal = -5.056, MI = -33.138\n",
      "Pre-Train Epoch 0, Batch 220: Loss = 643.208, a = 100.586 b = 25.514, c = 0.872, Joint = -1.182, Marginal = -6.749, MI = -59.057\n",
      "Pre-Train Epoch 0, Batch 221: Loss = 524.603, a = 86.384 b = 20.439, c = -1.467, Joint = -1.899, Marginal = -6.926, MI = -35.002\n",
      "Pre-Train Epoch 0, Batch 222: Loss = 5738.135, a = 129.071 b = 26.722, c = -1.577, Joint = 6.069, Marginal = -7.280, MI = -18.687\n",
      "Pre-Train Epoch 0, Batch 223: Loss = 1381.701, a = 96.122 b = 25.214, c = 1.059, Joint = 1.335, Marginal = -5.614, MI = -72.537\n",
      "Pre-Train Epoch 0, Batch 224: Loss = 458.978, a = 53.104 b = 16.843, c = 1.735, Joint = -4.453, Marginal = -7.796, MI = -9.903\n",
      "Pre-Train Epoch 0, Batch 225: Loss = 872.945, a = 116.866 b = 27.996, c = -1.207, Joint = -0.675, Marginal = -6.732, MI = -50.447\n",
      "Pre-Train Epoch 0, Batch 226: Loss = 1541.994, a = 69.748 b = 18.220, c = -1.579, Joint = 0.438, Marginal = -8.316, MI = -10.895\n",
      "Pre-Train Epoch 0, Batch 227: Loss = 1041.490, a = 78.073 b = 18.539, c = -1.579, Joint = -0.290, Marginal = -5.795, MI = -6.305\n",
      "Pre-Train Epoch 0, Batch 228: Loss = 541.324, a = 79.657 b = 18.276, c = -0.792, Joint = -1.135, Marginal = -3.645, MI = -33.545\n",
      "Pre-Train Epoch 0, Batch 229: Loss = 331.194, a = 96.543 b = 23.685, c = -1.434, Joint = -3.808, Marginal = -5.353, MI = -36.799\n",
      "Pre-Train Epoch 0, Batch 230: Loss = 3201.239, a = 72.043 b = 20.351, c = -1.580, Joint = 1.080, Marginal = -4.762, MI = -36.859\n",
      "Pre-Train Epoch 0, Batch 231: Loss = 818.279, a = 146.422 b = 28.180, c = -1.581, Joint = -4.563, Marginal = -5.102, MI = -68.842\n",
      "Pre-Train Epoch 0, Batch 232: Loss = 800.270, a = 115.391 b = 25.003, c = -0.187, Joint = -2.805, Marginal = -6.825, MI = -66.912\n",
      "Pre-Train Epoch 0, Batch 233: Loss = 1018.836, a = 85.033 b = 23.102, c = -1.094, Joint = 3.327, Marginal = -2.268, MI = -52.428\n",
      "Pre-Train Epoch 0, Batch 234: Loss = 2175.025, a = 108.211 b = 27.024, c = -1.006, Joint = 3.664, Marginal = -7.239, MI = -60.176\n",
      "Pre-Train Epoch 0, Batch 235: Loss = 560.193, a = 45.642 b = 17.105, c = -1.303, Joint = -5.002, Marginal = -6.735, MI = -32.637\n",
      "Pre-Train Epoch 0, Batch 236: Loss = 1223.076, a = 75.249 b = 17.887, c = -1.582, Joint = 1.832, Marginal = -6.296, MI = -34.419\n",
      "Pre-Train Epoch 0, Batch 237: Loss = 441.570, a = 73.105 b = 19.489, c = -1.583, Joint = -0.048, Marginal = -4.236, MI = -41.087\n",
      "Pre-Train Epoch 0, Batch 238: Loss = 427.208, a = 83.257 b = 21.443, c = -1.583, Joint = -0.380, Marginal = -3.671, MI = -28.484\n",
      "Pre-Train Epoch 0, Batch 239: Loss = 647.335, a = 84.120 b = 23.813, c = -1.583, Joint = -2.113, Marginal = -7.779, MI = -23.976\n",
      "Pre-Train Epoch 0, Batch 240: Loss = 677.173, a = 99.918 b = 22.548, c = 0.332, Joint = -0.924, Marginal = -2.430, MI = -53.430\n",
      "Pre-Train Epoch 0, Batch 241: Loss = 591.889, a = 92.374 b = 24.954, c = 0.620, Joint = -2.884, Marginal = -5.397, MI = -33.790\n",
      "Pre-Train Epoch 0, Batch 242: Loss = 1473.248, a = 130.619 b = 27.020, c = -1.584, Joint = 3.642, Marginal = -0.861, MI = -76.799\n",
      "Pre-Train Epoch 0, Batch 243: Loss = 1422.856, a = 74.016 b = 22.864, c = -1.584, Joint = 0.175, Marginal = -0.831, MI = -37.979\n",
      "Pre-Train Epoch 0, Batch 244: Loss = 269.486, a = 78.983 b = 17.273, c = -1.585, Joint = -3.417, Marginal = -5.783, MI = -49.737\n",
      "Pre-Train Epoch 0, Batch 245: Loss = 369.316, a = 91.979 b = 22.396, c = -0.268, Joint = -3.674, Marginal = -5.264, MI = -31.365\n",
      "Pre-Train Epoch 0, Batch 246: Loss = 557.021, a = 78.801 b = 20.838, c = 1.710, Joint = 0.408, Marginal = -5.932, MI = -29.979\n",
      "Pre-Train Epoch 0, Batch 247: Loss = 330.252, a = 74.948 b = 19.520, c = 1.034, Joint = -1.086, Marginal = -5.306, MI = -54.352\n",
      "Pre-Train Epoch 0, Batch 248: Loss = 456.872, a = 117.761 b = 22.517, c = -1.586, Joint = -9.995, Marginal = -9.939, MI = -32.980\n",
      "Pre-Train Epoch 0, Batch 249: Loss = 603.714, a = 75.801 b = 20.484, c = 0.048, Joint = -3.259, Marginal = -6.934, MI = -9.598\n",
      "Pre-Train Epoch 0, Batch 250: Loss = 888.717, a = 89.896 b = 23.366, c = -0.587, Joint = -5.941, Marginal = -7.516, MI = -31.582\n",
      "Pre-Train Epoch 0, Batch 251: Loss = 182.416, a = 39.606 b = 13.751, c = -1.586, Joint = -5.671, Marginal = -4.031, MI = -26.006\n",
      "Pre-Train Epoch 0, Batch 252: Loss = 1118.323, a = 75.193 b = 20.344, c = -1.586, Joint = -1.911, Marginal = -6.362, MI = -15.458\n",
      "Pre-Train Epoch 0, Batch 253: Loss = 644.239, a = 68.661 b = 17.677, c = -1.586, Joint = -0.804, Marginal = -5.162, MI = -38.159\n",
      "Pre-Train Epoch 0, Batch 254: Loss = 263.500, a = 66.980 b = 19.270, c = -1.587, Joint = -4.464, Marginal = -6.982, MI = -52.399\n",
      "Pre-Train Epoch 0, Batch 255: Loss = 3114.861, a = 155.192 b = 32.476, c = -0.272, Joint = 3.306, Marginal = -8.636, MI = -30.553\n",
      "Pre-Train Epoch 0, Batch 256: Loss = 352.134, a = 66.013 b = 19.163, c = -1.587, Joint = -2.378, Marginal = -6.197, MI = -21.458\n",
      "Pre-Train Epoch 0, Batch 257: Loss = 747.199, a = 94.012 b = 23.784, c = -1.587, Joint = -4.190, Marginal = -9.154, MI = -24.683\n",
      "Pre-Train Epoch 0, Batch 258: Loss = 294.968, a = 65.344 b = 19.350, c = 1.205, Joint = -1.575, Marginal = -2.251, MI = -72.005\n",
      "Pre-Train Epoch 0, Batch 259: Loss = 499.941, a = 59.894 b = 18.664, c = 1.655, Joint = -3.825, Marginal = -7.070, MI = -52.787\n",
      "Pre-Train Epoch 0, Batch 260: Loss = 535.735, a = 73.967 b = 23.472, c = -1.588, Joint = -4.314, Marginal = -4.214, MI = -68.387\n",
      "Pre-Train Epoch 0, Batch 261: Loss = 469.390, a = 97.673 b = 25.263, c = -1.588, Joint = -1.582, Marginal = -8.930, MI = -9.465\n",
      "Pre-Train Epoch 0, Batch 262: Loss = 364.440, a = 61.230 b = 19.628, c = -1.588, Joint = -1.954, Marginal = -4.912, MI = -33.372\n",
      "Pre-Train Epoch 0, Batch 263: Loss = 1936.901, a = 85.540 b = 21.057, c = -1.588, Joint = 5.409, Marginal = -6.268, MI = -23.705\n",
      "Pre-Train Epoch 0, Batch 264: Loss = 499.262, a = 76.322 b = 23.907, c = -0.256, Joint = -0.814, Marginal = -3.477, MI = -69.093\n",
      "Pre-Train Epoch 0, Batch 265: Loss = 489.562, a = 72.241 b = 20.497, c = -0.463, Joint = -6.758, Marginal = -8.772, MI = -16.280\n",
      "Pre-Train Epoch 0, Batch 266: Loss = 194.124, a = 42.168 b = 16.199, c = -1.589, Joint = -4.494, Marginal = -5.873, MI = -23.208\n",
      "Pre-Train Epoch 0, Batch 267: Loss = 689.384, a = 71.421 b = 21.711, c = -1.589, Joint = -0.278, Marginal = -5.045, MI = -66.303\n",
      "Pre-Train Epoch 0, Batch 268: Loss = 484.168, a = 65.163 b = 21.482, c = -1.589, Joint = 1.121, Marginal = -4.634, MI = -65.884\n",
      "Pre-Train Epoch 0, Batch 269: Loss = 1353.626, a = 76.145 b = 23.200, c = -0.667, Joint = 1.131, Marginal = -2.276, MI = -6.575\n",
      "Pre-Train Epoch 0, Batch 270: Loss = 961.401, a = 65.847 b = 19.551, c = 1.162, Joint = -1.257, Marginal = -7.974, MI = -33.076\n",
      "Pre-Train Epoch 0, Batch 271: Loss = 2854.615, a = 82.041 b = 20.600, c = -1.589, Joint = 4.466, Marginal = -1.942, MI = -9.507\n",
      "Pre-Train Epoch 0, Batch 272: Loss = 615.873, a = 51.050 b = 19.501, c = -1.590, Joint = 2.046, Marginal = 0.040, MI = -55.877\n",
      "Pre-Train Epoch 0, Batch 273: Loss = 633.031, a = 50.403 b = 19.095, c = -1.590, Joint = -1.570, Marginal = -4.238, MI = -51.477\n",
      "Pre-Train Epoch 0, Batch 274: Loss = 731.568, a = 53.093 b = 17.926, c = -0.962, Joint = 0.710, Marginal = -4.125, MI = -32.990\n",
      "Pre-Train Epoch 0, Batch 275: Loss = 199.703, a = 48.517 b = 15.459, c = -1.095, Joint = -4.446, Marginal = -6.759, MI = -6.855\n",
      "Pre-Train Epoch 0, Batch 276: Loss = 446.394, a = 66.509 b = 19.726, c = -0.733, Joint = -1.343, Marginal = -6.255, MI = -23.946\n",
      "Pre-Train Epoch 0, Batch 277: Loss = 2263.933, a = 97.296 b = 27.408, c = -0.243, Joint = 1.975, Marginal = -2.578, MI = -76.691\n",
      "Pre-Train Epoch 0, Batch 278: Loss = 637.420, a = 83.803 b = 28.792, c = -1.591, Joint = -3.484, Marginal = -6.911, MI = -77.695\n",
      "Pre-Train Epoch 0, Batch 279: Loss = 304.171, a = 46.316 b = 17.129, c = -1.592, Joint = -5.074, Marginal = -5.610, MI = -30.095\n",
      "Pre-Train Epoch 0, Batch 280: Loss = 616.048, a = 74.984 b = 21.211, c = -1.592, Joint = 0.964, Marginal = -4.581, MI = -72.460\n",
      "Pre-Train Epoch 0, Batch 281: Loss = 678.568, a = 99.970 b = 26.902, c = 0.165, Joint = -2.401, Marginal = -6.213, MI = -25.283\n",
      "Pre-Train Epoch 0, Batch 282: Loss = 2557.202, a = 121.142 b = 31.046, c = -0.007, Joint = 6.350, Marginal = -4.084, MI = -32.631\n",
      "Pre-Train Epoch 0, Batch 283: Loss = 286.467, a = 61.847 b = 21.048, c = -0.486, Joint = -5.836, Marginal = -3.473, MI = -8.724\n",
      "Pre-Train Epoch 0, Batch 284: Loss = 603.779, a = 47.992 b = 18.304, c = -1.593, Joint = 1.999, Marginal = -1.462, MI = -11.410\n",
      "Pre-Train Epoch 0, Batch 285: Loss = 394.467, a = 71.529 b = 23.742, c = 1.100, Joint = -1.861, Marginal = -5.621, MI = -50.036\n",
      "Pre-Train Epoch 0, Batch 286: Loss = 1178.862, a = 66.469 b = 19.472, c = -0.026, Joint = 3.909, Marginal = 0.250, MI = -36.564\n",
      "Pre-Train Epoch 0, Batch 287: Loss = 2797.502, a = 58.595 b = 21.153, c = -0.605, Joint = 9.007, Marginal = -1.437, MI = -60.806\n",
      "Pre-Train Epoch 0, Batch 288: Loss = 488.810, a = 78.466 b = 25.784, c = 1.504, Joint = -3.471, Marginal = -5.682, MI = -33.327\n",
      "Pre-Train Epoch 0, Batch 289: Loss = 290.065, a = 41.234 b = 17.242, c = -0.428, Joint = -3.879, Marginal = -4.358, MI = -9.805\n",
      "Pre-Train Epoch 0, Batch 290: Loss = 213.923, a = 55.899 b = 16.558, c = -1.347, Joint = -2.840, Marginal = -6.110, MI = -18.872\n",
      "Pre-Train Epoch 0, Batch 291: Loss = 1078.994, a = 54.538 b = 18.465, c = -1.595, Joint = -0.319, Marginal = -5.993, MI = -14.940\n",
      "Pre-Train Epoch 0, Batch 292: Loss = 542.961, a = 71.928 b = 23.744, c = 1.368, Joint = -6.955, Marginal = -5.008, MI = -84.424\n",
      "Pre-Train Epoch 0, Batch 293: Loss = 333.425, a = 70.294 b = 23.430, c = -0.293, Joint = -3.681, Marginal = -5.470, MI = -54.917\n",
      "Pre-Train Epoch 0, Batch 294: Loss = 428.370, a = 75.236 b = 22.633, c = -0.728, Joint = 1.006, Marginal = -2.160, MI = -25.867\n",
      "Pre-Train Epoch 0, Batch 295: Loss = 228.781, a = 48.904 b = 16.649, c = 0.131, Joint = -1.324, Marginal = -1.576, MI = -35.047\n",
      "Pre-Train Epoch 0, Batch 296: Loss = 690.016, a = 79.878 b = 22.751, c = 0.147, Joint = 0.363, Marginal = -5.006, MI = -33.044\n",
      "Pre-Train Epoch 0, Batch 297: Loss = 655.041, a = 90.992 b = 28.650, c = -0.326, Joint = -3.455, Marginal = -5.897, MI = -38.554\n",
      "Pre-Train Epoch 0, Batch 298: Loss = 392.739, a = 49.462 b = 18.042, c = -1.535, Joint = -1.930, Marginal = -5.657, MI = -11.364\n",
      "Pre-Train Epoch 0, Batch 299: Loss = 1000.059, a = 81.781 b = 23.644, c = -0.767, Joint = 1.833, Marginal = -4.466, MI = -55.248\n",
      "Pre-Train Epoch 0, Batch 300: Loss = 230.732, a = 41.825 b = 14.489, c = -1.153, Joint = -1.302, Marginal = -5.001, MI = -31.137\n",
      "Pre-Train Epoch 0, Batch 301: Loss = 446.619, a = 85.373 b = 25.638, c = -0.761, Joint = -1.398, Marginal = -5.335, MI = -35.302\n",
      "Pre-Train Epoch 0, Batch 302: Loss = 1868.146, a = 84.375 b = 24.483, c = -1.597, Joint = 2.080, Marginal = -0.769, MI = -27.540\n",
      "Pre-Train Epoch 0, Batch 303: Loss = 611.354, a = 69.223 b = 22.972, c = 2.230, Joint = -0.103, Marginal = -1.698, MI = -75.193\n",
      "Pre-Train Epoch 0, Batch 304: Loss = 747.969, a = 60.551 b = 20.664, c = -1.597, Joint = 4.056, Marginal = -5.414, MI = -18.565\n",
      "Pre-Train Epoch 0, Batch 305: Loss = 311.477, a = 36.371 b = 15.756, c = -1.598, Joint = -1.417, Marginal = -1.175, MI = -33.944\n",
      "Pre-Train Epoch 0, Batch 306: Loss = 583.271, a = 71.226 b = 24.056, c = 0.716, Joint = -3.690, Marginal = -4.105, MI = -54.634\n",
      "\n",
      "Start of epoch 1\n",
      "Pre-Train Epoch 1, Batch 0: Loss = 389.041, a = 61.770 b = 20.097, c = -1.068, Joint = -1.753, Marginal = -4.206, MI = -79.944\n",
      "Pre-Train Epoch 1, Batch 1: Loss = 300.322, a = 52.957 b = 18.794, c = -1.151, Joint = 1.083, Marginal = -3.315, MI = -20.702\n",
      "Pre-Train Epoch 1, Batch 2: Loss = 344.518, a = 96.585 b = 27.469, c = -1.598, Joint = -4.235, Marginal = -7.894, MI = -32.606\n",
      "Pre-Train Epoch 1, Batch 3: Loss = 311.505, a = 71.957 b = 22.306, c = 0.746, Joint = -1.945, Marginal = -6.578, MI = -44.399\n",
      "Pre-Train Epoch 1, Batch 4: Loss = 238.136, a = 47.811 b = 17.638, c = -1.599, Joint = -2.539, Marginal = -3.985, MI = -52.894\n",
      "Pre-Train Epoch 1, Batch 5: Loss = 197.095, a = 51.578 b = 20.105, c = 0.154, Joint = -3.326, Marginal = -3.639, MI = -46.242\n",
      "Pre-Train Epoch 1, Batch 6: Loss = 668.020, a = 54.996 b = 20.940, c = -1.356, Joint = -0.445, Marginal = -4.957, MI = -33.655\n",
      "Pre-Train Epoch 1, Batch 7: Loss = 182.294, a = 32.197 b = 13.395, c = -1.372, Joint = -1.060, Marginal = -2.627, MI = -18.384\n",
      "Pre-Train Epoch 1, Batch 8: Loss = 704.086, a = 43.962 b = 21.162, c = -0.638, Joint = -1.251, Marginal = -5.426, MI = -30.877\n",
      "Pre-Train Epoch 1, Batch 9: Loss = 416.293, a = 52.341 b = 18.231, c = -0.949, Joint = -4.019, Marginal = -3.451, MI = -26.752\n",
      "Pre-Train Epoch 1, Batch 10: Loss = 767.923, a = 47.557 b = 20.278, c = -1.600, Joint = 2.525, Marginal = -3.484, MI = -33.271\n",
      "Pre-Train Epoch 1, Batch 11: Loss = 400.308, a = 56.669 b = 21.616, c = -1.048, Joint = -2.269, Marginal = -6.266, MI = -30.906\n",
      "Pre-Train Epoch 1, Batch 12: Loss = 183.792, a = 25.011 b = 14.162, c = -0.962, Joint = -0.042, Marginal = -1.736, MI = -64.965\n",
      "Pre-Train Epoch 1, Batch 13: Loss = 334.488, a = 42.400 b = 20.087, c = 0.566, Joint = 0.018, Marginal = -2.393, MI = -26.908\n",
      "Pre-Train Epoch 1, Batch 14: Loss = 545.445, a = 87.216 b = 30.209, c = -0.330, Joint = -0.011, Marginal = -4.042, MI = -41.197\n",
      "Pre-Train Epoch 1, Batch 15: Loss = 207.351, a = 27.139 b = 13.374, c = -1.600, Joint = -2.702, Marginal = -3.189, MI = -30.891\n",
      "Pre-Train Epoch 1, Batch 16: Loss = 508.687, a = 44.414 b = 16.420, c = -1.582, Joint = 0.560, Marginal = -4.751, MI = -16.978\n",
      "Pre-Train Epoch 1, Batch 17: Loss = 345.625, a = 49.666 b = 21.394, c = -0.690, Joint = -1.264, Marginal = -3.604, MI = -24.138\n",
      "Pre-Train Epoch 1, Batch 18: Loss = 428.730, a = 43.667 b = 17.937, c = -0.422, Joint = -2.513, Marginal = -1.858, MI = -30.632\n",
      "Pre-Train Epoch 1, Batch 19: Loss = 195.202, a = 40.838 b = 17.493, c = -0.452, Joint = -2.778, Marginal = -3.765, MI = -53.642\n",
      "Pre-Train Epoch 1, Batch 20: Loss = 272.541, a = 59.100 b = 22.080, c = -0.913, Joint = -3.394, Marginal = -3.877, MI = -67.992\n",
      "Pre-Train Epoch 1, Batch 21: Loss = 197.948, a = 38.292 b = 18.118, c = -0.557, Joint = -0.738, Marginal = -5.508, MI = -12.329\n",
      "Pre-Train Epoch 1, Batch 22: Loss = 305.504, a = 40.894 b = 16.413, c = 0.928, Joint = -1.632, Marginal = -5.864, MI = -15.923\n",
      "Pre-Train Epoch 1, Batch 23: Loss = 2229.670, a = 59.955 b = 20.917, c = -0.988, Joint = 1.635, Marginal = -7.526, MI = -20.537\n",
      "Pre-Train Epoch 1, Batch 24: Loss = 1498.485, a = 74.499 b = 23.773, c = -1.066, Joint = 7.097, Marginal = -4.244, MI = -20.419\n",
      "Pre-Train Epoch 1, Batch 25: Loss = 586.657, a = 70.585 b = 21.987, c = -1.521, Joint = -3.724, Marginal = -2.297, MI = -16.378\n",
      "Pre-Train Epoch 1, Batch 26: Loss = 442.224, a = 42.854 b = 17.979, c = 0.560, Joint = 2.799, Marginal = -3.133, MI = -42.559\n",
      "Pre-Train Epoch 1, Batch 27: Loss = 340.396, a = 37.470 b = 16.759, c = -0.879, Joint = -2.711, Marginal = -3.762, MI = -32.799\n",
      "Pre-Train Epoch 1, Batch 28: Loss = 464.429, a = 40.537 b = 21.253, c = -1.138, Joint = -2.315, Marginal = -4.765, MI = -32.584\n",
      "Pre-Train Epoch 1, Batch 29: Loss = 3811.018, a = 78.416 b = 27.977, c = -1.603, Joint = 7.350, Marginal = -2.861, MI = -27.071\n",
      "Pre-Train Epoch 1, Batch 30: Loss = 491.921, a = 78.162 b = 27.785, c = 0.059, Joint = -2.727, Marginal = -5.930, MI = -52.252\n",
      "Pre-Train Epoch 1, Batch 31: Loss = 306.246, a = 35.140 b = 17.759, c = -1.603, Joint = -0.900, Marginal = -3.490, MI = -30.239\n",
      "Pre-Train Epoch 1, Batch 32: Loss = 396.058, a = 56.878 b = 21.541, c = -1.122, Joint = -1.576, Marginal = -3.870, MI = -12.168\n",
      "Pre-Train Epoch 1, Batch 33: Loss = 331.762, a = 45.378 b = 20.054, c = -1.604, Joint = -0.189, Marginal = -4.709, MI = -45.436\n",
      "Pre-Train Epoch 1, Batch 34: Loss = 305.548, a = 48.158 b = 19.775, c = -0.587, Joint = -2.705, Marginal = -5.182, MI = -54.960\n",
      "Pre-Train Epoch 1, Batch 35: Loss = 365.838, a = 60.123 b = 20.046, c = -0.014, Joint = -1.357, Marginal = -3.647, MI = -58.104\n",
      "Pre-Train Epoch 1, Batch 36: Loss = 255.899, a = 39.041 b = 17.782, c = -0.531, Joint = -1.689, Marginal = -3.723, MI = -22.820\n",
      "Pre-Train Epoch 1, Batch 37: Loss = 557.224, a = 69.102 b = 25.923, c = 1.557, Joint = -1.619, Marginal = -2.212, MI = -48.936\n",
      "Pre-Train Epoch 1, Batch 38: Loss = 814.831, a = 70.341 b = 21.868, c = -1.605, Joint = -1.608, Marginal = -5.723, MI = -18.419\n",
      "Pre-Train Epoch 1, Batch 39: Loss = 408.027, a = 65.927 b = 26.649, c = -1.605, Joint = -2.964, Marginal = -6.536, MI = -25.815\n",
      "Pre-Train Epoch 1, Batch 40: Loss = 373.162, a = 42.054 b = 17.621, c = -1.381, Joint = -0.694, Marginal = -6.018, MI = -11.344\n",
      "Pre-Train Epoch 1, Batch 41: Loss = 347.727, a = 37.933 b = 18.882, c = -1.606, Joint = -0.982, Marginal = -2.084, MI = -54.337\n",
      "Pre-Train Epoch 1, Batch 42: Loss = 337.526, a = 46.977 b = 21.228, c = 0.871, Joint = -3.690, Marginal = -6.934, MI = -55.145\n",
      "Pre-Train Epoch 1, Batch 43: Loss = 416.519, a = 38.443 b = 17.515, c = -0.374, Joint = 0.081, Marginal = -3.901, MI = -15.713\n",
      "Pre-Train Epoch 1, Batch 44: Loss = 162.426, a = 42.653 b = 16.045, c = -1.606, Joint = -3.509, Marginal = -4.757, MI = -25.689\n",
      "Pre-Train Epoch 1, Batch 45: Loss = 400.127, a = 51.134 b = 20.658, c = 0.361, Joint = 0.400, Marginal = -0.904, MI = -24.055\n",
      "Pre-Train Epoch 1, Batch 46: Loss = 219.252, a = 48.793 b = 19.894, c = -0.795, Joint = -1.629, Marginal = -6.030, MI = -26.567\n",
      "Pre-Train Epoch 1, Batch 47: Loss = 451.562, a = 52.055 b = 20.878, c = -1.607, Joint = -1.340, Marginal = -4.548, MI = -28.774\n",
      "Pre-Train Epoch 1, Batch 48: Loss = 800.217, a = 65.285 b = 22.745, c = -1.607, Joint = 2.214, Marginal = -2.324, MI = -20.521\n",
      "Pre-Train Epoch 1, Batch 49: Loss = 139.731, a = 24.708 b = 14.088, c = -1.607, Joint = -1.595, Marginal = -2.977, MI = -63.795\n",
      "Pre-Train Epoch 1, Batch 50: Loss = 495.481, a = 35.750 b = 17.114, c = -1.607, Joint = 1.975, Marginal = 0.982, MI = -42.084\n",
      "Pre-Train Epoch 1, Batch 51: Loss = 200.482, a = 33.513 b = 15.458, c = -1.608, Joint = -0.943, Marginal = -5.321, MI = -4.683\n",
      "Pre-Train Epoch 1, Batch 52: Loss = 315.905, a = 47.848 b = 21.086, c = -1.608, Joint = -4.079, Marginal = -7.944, MI = -17.610\n",
      "Pre-Train Epoch 1, Batch 53: Loss = 1798.871, a = 56.790 b = 21.425, c = -1.608, Joint = 4.186, Marginal = -3.583, MI = -20.414\n",
      "Pre-Train Epoch 1, Batch 54: Loss = 231.897, a = 40.238 b = 18.722, c = -1.608, Joint = -2.627, Marginal = -3.596, MI = -6.285\n",
      "Pre-Train Epoch 1, Batch 55: Loss = 696.734, a = 50.059 b = 22.233, c = 0.384, Joint = 1.466, Marginal = -4.715, MI = -55.115\n",
      "Pre-Train Epoch 1, Batch 56: Loss = 363.361, a = 38.565 b = 17.741, c = -1.608, Joint = -5.765, Marginal = -4.665, MI = -13.511\n",
      "Pre-Train Epoch 1, Batch 57: Loss = 2728.989, a = 57.248 b = 22.973, c = -0.732, Joint = 2.732, Marginal = -0.008, MI = -58.167\n",
      "Pre-Train Epoch 1, Batch 58: Loss = 446.952, a = 56.625 b = 24.043, c = -0.392, Joint = -1.250, Marginal = -6.423, MI = -71.442\n",
      "Pre-Train Epoch 1, Batch 59: Loss = 921.134, a = 67.278 b = 24.925, c = -0.587, Joint = 1.231, Marginal = -3.385, MI = -54.572\n",
      "Pre-Train Epoch 1, Batch 60: Loss = 1486.640, a = 45.668 b = 22.332, c = -0.718, Joint = 4.704, Marginal = -3.086, MI = -23.420\n",
      "Pre-Train Epoch 1, Batch 61: Loss = 309.578, a = 48.342 b = 20.925, c = -0.159, Joint = -4.554, Marginal = -4.426, MI = -52.819\n",
      "Pre-Train Epoch 1, Batch 62: Loss = 332.993, a = 44.711 b = 18.843, c = -1.393, Joint = -0.695, Marginal = -4.076, MI = -16.926\n",
      "Pre-Train Epoch 1, Batch 63: Loss = 437.134, a = 55.045 b = 23.173, c = -1.099, Joint = -4.920, Marginal = -8.956, MI = -24.349\n",
      "Pre-Train Epoch 1, Batch 64: Loss = 947.689, a = 38.307 b = 19.610, c = -0.484, Joint = 2.757, Marginal = -4.074, MI = -1.860\n",
      "Pre-Train Epoch 1, Batch 65: Loss = 912.057, a = 62.068 b = 26.007, c = -1.611, Joint = -0.628, Marginal = -8.904, MI = -10.843\n",
      "Pre-Train Epoch 1, Batch 66: Loss = 1191.957, a = 52.793 b = 24.317, c = 1.382, Joint = 0.070, Marginal = -6.442, MI = -33.547\n",
      "Pre-Train Epoch 1, Batch 67: Loss = 410.900, a = 50.621 b = 22.571, c = -1.612, Joint = -2.358, Marginal = -6.387, MI = -52.371\n",
      "Pre-Train Epoch 1, Batch 68: Loss = 353.573, a = 42.836 b = 20.921, c = -0.842, Joint = 0.209, Marginal = -2.849, MI = -30.291\n",
      "Pre-Train Epoch 1, Batch 69: Loss = 1056.182, a = 50.747 b = 22.832, c = -1.612, Joint = 4.288, Marginal = -1.993, MI = -69.544\n",
      "Pre-Train Epoch 1, Batch 70: Loss = 316.816, a = 28.077 b = 17.045, c = -1.594, Joint = 1.733, Marginal = -0.752, MI = -72.694\n",
      "Pre-Train Epoch 1, Batch 71: Loss = 286.472, a = 43.243 b = 18.355, c = 1.516, Joint = -2.942, Marginal = -4.733, MI = -17.333\n",
      "Pre-Train Epoch 1, Batch 72: Loss = 664.043, a = 38.553 b = 18.913, c = -0.270, Joint = -0.141, Marginal = -3.704, MI = -6.041\n",
      "Pre-Train Epoch 1, Batch 73: Loss = 566.105, a = 46.328 b = 20.653, c = -1.613, Joint = 1.630, Marginal = -4.708, MI = -30.131\n",
      "Pre-Train Epoch 1, Batch 74: Loss = 153.860, a = 24.018 b = 14.509, c = -1.614, Joint = -3.110, Marginal = -4.667, MI = -10.782\n",
      "Pre-Train Epoch 1, Batch 75: Loss = 518.789, a = 59.428 b = 22.878, c = -0.659, Joint = -4.280, Marginal = -6.080, MI = -54.127\n",
      "Pre-Train Epoch 1, Batch 76: Loss = 721.749, a = 69.708 b = 23.799, c = -1.614, Joint = -2.008, Marginal = -5.539, MI = -31.261\n",
      "Pre-Train Epoch 1, Batch 77: Loss = 522.349, a = 49.894 b = 22.683, c = -1.614, Joint = -3.120, Marginal = -7.375, MI = -50.720\n",
      "Pre-Train Epoch 1, Batch 78: Loss = 282.077, a = 48.630 b = 19.752, c = -1.615, Joint = -1.535, Marginal = -4.146, MI = -65.196\n",
      "Pre-Train Epoch 1, Batch 79: Loss = 209.701, a = 56.968 b = 22.294, c = 0.053, Joint = -4.186, Marginal = -5.554, MI = -40.285\n",
      "Pre-Train Epoch 1, Batch 80: Loss = 415.032, a = 51.943 b = 22.886, c = -1.615, Joint = 0.174, Marginal = -4.334, MI = -83.293\n",
      "Pre-Train Epoch 1, Batch 81: Loss = 664.329, a = 36.278 b = 19.460, c = -1.615, Joint = -0.247, Marginal = -4.781, MI = -35.532\n",
      "Pre-Train Epoch 1, Batch 82: Loss = 366.517, a = 53.735 b = 20.690, c = -1.572, Joint = -2.040, Marginal = -4.122, MI = -63.228\n",
      "Pre-Train Epoch 1, Batch 83: Loss = 1722.898, a = 47.108 b = 19.854, c = -0.267, Joint = 4.250, Marginal = -2.604, MI = -48.902\n",
      "Pre-Train Epoch 1, Batch 84: Loss = 626.529, a = 42.934 b = 20.874, c = -0.691, Joint = 0.829, Marginal = -3.784, MI = -70.487\n",
      "Pre-Train Epoch 1, Batch 85: Loss = 537.476, a = 49.666 b = 21.906, c = -1.082, Joint = 1.392, Marginal = -5.687, MI = -46.694\n",
      "Pre-Train Epoch 1, Batch 86: Loss = 230.899, a = 51.214 b = 22.410, c = -1.616, Joint = -2.157, Marginal = 0.022, MI = -44.583\n",
      "Pre-Train Epoch 1, Batch 87: Loss = 396.744, a = 35.120 b = 18.785, c = 0.301, Joint = 0.113, Marginal = -2.053, MI = -67.278\n",
      "Pre-Train Epoch 1, Batch 88: Loss = 183.476, a = 38.581 b = 17.232, c = -0.411, Joint = -3.586, Marginal = -5.510, MI = -12.482\n",
      "Pre-Train Epoch 1, Batch 89: Loss = 720.331, a = 58.961 b = 25.617, c = -1.617, Joint = 1.071, Marginal = -5.807, MI = -64.765\n",
      "Pre-Train Epoch 1, Batch 90: Loss = 210.486, a = 47.016 b = 21.715, c = -1.020, Joint = -5.818, Marginal = -6.689, MI = -21.930\n",
      "Pre-Train Epoch 1, Batch 91: Loss = 234.567, a = 28.533 b = 16.089, c = -1.517, Joint = -0.412, Marginal = -2.947, MI = -21.468\n",
      "Pre-Train Epoch 1, Batch 92: Loss = 962.377, a = 50.296 b = 23.771, c = -1.097, Joint = 2.112, Marginal = -3.913, MI = -15.602\n",
      "Pre-Train Epoch 1, Batch 93: Loss = 3368.414, a = 47.068 b = 23.209, c = 0.291, Joint = 3.666, Marginal = -3.643, MI = -32.769\n",
      "Pre-Train Epoch 1, Batch 94: Loss = 485.083, a = 45.024 b = 21.681, c = -0.986, Joint = -0.019, Marginal = -4.971, MI = -29.259\n",
      "Pre-Train Epoch 1, Batch 95: Loss = 380.114, a = 60.453 b = 23.305, c = -1.618, Joint = -4.855, Marginal = -6.617, MI = -55.433\n",
      "Pre-Train Epoch 1, Batch 96: Loss = 245.021, a = 37.032 b = 20.540, c = 0.543, Joint = -3.707, Marginal = -5.595, MI = -31.617\n",
      "Pre-Train Epoch 1, Batch 97: Loss = 194.755, a = 28.809 b = 16.757, c = -0.612, Joint = -1.728, Marginal = -4.333, MI = -22.632\n",
      "Pre-Train Epoch 1, Batch 98: Loss = 398.836, a = 39.883 b = 18.640, c = -1.338, Joint = -1.161, Marginal = -5.333, MI = -11.784\n",
      "Pre-Train Epoch 1, Batch 99: Loss = 589.925, a = 39.603 b = 20.084, c = 0.380, Joint = 1.037, Marginal = -5.251, MI = -40.845\n",
      "Pre-Train Epoch 1, Batch 100: Loss = 992.206, a = 51.610 b = 23.713, c = -0.672, Joint = 4.224, Marginal = -5.793, MI = -17.753\n",
      "Pre-Train Epoch 1, Batch 101: Loss = 178.782, a = 33.955 b = 16.803, c = 0.873, Joint = -4.159, Marginal = -5.912, MI = -9.603\n",
      "Pre-Train Epoch 1, Batch 102: Loss = 271.164, a = 40.540 b = 18.179, c = -1.620, Joint = -3.497, Marginal = -6.876, MI = -22.899\n",
      "Pre-Train Epoch 1, Batch 103: Loss = 525.706, a = 40.698 b = 22.501, c = -0.554, Joint = 0.458, Marginal = -4.425, MI = -42.420\n",
      "Pre-Train Epoch 1, Batch 104: Loss = 226.877, a = 38.232 b = 18.703, c = -1.168, Joint = -6.919, Marginal = -6.458, MI = -14.957\n",
      "Pre-Train Epoch 1, Batch 105: Loss = 202.751, a = 33.148 b = 17.485, c = -0.057, Joint = -4.515, Marginal = -5.714, MI = -24.013\n",
      "Pre-Train Epoch 1, Batch 106: Loss = 341.902, a = 44.916 b = 20.806, c = 0.316, Joint = -4.470, Marginal = -5.774, MI = -33.286\n",
      "Pre-Train Epoch 1, Batch 107: Loss = 698.303, a = 32.914 b = 18.151, c = -1.040, Joint = -1.014, Marginal = -4.874, MI = -78.805\n",
      "Pre-Train Epoch 1, Batch 108: Loss = 666.035, a = 38.289 b = 19.490, c = -1.621, Joint = -0.117, Marginal = -2.220, MI = -79.783\n",
      "Pre-Train Epoch 1, Batch 109: Loss = 135.800, a = 32.237 b = 17.200, c = 0.421, Joint = -5.209, Marginal = -5.814, MI = -15.225\n",
      "Pre-Train Epoch 1, Batch 110: Loss = 469.375, a = 27.523 b = 15.935, c = -0.977, Joint = 0.543, Marginal = -4.766, MI = -29.435\n",
      "Pre-Train Epoch 1, Batch 111: Loss = 559.724, a = 41.874 b = 20.438, c = -1.334, Joint = -6.500, Marginal = -7.255, MI = -63.533\n",
      "Pre-Train Epoch 1, Batch 112: Loss = 268.397, a = 39.180 b = 22.468, c = 0.732, Joint = -1.857, Marginal = -3.075, MI = -46.899\n",
      "Pre-Train Epoch 1, Batch 113: Loss = 214.098, a = 35.442 b = 18.927, c = -0.407, Joint = -4.100, Marginal = -5.815, MI = -11.398\n",
      "Pre-Train Epoch 1, Batch 114: Loss = 390.746, a = 34.597 b = 19.101, c = -0.762, Joint = -0.187, Marginal = -3.062, MI = -19.697\n",
      "Pre-Train Epoch 1, Batch 115: Loss = 385.693, a = 43.348 b = 23.659, c = 0.302, Joint = -1.040, Marginal = -3.488, MI = -33.512\n",
      "Pre-Train Epoch 1, Batch 116: Loss = 602.746, a = 33.972 b = 18.644, c = -1.622, Joint = 0.007, Marginal = -5.077, MI = -8.577\n",
      "Pre-Train Epoch 1, Batch 117: Loss = 556.979, a = 44.147 b = 21.941, c = 0.026, Joint = -0.367, Marginal = -6.754, MI = -31.535\n",
      "Pre-Train Epoch 1, Batch 118: Loss = 133.641, a = 28.567 b = 15.270, c = -1.622, Joint = -2.780, Marginal = -3.210, MI = -30.876\n",
      "Pre-Train Epoch 1, Batch 119: Loss = 449.708, a = 47.430 b = 20.125, c = -1.622, Joint = -1.319, Marginal = -3.181, MI = -57.622\n",
      "Pre-Train Epoch 1, Batch 120: Loss = 546.590, a = 51.192 b = 25.731, c = -1.622, Joint = 2.434, Marginal = -1.124, MI = -66.576\n",
      "Pre-Train Epoch 1, Batch 121: Loss = 542.435, a = 45.448 b = 24.430, c = 0.493, Joint = -1.264, Marginal = -4.400, MI = -21.636\n",
      "Pre-Train Epoch 1, Batch 122: Loss = 172.435, a = 32.177 b = 19.090, c = -1.484, Joint = -3.910, Marginal = -4.090, MI = -52.266\n",
      "Pre-Train Epoch 1, Batch 123: Loss = 745.227, a = 51.158 b = 25.390, c = -1.623, Joint = 1.736, Marginal = -1.731, MI = -63.035\n",
      "Pre-Train Epoch 1, Batch 124: Loss = 192.579, a = 33.609 b = 17.233, c = -1.623, Joint = -5.310, Marginal = -6.441, MI = -8.910\n",
      "Pre-Train Epoch 1, Batch 125: Loss = 528.860, a = 38.136 b = 22.186, c = -1.387, Joint = 1.625, Marginal = -3.365, MI = -50.471\n",
      "Pre-Train Epoch 1, Batch 126: Loss = 163.225, a = 32.253 b = 16.886, c = -0.489, Joint = -3.910, Marginal = -2.103, MI = -20.306\n",
      "Pre-Train Epoch 1, Batch 127: Loss = 511.932, a = 48.219 b = 24.091, c = -1.623, Joint = -3.651, Marginal = -7.647, MI = -11.971\n",
      "Pre-Train Epoch 1, Batch 128: Loss = 564.279, a = 47.751 b = 23.441, c = -1.139, Joint = 1.983, Marginal = -2.018, MI = -41.611\n",
      "Pre-Train Epoch 1, Batch 129: Loss = 412.872, a = 38.227 b = 20.184, c = -1.624, Joint = 0.649, Marginal = -4.698, MI = -33.310\n",
      "Pre-Train Epoch 1, Batch 130: Loss = 348.674, a = 35.767 b = 18.959, c = -1.624, Joint = -2.789, Marginal = -4.845, MI = -70.521\n",
      "Pre-Train Epoch 1, Batch 131: Loss = 195.477, a = 20.054 b = 14.513, c = -0.644, Joint = -1.334, Marginal = -5.386, MI = -2.766\n",
      "Pre-Train Epoch 1, Batch 132: Loss = 301.298, a = 27.638 b = 18.749, c = -1.624, Joint = -1.811, Marginal = -2.309, MI = -44.155\n",
      "Pre-Train Epoch 1, Batch 133: Loss = 225.427, a = 48.049 b = 22.110, c = -0.703, Joint = -2.868, Marginal = -4.644, MI = -8.946\n",
      "Pre-Train Epoch 1, Batch 134: Loss = 575.624, a = 30.321 b = 17.486, c = -1.624, Joint = -0.101, Marginal = -2.673, MI = -64.518\n",
      "Pre-Train Epoch 1, Batch 135: Loss = 594.068, a = 50.195 b = 24.334, c = -0.144, Joint = -1.646, Marginal = -5.435, MI = -13.016\n",
      "Pre-Train Epoch 1, Batch 136: Loss = 344.521, a = 34.184 b = 20.102, c = -1.349, Joint = -1.430, Marginal = -5.573, MI = -25.821\n",
      "Pre-Train Epoch 1, Batch 137: Loss = 946.595, a = 49.253 b = 25.515, c = -0.960, Joint = 1.630, Marginal = -6.590, MI = -55.395\n",
      "Pre-Train Epoch 1, Batch 138: Loss = 247.936, a = 41.901 b = 19.580, c = -1.625, Joint = -4.628, Marginal = -7.155, MI = -13.366\n",
      "Pre-Train Epoch 1, Batch 139: Loss = 159.197, a = 34.705 b = 18.751, c = -0.923, Joint = -1.472, Marginal = -4.739, MI = -14.369\n",
      "Pre-Train Epoch 1, Batch 140: Loss = 626.451, a = 49.524 b = 20.607, c = -1.418, Joint = -2.731, Marginal = -3.867, MI = -57.564\n",
      "Pre-Train Epoch 1, Batch 141: Loss = 187.679, a = 28.291 b = 18.765, c = -1.626, Joint = -3.498, Marginal = -4.540, MI = -41.244\n",
      "Pre-Train Epoch 1, Batch 142: Loss = 334.267, a = 31.272 b = 17.674, c = -1.334, Joint = -0.869, Marginal = -3.824, MI = -22.427\n",
      "Pre-Train Epoch 1, Batch 143: Loss = 437.147, a = 29.900 b = 17.496, c = -1.626, Joint = -1.565, Marginal = -3.626, MI = -28.010\n",
      "Pre-Train Epoch 1, Batch 144: Loss = 338.587, a = 39.532 b = 23.374, c = -1.584, Joint = -4.297, Marginal = -5.136, MI = -16.559\n",
      "Pre-Train Epoch 1, Batch 145: Loss = 179.260, a = 19.286 b = 14.628, c = 0.071, Joint = -3.938, Marginal = -5.598, MI = -23.397\n",
      "Pre-Train Epoch 1, Batch 146: Loss = 894.332, a = 46.940 b = 30.517, c = -0.819, Joint = 3.560, Marginal = -3.449, MI = -52.848\n",
      "Pre-Train Epoch 1, Batch 147: Loss = 1088.005, a = 32.509 b = 22.204, c = -1.626, Joint = 1.504, Marginal = -3.491, MI = -65.780\n",
      "Pre-Train Epoch 1, Batch 148: Loss = 285.524, a = 34.030 b = 21.997, c = 2.151, Joint = 0.718, Marginal = -2.712, MI = -13.054\n",
      "Pre-Train Epoch 1, Batch 149: Loss = 278.080, a = 28.240 b = 16.807, c = -1.020, Joint = -2.038, Marginal = -5.794, MI = -9.973\n",
      "Pre-Train Epoch 1, Batch 150: Loss = 225.358, a = 28.212 b = 17.695, c = -0.682, Joint = -3.723, Marginal = -5.805, MI = -43.261\n",
      "Pre-Train Epoch 1, Batch 151: Loss = 320.927, a = 43.412 b = 22.803, c = -1.627, Joint = -3.455, Marginal = -6.604, MI = -52.628\n",
      "Pre-Train Epoch 1, Batch 152: Loss = 492.922, a = 52.654 b = 24.229, c = 0.044, Joint = -3.548, Marginal = -5.159, MI = -65.276\n",
      "Pre-Train Epoch 1, Batch 153: Loss = 126.204, a = 20.380 b = 14.574, c = 1.011, Joint = -2.296, Marginal = -3.326, MI = -19.960\n",
      "Pre-Train Epoch 1, Batch 154: Loss = 208.710, a = 19.385 b = 18.222, c = -0.457, Joint = -2.333, Marginal = -3.907, MI = -31.502\n",
      "Pre-Train Epoch 1, Batch 155: Loss = 221.004, a = 28.301 b = 16.526, c = -0.609, Joint = -3.886, Marginal = -6.311, MI = -19.388\n",
      "Pre-Train Epoch 1, Batch 156: Loss = 358.025, a = 47.969 b = 23.159, c = -1.628, Joint = -3.920, Marginal = -4.379, MI = -20.541\n",
      "Pre-Train Epoch 1, Batch 157: Loss = 652.868, a = 44.442 b = 24.527, c = -1.628, Joint = 1.465, Marginal = -2.106, MI = -45.136\n",
      "Pre-Train Epoch 1, Batch 158: Loss = 225.235, a = 30.045 b = 17.732, c = -1.131, Joint = -4.003, Marginal = -3.720, MI = -69.880\n",
      "Pre-Train Epoch 1, Batch 159: Loss = 330.770, a = 37.606 b = 20.877, c = -0.754, Joint = -1.800, Marginal = -5.974, MI = -11.766\n",
      "Pre-Train Epoch 1, Batch 160: Loss = 336.718, a = 40.340 b = 20.840, c = -1.628, Joint = -1.435, Marginal = -8.147, MI = -6.789\n",
      "Pre-Train Epoch 1, Batch 161: Loss = 144.205, a = 26.589 b = 17.249, c = 0.115, Joint = -3.525, Marginal = -5.326, MI = -11.392\n",
      "Pre-Train Epoch 1, Batch 162: Loss = 105.601, a = 25.412 b = 16.346, c = -0.852, Joint = -2.858, Marginal = -4.560, MI = -7.658\n",
      "Pre-Train Epoch 1, Batch 163: Loss = 219.502, a = 25.223 b = 16.396, c = -1.628, Joint = 1.003, Marginal = -5.010, MI = -10.758\n",
      "Pre-Train Epoch 1, Batch 164: Loss = 431.995, a = 28.981 b = 19.903, c = -0.907, Joint = -1.311, Marginal = -3.879, MI = -12.506\n",
      "Pre-Train Epoch 1, Batch 165: Loss = 462.156, a = 34.558 b = 20.859, c = -1.628, Joint = -3.571, Marginal = -4.221, MI = -30.997\n",
      "Pre-Train Epoch 1, Batch 166: Loss = 144.372, a = 32.395 b = 18.872, c = -1.064, Joint = -4.936, Marginal = -5.774, MI = -9.690\n",
      "Pre-Train Epoch 1, Batch 167: Loss = 178.685, a = 22.714 b = 13.228, c = 0.887, Joint = -3.294, Marginal = -4.458, MI = -15.669\n",
      "Pre-Train Epoch 1, Batch 168: Loss = 339.546, a = 13.224 b = 12.188, c = -0.954, Joint = 1.251, Marginal = -2.648, MI = -52.456\n",
      "Pre-Train Epoch 1, Batch 169: Loss = 173.937, a = 24.780 b = 18.607, c = -0.994, Joint = -1.805, Marginal = -4.723, MI = -22.229\n",
      "Pre-Train Epoch 1, Batch 170: Loss = 146.826, a = 21.819 b = 15.975, c = -1.629, Joint = -2.471, Marginal = -3.556, MI = -21.095\n",
      "Pre-Train Epoch 1, Batch 171: Loss = 272.150, a = 28.546 b = 19.058, c = -1.629, Joint = -0.082, Marginal = -3.615, MI = -36.925\n",
      "Pre-Train Epoch 1, Batch 172: Loss = 347.131, a = 32.826 b = 22.041, c = -1.122, Joint = -3.596, Marginal = -5.245, MI = -38.436\n",
      "Pre-Train Epoch 1, Batch 173: Loss = 499.989, a = 36.814 b = 21.591, c = -0.136, Joint = 1.201, Marginal = -4.965, MI = -49.263\n",
      "Pre-Train Epoch 1, Batch 174: Loss = 351.708, a = 22.763 b = 19.109, c = -0.143, Joint = -1.763, Marginal = -3.259, MI = -71.868\n",
      "Pre-Train Epoch 1, Batch 175: Loss = 116.065, a = 20.444 b = 17.245, c = -1.629, Joint = -2.644, Marginal = -4.581, MI = -7.744\n",
      "Pre-Train Epoch 1, Batch 176: Loss = 467.497, a = 31.392 b = 19.114, c = -1.629, Joint = 1.863, Marginal = -2.110, MI = -17.466\n",
      "Pre-Train Epoch 1, Batch 177: Loss = 174.230, a = 28.062 b = 20.152, c = 0.455, Joint = -4.932, Marginal = -6.846, MI = -10.653\n",
      "Pre-Train Epoch 1, Batch 178: Loss = 945.936, a = 38.246 b = 24.028, c = -1.629, Joint = -2.651, Marginal = -5.165, MI = -35.713\n",
      "Pre-Train Epoch 1, Batch 179: Loss = 1010.266, a = 32.239 b = 24.148, c = -0.682, Joint = 3.393, Marginal = -1.592, MI = -2.540\n",
      "Pre-Train Epoch 1, Batch 180: Loss = 217.638, a = 31.670 b = 23.542, c = -1.630, Joint = -3.169, Marginal = -4.887, MI = -32.117\n",
      "Pre-Train Epoch 1, Batch 181: Loss = 323.319, a = 23.819 b = 17.825, c = -1.091, Joint = -0.359, Marginal = -2.655, MI = -17.521\n",
      "Pre-Train Epoch 1, Batch 182: Loss = 370.913, a = 27.753 b = 19.143, c = 1.550, Joint = 1.068, Marginal = -5.394, MI = -8.643\n",
      "Pre-Train Epoch 1, Batch 183: Loss = 176.176, a = 25.768 b = 15.950, c = -1.630, Joint = -2.044, Marginal = -3.153, MI = -23.860\n",
      "Pre-Train Epoch 1, Batch 184: Loss = 204.595, a = 25.267 b = 21.337, c = -1.076, Joint = -1.117, Marginal = -3.716, MI = -34.496\n",
      "Pre-Train Epoch 1, Batch 185: Loss = 301.353, a = 24.215 b = 17.445, c = -0.841, Joint = -2.953, Marginal = -2.652, MI = -30.958\n",
      "Pre-Train Epoch 1, Batch 186: Loss = 513.266, a = 26.737 b = 20.831, c = -0.354, Joint = -1.262, Marginal = -5.606, MI = -37.934\n",
      "Pre-Train Epoch 1, Batch 187: Loss = 548.561, a = 18.647 b = 15.560, c = -0.191, Joint = -1.176, Marginal = -3.744, MI = -6.369\n",
      "Pre-Train Epoch 1, Batch 188: Loss = 708.766, a = 17.008 b = 20.195, c = -1.631, Joint = 4.540, Marginal = -0.716, MI = -22.199\n",
      "Pre-Train Epoch 1, Batch 189: Loss = 211.766, a = 21.460 b = 19.519, c = -0.329, Joint = 0.296, Marginal = -2.711, MI = -77.956\n",
      "Pre-Train Epoch 1, Batch 190: Loss = 396.560, a = 40.430 b = 22.163, c = -1.631, Joint = -2.477, Marginal = -4.865, MI = -18.114\n",
      "Pre-Train Epoch 1, Batch 191: Loss = 530.090, a = 29.885 b = 20.220, c = -1.631, Joint = -1.966, Marginal = -1.724, MI = -47.416\n",
      "Pre-Train Epoch 1, Batch 192: Loss = 152.456, a = 20.236 b = 15.232, c = -0.914, Joint = -2.048, Marginal = -3.395, MI = -21.903\n",
      "Pre-Train Epoch 1, Batch 193: Loss = 246.352, a = 29.934 b = 19.703, c = -0.157, Joint = -2.746, Marginal = -3.955, MI = -36.672\n",
      "Pre-Train Epoch 1, Batch 194: Loss = 128.205, a = 13.492 b = 14.704, c = -1.632, Joint = -0.853, Marginal = -4.317, MI = -18.741\n",
      "Pre-Train Epoch 1, Batch 195: Loss = 362.440, a = 32.031 b = 18.377, c = -1.632, Joint = -1.196, Marginal = -5.709, MI = -51.843\n",
      "Pre-Train Epoch 1, Batch 196: Loss = 266.123, a = 21.177 b = 16.779, c = -0.365, Joint = -2.935, Marginal = -5.346, MI = -23.253\n",
      "Pre-Train Epoch 1, Batch 197: Loss = 135.057, a = 17.473 b = 12.757, c = -0.539, Joint = -2.933, Marginal = -5.004, MI = -12.947\n",
      "Pre-Train Epoch 1, Batch 198: Loss = 117.107, a = 13.345 b = 14.517, c = -1.632, Joint = -1.880, Marginal = -3.774, MI = -4.447\n",
      "Pre-Train Epoch 1, Batch 199: Loss = 169.569, a = 28.433 b = 17.885, c = -1.144, Joint = -2.416, Marginal = -4.879, MI = -21.593\n",
      "Pre-Train Epoch 1, Batch 200: Loss = 281.728, a = 27.762 b = 19.603, c = -0.587, Joint = -2.045, Marginal = -5.668, MI = -14.716\n",
      "Pre-Train Epoch 1, Batch 201: Loss = 999.028, a = 25.123 b = 21.291, c = -0.962, Joint = 2.315, Marginal = -2.702, MI = -46.260\n",
      "Pre-Train Epoch 1, Batch 202: Loss = 385.125, a = 20.523 b = 17.707, c = -1.603, Joint = -3.287, Marginal = -7.063, MI = -17.935\n",
      "Pre-Train Epoch 1, Batch 203: Loss = 303.211, a = 32.705 b = 21.614, c = -0.667, Joint = -3.134, Marginal = -4.997, MI = -84.998\n",
      "Pre-Train Epoch 1, Batch 204: Loss = 97.152, a = 16.924 b = 13.439, c = -1.633, Joint = -3.032, Marginal = -2.735, MI = -17.038\n",
      "Pre-Train Epoch 1, Batch 205: Loss = 161.316, a = 18.501 b = 17.108, c = 0.046, Joint = -1.879, Marginal = -3.992, MI = -28.440\n",
      "Pre-Train Epoch 1, Batch 206: Loss = 237.463, a = 25.744 b = 17.657, c = -0.913, Joint = -5.236, Marginal = -4.409, MI = -18.548\n",
      "Pre-Train Epoch 1, Batch 207: Loss = 120.106, a = 22.131 b = 14.498, c = 0.066, Joint = -2.371, Marginal = -3.382, MI = -33.486\n",
      "Pre-Train Epoch 1, Batch 208: Loss = 146.198, a = 24.048 b = 14.758, c = 1.078, Joint = -2.292, Marginal = -4.510, MI = -4.844\n",
      "Pre-Train Epoch 1, Batch 209: Loss = 158.889, a = 17.707 b = 16.854, c = -1.633, Joint = -0.636, Marginal = -2.098, MI = -39.241\n",
      "Pre-Train Epoch 1, Batch 210: Loss = 615.548, a = 20.974 b = 19.550, c = -0.445, Joint = 0.615, Marginal = -4.088, MI = -70.505\n",
      "Pre-Train Epoch 1, Batch 211: Loss = 206.819, a = 19.868 b = 16.471, c = -1.415, Joint = 1.134, Marginal = -2.946, MI = -10.569\n",
      "Pre-Train Epoch 1, Batch 212: Loss = 131.641, a = 16.540 b = 15.138, c = -1.634, Joint = -4.551, Marginal = -5.738, MI = -10.308\n",
      "Pre-Train Epoch 1, Batch 213: Loss = 362.225, a = 24.679 b = 19.302, c = -1.559, Joint = -1.070, Marginal = -4.723, MI = -42.460\n",
      "Pre-Train Epoch 1, Batch 214: Loss = 314.624, a = 19.586 b = 19.118, c = -1.634, Joint = 0.148, Marginal = -2.176, MI = -35.050\n",
      "Pre-Train Epoch 1, Batch 215: Loss = 358.994, a = 30.290 b = 19.219, c = -1.506, Joint = -3.763, Marginal = -5.470, MI = -22.870\n",
      "Pre-Train Epoch 1, Batch 216: Loss = 703.545, a = 22.404 b = 15.712, c = 0.562, Joint = 0.023, Marginal = -5.579, MI = -8.236\n",
      "Pre-Train Epoch 1, Batch 217: Loss = 544.461, a = 23.371 b = 21.464, c = -1.634, Joint = -0.586, Marginal = -1.221, MI = -36.628\n",
      "Pre-Train Epoch 1, Batch 218: Loss = 346.698, a = 20.285 b = 19.749, c = -1.634, Joint = 1.123, Marginal = -2.327, MI = -37.049\n",
      "Pre-Train Epoch 1, Batch 219: Loss = 196.877, a = 17.795 b = 19.649, c = -0.351, Joint = -1.217, Marginal = -4.434, MI = -24.472\n",
      "Pre-Train Epoch 1, Batch 220: Loss = 263.473, a = 20.860 b = 20.776, c = 0.363, Joint = -2.207, Marginal = -4.401, MI = -36.865\n",
      "Pre-Train Epoch 1, Batch 221: Loss = 164.346, a = 20.300 b = 16.426, c = -1.585, Joint = -2.167, Marginal = -4.489, MI = -20.140\n",
      "Pre-Train Epoch 1, Batch 222: Loss = 1831.858, a = 25.594 b = 21.539, c = -1.635, Joint = 2.409, Marginal = -3.766, MI = -38.210\n",
      "Pre-Train Epoch 1, Batch 223: Loss = 492.423, a = 26.841 b = 20.334, c = 0.461, Joint = -0.078, Marginal = -4.578, MI = -52.086\n",
      "Pre-Train Epoch 1, Batch 224: Loss = 188.501, a = 22.479 b = 13.526, c = 1.040, Joint = -3.560, Marginal = -5.010, MI = -4.314\n",
      "Pre-Train Epoch 1, Batch 225: Loss = 316.122, a = 25.211 b = 22.717, c = -1.280, Joint = -1.883, Marginal = -7.514, MI = -9.652\n",
      "Pre-Train Epoch 1, Batch 226: Loss = 489.771, a = 23.506 b = 14.638, c = -1.636, Joint = -0.475, Marginal = -3.015, MI = -66.691\n",
      "Pre-Train Epoch 1, Batch 227: Loss = 264.190, a = 18.684 b = 15.014, c = -1.636, Joint = -1.201, Marginal = -4.085, MI = -28.416\n",
      "Pre-Train Epoch 1, Batch 228: Loss = 192.549, a = 14.600 b = 14.889, c = -0.961, Joint = -1.630, Marginal = -3.580, MI = -9.251\n",
      "Pre-Train Epoch 1, Batch 229: Loss = 137.066, a = 26.400 b = 19.080, c = -1.525, Joint = -3.554, Marginal = -5.458, MI = -8.744\n",
      "Pre-Train Epoch 1, Batch 230: Loss = 1240.464, a = 17.579 b = 16.656, c = -1.637, Joint = 0.043, Marginal = -2.795, MI = -48.900\n",
      "Pre-Train Epoch 1, Batch 231: Loss = 290.791, a = 28.446 b = 22.680, c = -1.637, Joint = -4.934, Marginal = -6.902, MI = -21.239\n",
      "Pre-Train Epoch 1, Batch 232: Loss = 261.632, a = 23.876 b = 20.232, c = -0.442, Joint = -3.131, Marginal = -5.451, MI = -42.112\n",
      "Pre-Train Epoch 1, Batch 233: Loss = 384.511, a = 19.109 b = 18.968, c = -1.182, Joint = 1.393, Marginal = -2.891, MI = -41.632\n",
      "Pre-Train Epoch 1, Batch 234: Loss = 787.739, a = 30.359 b = 21.969, c = -1.135, Joint = 1.283, Marginal = -0.257, MI = -15.675\n",
      "Pre-Train Epoch 1, Batch 235: Loss = 267.456, a = 19.026 b = 14.033, c = -1.420, Joint = -3.984, Marginal = -3.607, MI = -10.862\n",
      "Pre-Train Epoch 1, Batch 236: Loss = 453.647, a = 18.971 b = 14.545, c = -1.638, Joint = 0.134, Marginal = -3.555, MI = -28.773\n",
      "Pre-Train Epoch 1, Batch 237: Loss = 165.147, a = 11.973 b = 16.152, c = -1.638, Joint = -0.949, Marginal = -3.922, MI = -19.671\n",
      "Pre-Train Epoch 1, Batch 238: Loss = 138.601, a = 18.512 b = 17.639, c = -1.638, Joint = -1.282, Marginal = -3.564, MI = -21.515\n",
      "Pre-Train Epoch 1, Batch 239: Loss = 274.214, a = 26.518 b = 19.410, c = -1.638, Joint = -2.048, Marginal = -5.262, MI = -39.842\n",
      "Pre-Train Epoch 1, Batch 240: Loss = 250.456, a = 24.246 b = 18.293, c = -0.068, Joint = -1.773, Marginal = -5.609, MI = -29.227\n",
      "Pre-Train Epoch 1, Batch 241: Loss = 238.369, a = 28.773 b = 20.351, c = 0.136, Joint = -3.096, Marginal = -5.001, MI = -16.159\n",
      "Pre-Train Epoch 1, Batch 242: Loss = 475.164, a = 29.736 b = 21.987, c = -1.639, Joint = 0.434, Marginal = -5.794, MI = -9.087\n",
      "Pre-Train Epoch 1, Batch 243: Loss = 512.301, a = 32.915 b = 18.568, c = -1.639, Joint = -0.290, Marginal = -3.924, MI = -19.827\n",
      "Pre-Train Epoch 1, Batch 244: Loss = 103.787, a = 18.529 b = 14.008, c = -1.639, Joint = -3.001, Marginal = -3.723, MI = -10.690\n",
      "Pre-Train Epoch 1, Batch 245: Loss = 131.623, a = 25.253 b = 18.289, c = -0.603, Joint = -3.511, Marginal = -1.762, MI = -44.701\n",
      "Pre-Train Epoch 1, Batch 246: Loss = 194.007, a = 16.128 b = 17.267, c = 0.996, Joint = -0.685, Marginal = -2.687, MI = -45.050\n",
      "Pre-Train Epoch 1, Batch 247: Loss = 123.823, a = 11.525 b = 16.195, c = 0.528, Joint = -1.804, Marginal = -3.700, MI = -19.637\n",
      "Pre-Train Epoch 1, Batch 248: Loss = 228.179, a = 24.866 b = 18.128, c = -1.640, Joint = -7.624, Marginal = -8.117, MI = -17.206\n",
      "Pre-Train Epoch 1, Batch 249: Loss = 211.357, a = 16.105 b = 16.849, c = -0.329, Joint = -2.947, Marginal = -2.916, MI = -45.037\n",
      "Pre-Train Epoch 1, Batch 250: Loss = 393.707, a = 34.309 b = 18.887, c = -0.795, Joint = -4.908, Marginal = -4.728, MI = -32.360\n",
      "Pre-Train Epoch 1, Batch 251: Loss = 77.050, a = 7.315 b = 11.447, c = -1.640, Joint = -3.962, Marginal = -1.934, MI = -18.791\n",
      "Pre-Train Epoch 1, Batch 252: Loss = 473.693, a = 24.067 b = 16.645, c = -1.640, Joint = -2.092, Marginal = -3.829, MI = -6.485\n",
      "Pre-Train Epoch 1, Batch 253: Loss = 215.687, a = 13.031 b = 14.593, c = -1.640, Joint = -1.592, Marginal = -2.428, MI = -18.048\n",
      "Pre-Train Epoch 1, Batch 254: Loss = 108.047, a = 17.352 b = 15.805, c = -1.640, Joint = -3.789, Marginal = -4.492, MI = -18.525\n",
      "Pre-Train Epoch 1, Batch 255: Loss = 1087.542, a = 39.902 b = 26.349, c = -0.529, Joint = 0.141, Marginal = -3.267, MI = -70.916\n",
      "Pre-Train Epoch 1, Batch 256: Loss = 154.829, a = 16.588 b = 15.800, c = -1.640, Joint = -2.434, Marginal = -3.122, MI = -19.091\n",
      "Pre-Train Epoch 1, Batch 257: Loss = 281.928, a = 28.566 b = 19.322, c = -1.640, Joint = -3.737, Marginal = -8.058, MI = -9.677\n",
      "Pre-Train Epoch 1, Batch 258: Loss = 118.728, a = 19.120 b = 15.921, c = 0.666, Joint = -2.072, Marginal = -3.984, MI = -12.661\n",
      "Pre-Train Epoch 1, Batch 259: Loss = 205.689, a = 17.010 b = 15.268, c = 1.035, Joint = -2.882, Marginal = -4.043, MI = -19.898\n",
      "Pre-Train Epoch 1, Batch 260: Loss = 216.187, a = 17.654 b = 19.463, c = -1.640, Joint = -3.887, Marginal = -4.450, MI = -23.312\n",
      "Pre-Train Epoch 1, Batch 261: Loss = 170.931, a = 24.778 b = 20.686, c = -1.640, Joint = -2.729, Marginal = -4.943, MI = -33.530\n",
      "Pre-Train Epoch 1, Batch 262: Loss = 155.527, a = 16.788 b = 16.176, c = -1.640, Joint = -1.811, Marginal = -4.822, MI = -15.191\n",
      "Pre-Train Epoch 1, Batch 263: Loss = 614.832, a = 16.728 b = 17.394, c = -1.573, Joint = 1.894, Marginal = -3.769, MI = -19.095\n",
      "Pre-Train Epoch 1, Batch 264: Loss = 206.958, a = 13.279 b = 19.900, c = -0.582, Joint = -1.611, Marginal = -2.989, MI = -62.506\n",
      "Pre-Train Epoch 1, Batch 265: Loss = 245.584, a = 26.811 b = 16.675, c = -0.586, Joint = -5.537, Marginal = -4.627, MI = -34.073\n",
      "Pre-Train Epoch 1, Batch 266: Loss = 88.931, a = 16.136 b = 13.331, c = -1.641, Joint = -3.839, Marginal = -4.441, MI = -13.676\n",
      "Pre-Train Epoch 1, Batch 267: Loss = 267.363, a = 19.815 b = 17.841, c = -1.641, Joint = -1.109, Marginal = -4.133, MI = -31.138\n",
      "Pre-Train Epoch 1, Batch 268: Loss = 190.290, a = 21.174 b = 17.738, c = -1.641, Joint = -0.420, Marginal = -4.139, MI = -21.089\n",
      "Pre-Train Epoch 1, Batch 269: Loss = 508.391, a = 17.998 b = 19.127, c = -0.823, Joint = -0.420, Marginal = -3.918, MI = -29.414\n",
      "Pre-Train Epoch 1, Batch 270: Loss = 359.983, a = 21.996 b = 16.008, c = 0.814, Joint = -2.065, Marginal = -4.383, MI = -24.729\n",
      "Pre-Train Epoch 1, Batch 271: Loss = 924.591, a = 13.273 b = 16.976, c = -1.641, Joint = 1.093, Marginal = -5.282, MI = -13.296\n",
      "Pre-Train Epoch 1, Batch 272: Loss = 242.472, a = 21.093 b = 16.150, c = -1.641, Joint = 0.387, Marginal = -3.491, MI = -6.109\n",
      "Pre-Train Epoch 1, Batch 273: Loss = 267.662, a = 19.852 b = 15.707, c = -1.641, Joint = -1.825, Marginal = -3.912, MI = -27.507\n",
      "Pre-Train Epoch 1, Batch 274: Loss = 299.836, a = 12.067 b = 14.925, c = -1.222, Joint = -0.603, Marginal = -3.518, MI = -68.792\n",
      "Pre-Train Epoch 1, Batch 275: Loss = 91.969, a = 11.419 b = 12.760, c = -1.287, Joint = -4.268, Marginal = -4.366, MI = -7.330\n",
      "Pre-Train Epoch 1, Batch 276: Loss = 177.230, a = 13.834 b = 16.355, c = -0.972, Joint = -2.369, Marginal = -4.951, MI = -8.732\n",
      "Pre-Train Epoch 1, Batch 277: Loss = 926.749, a = 17.368 b = 22.672, c = -0.535, Joint = -0.189, Marginal = -7.040, MI = -18.434\n",
      "Pre-Train Epoch 1, Batch 278: Loss = 276.487, a = 29.233 b = 23.783, c = -1.642, Joint = -4.019, Marginal = -6.440, MI = -12.359\n",
      "Pre-Train Epoch 1, Batch 279: Loss = 152.620, a = 14.053 b = 14.163, c = -1.642, Joint = -4.325, Marginal = -3.844, MI = -17.622\n",
      "Pre-Train Epoch 1, Batch 280: Loss = 178.932, a = 15.200 b = 17.563, c = -1.642, Joint = -0.796, Marginal = -1.743, MI = -51.101\n",
      "Pre-Train Epoch 1, Batch 281: Loss = 263.146, a = 23.426 b = 22.079, c = -0.187, Joint = -3.583, Marginal = -4.719, MI = -20.256\n",
      "Pre-Train Epoch 1, Batch 282: Loss = 944.862, a = 29.182 b = 25.535, c = -0.226, Joint = 2.165, Marginal = -2.787, MI = -51.110\n",
      "Pre-Train Epoch 1, Batch 283: Loss = 149.955, a = 22.903 b = 17.304, c = -0.727, Joint = -4.824, Marginal = -3.803, MI = -81.374\n",
      "Pre-Train Epoch 1, Batch 284: Loss = 233.922, a = 11.261 b = 15.377, c = -1.643, Joint = 0.467, Marginal = -3.511, MI = -13.297\n",
      "Pre-Train Epoch 1, Batch 285: Loss = 167.937, a = 18.489 b = 19.735, c = 0.720, Joint = -2.341, Marginal = -4.408, MI = -21.743\n",
      "Pre-Train Epoch 1, Batch 286: Loss = 377.565, a = 17.792 b = 16.167, c = -0.289, Joint = 1.031, Marginal = -1.132, MI = -52.101\n",
      "Pre-Train Epoch 1, Batch 287: Loss = 1256.592, a = 14.758 b = 17.771, c = -0.919, Joint = 5.370, Marginal = -3.828, MI = -42.072\n",
      "Pre-Train Epoch 1, Batch 288: Loss = 234.197, a = 31.186 b = 21.327, c = 0.899, Joint = -3.805, Marginal = -5.617, MI = -39.496\n",
      "Pre-Train Epoch 1, Batch 289: Loss = 142.600, a = 23.014 b = 14.288, c = -0.637, Joint = -3.044, Marginal = -3.258, MI = -10.168\n",
      "Pre-Train Epoch 1, Batch 290: Loss = 94.355, a = 13.988 b = 13.677, c = -1.478, Joint = -2.899, Marginal = -5.001, MI = -4.166\n",
      "Pre-Train Epoch 1, Batch 291: Loss = 501.243, a = 13.709 b = 15.391, c = -1.644, Joint = -1.205, Marginal = -3.899, MI = -15.397\n",
      "Pre-Train Epoch 1, Batch 292: Loss = 246.218, a = 35.230 b = 19.557, c = 0.695, Joint = -5.796, Marginal = -3.041, MI = -71.064\n",
      "Pre-Train Epoch 1, Batch 293: Loss = 139.472, a = 25.001 b = 19.369, c = -0.543, Joint = -3.648, Marginal = -6.164, MI = -13.486\n",
      "Pre-Train Epoch 1, Batch 294: Loss = 158.229, a = 16.060 b = 18.907, c = -0.845, Joint = -0.734, Marginal = -3.448, MI = -46.139\n",
      "Pre-Train Epoch 1, Batch 295: Loss = 102.855, a = 12.430 b = 13.891, c = -0.197, Joint = -1.567, Marginal = -3.404, MI = -16.514\n",
      "Pre-Train Epoch 1, Batch 296: Loss = 264.117, a = 19.599 b = 18.866, c = -0.159, Joint = -0.906, Marginal = -4.011, MI = -35.376\n",
      "Pre-Train Epoch 1, Batch 297: Loss = 281.023, a = 42.069 b = 23.543, c = -0.563, Joint = -3.557, Marginal = -6.532, MI = -16.452\n",
      "Pre-Train Epoch 1, Batch 298: Loss = 174.609, a = 19.698 b = 14.959, c = -1.573, Joint = -2.225, Marginal = -4.700, MI = -5.763\n",
      "Pre-Train Epoch 1, Batch 299: Loss = 410.331, a = 14.542 b = 19.753, c = -0.976, Joint = -0.346, Marginal = -3.419, MI = -29.803\n",
      "Pre-Train Epoch 1, Batch 300: Loss = 111.935, a = 12.922 b = 12.055, c = -1.224, Joint = -1.592, Marginal = -4.546, MI = -5.284\n",
      "Pre-Train Epoch 1, Batch 301: Loss = 175.726, a = 28.299 b = 21.222, c = -0.922, Joint = -2.506, Marginal = -2.140, MI = -46.126\n",
      "Pre-Train Epoch 1, Batch 302: Loss = 748.870, a = 21.496 b = 20.343, c = -1.645, Joint = 0.097, Marginal = -1.952, MI = -36.095\n",
      "Pre-Train Epoch 1, Batch 303: Loss = 260.822, a = 25.920 b = 19.094, c = 1.565, Joint = -1.251, Marginal = -5.331, MI = -15.662\n",
      "Pre-Train Epoch 1, Batch 304: Loss = 293.130, a = 11.065 b = 17.429, c = -1.645, Joint = 1.360, Marginal = -3.611, MI = -10.807\n",
      "Pre-Train Epoch 1, Batch 305: Loss = 146.115, a = 16.260 b = 13.240, c = -1.645, Joint = -1.580, Marginal = -2.139, MI = -18.371\n",
      "Pre-Train Epoch 1, Batch 306: Loss = 253.744, a = 16.884 b = 20.060, c = 0.231, Joint = -3.838, Marginal = -6.127, MI = -35.587\n",
      "\n",
      "Start of epoch 2\n",
      "Pre-Train Epoch 2, Batch 0: Loss = 148.807, a = 18.194 b = 16.634, c = -1.162, Joint = -2.167, Marginal = -4.151, MI = -42.480\n",
      "Pre-Train Epoch 2, Batch 1: Loss = 94.899, a = 10.754 b = 15.843, c = -1.271, Joint = -0.514, Marginal = -3.940, MI = -25.472\n",
      "Pre-Train Epoch 2, Batch 2: Loss = 143.446, a = 26.916 b = 22.582, c = -1.646, Joint = -4.564, Marginal = -5.665, MI = -32.952\n",
      "Pre-Train Epoch 2, Batch 3: Loss = 135.130, a = 19.325 b = 18.493, c = 0.208, Joint = -2.951, Marginal = -6.330, MI = -4.581\n",
      "Pre-Train Epoch 2, Batch 4: Loss = 95.863, a = 12.674 b = 14.697, c = -1.646, Joint = -2.734, Marginal = -3.726, MI = -20.279\n",
      "Pre-Train Epoch 2, Batch 5: Loss = 97.300, a = 15.551 b = 16.856, c = -0.192, Joint = -3.557, Marginal = -3.215, MI = -26.192\n",
      "Pre-Train Epoch 2, Batch 6: Loss = 251.570, a = 14.101 b = 17.552, c = -1.428, Joint = -1.407, Marginal = -2.878, MI = -24.417\n",
      "Pre-Train Epoch 2, Batch 7: Loss = 83.825, a = 9.288 b = 11.214, c = -1.466, Joint = -1.682, Marginal = -2.797, MI = -30.666\n",
      "Pre-Train Epoch 2, Batch 8: Loss = 352.827, a = 15.681 b = 17.849, c = -0.794, Joint = -1.981, Marginal = -3.230, MI = -26.353\n",
      "Pre-Train Epoch 2, Batch 9: Loss = 175.041, a = 17.193 b = 15.109, c = -1.200, Joint = -4.013, Marginal = -6.000, MI = -11.796\n",
      "Pre-Train Epoch 2, Batch 10: Loss = 303.016, a = 11.762 b = 17.165, c = -1.646, Joint = 0.431, Marginal = -4.371, MI = -10.686\n",
      "Pre-Train Epoch 2, Batch 11: Loss = 159.444, a = 16.073 b = 18.055, c = -1.228, Joint = -2.916, Marginal = -2.714, MI = -31.922\n",
      "Pre-Train Epoch 2, Batch 12: Loss = 84.792, a = 5.118 b = 12.103, c = -1.043, Joint = -0.625, Marginal = -2.104, MI = -12.387\n",
      "Pre-Train Epoch 2, Batch 13: Loss = 128.562, a = 6.951 b = 17.065, c = 0.203, Joint = -0.968, Marginal = -3.718, MI = -17.093\n",
      "Pre-Train Epoch 2, Batch 14: Loss = 215.825, a = 19.391 b = 25.235, c = -0.621, Joint = -1.919, Marginal = -6.405, MI = -36.211\n",
      "Pre-Train Epoch 2, Batch 15: Loss = 96.353, a = 11.330 b = 11.205, c = -1.646, Joint = -2.334, Marginal = -1.727, MI = -78.334\n",
      "Pre-Train Epoch 2, Batch 16: Loss = 217.019, a = 16.712 b = 13.649, c = -1.512, Joint = -0.444, Marginal = -2.711, MI = -10.230\n",
      "Pre-Train Epoch 2, Batch 17: Loss = 143.693, a = 11.047 b = 18.074, c = -0.852, Joint = -2.304, Marginal = -2.306, MI = -48.059\n",
      "Pre-Train Epoch 2, Batch 18: Loss = 199.453, a = 19.184 b = 14.925, c = -0.629, Joint = -2.422, Marginal = -3.544, MI = -38.828\n",
      "Pre-Train Epoch 2, Batch 19: Loss = 100.755, a = 19.091 b = 14.551, c = -0.637, Joint = -2.736, Marginal = -4.708, MI = -13.750\n",
      "Pre-Train Epoch 2, Batch 20: Loss = 132.458, a = 16.375 b = 18.409, c = -1.067, Joint = -3.736, Marginal = -3.677, MI = -44.964\n",
      "Pre-Train Epoch 2, Batch 21: Loss = 80.061, a = 7.318 b = 15.362, c = -0.754, Joint = -1.590, Marginal = -2.344, MI = -31.226\n",
      "Pre-Train Epoch 2, Batch 22: Loss = 131.085, a = 10.933 b = 13.730, c = 0.520, Joint = -1.988, Marginal = -3.246, MI = -39.193\n",
      "Pre-Train Epoch 2, Batch 23: Loss = 935.445, a = 16.857 b = 17.382, c = -1.212, Joint = -0.380, Marginal = -4.651, MI = -22.366\n",
      "Pre-Train Epoch 2, Batch 24: Loss = 425.795, a = 17.796 b = 19.839, c = -1.158, Joint = 2.748, Marginal = -3.478, MI = -14.344\n",
      "Pre-Train Epoch 2, Batch 25: Loss = 285.361, a = 19.952 b = 18.116, c = -1.634, Joint = -4.017, Marginal = -4.613, MI = -24.990\n",
      "Pre-Train Epoch 2, Batch 26: Loss = 156.150, a = 8.747 b = 15.213, c = 0.220, Joint = 0.536, Marginal = -5.181, MI = -9.080\n",
      "Pre-Train Epoch 2, Batch 27: Loss = 171.752, a = 12.642 b = 14.133, c = -0.966, Joint = -2.750, Marginal = -3.829, MI = -9.904\n",
      "Pre-Train Epoch 2, Batch 28: Loss = 232.970, a = 21.353 b = 17.888, c = -1.177, Joint = -2.473, Marginal = -2.617, MI = -34.631\n",
      "Pre-Train Epoch 2, Batch 29: Loss = 1512.240, a = 19.553 b = 23.415, c = -1.647, Joint = 3.223, Marginal = -3.174, MI = -35.747\n",
      "Pre-Train Epoch 2, Batch 30: Loss = 228.831, a = 32.205 b = 23.000, c = -0.170, Joint = -3.218, Marginal = -4.330, MI = -19.073\n",
      "Pre-Train Epoch 2, Batch 31: Loss = 132.093, a = 11.208 b = 15.040, c = -1.648, Joint = -1.360, Marginal = -2.838, MI = -10.802\n",
      "Pre-Train Epoch 2, Batch 32: Loss = 149.107, a = 18.274 b = 17.993, c = -1.299, Joint = -2.400, Marginal = -3.794, MI = -42.684\n",
      "Pre-Train Epoch 2, Batch 33: Loss = 146.263, a = 14.037 b = 16.887, c = -1.648, Joint = -1.237, Marginal = -3.660, MI = -15.919\n",
      "Pre-Train Epoch 2, Batch 34: Loss = 141.647, a = 18.054 b = 16.616, c = -0.838, Joint = -2.731, Marginal = -3.822, MI = -18.985\n",
      "Pre-Train Epoch 2, Batch 35: Loss = 138.831, a = 18.404 b = 16.658, c = -0.282, Joint = -2.109, Marginal = -5.617, MI = -9.673\n",
      "Pre-Train Epoch 2, Batch 36: Loss = 105.563, a = 10.561 b = 15.008, c = -0.672, Joint = -2.105, Marginal = -3.342, MI = -8.897\n",
      "Pre-Train Epoch 2, Batch 37: Loss = 238.887, a = 21.100 b = 21.718, c = 0.984, Joint = -2.713, Marginal = -4.282, MI = -23.820\n",
      "Pre-Train Epoch 2, Batch 38: Loss = 285.607, a = 21.434 b = 18.103, c = -1.649, Joint = -2.696, Marginal = -4.447, MI = -12.017\n",
      "Pre-Train Epoch 2, Batch 39: Loss = 203.602, a = 17.530 b = 22.444, c = -1.649, Joint = -3.464, Marginal = -3.865, MI = -27.940\n",
      "Pre-Train Epoch 2, Batch 40: Loss = 175.978, a = 11.284 b = 14.836, c = -1.390, Joint = -1.366, Marginal = -3.387, MI = -25.075\n",
      "Pre-Train Epoch 2, Batch 41: Loss = 148.754, a = 12.540 b = 16.017, c = -1.649, Joint = -1.409, Marginal = -3.984, MI = -12.386\n",
      "Pre-Train Epoch 2, Batch 42: Loss = 159.615, a = 16.930 b = 17.853, c = 0.455, Joint = -3.413, Marginal = -2.558, MI = -70.528\n",
      "Pre-Train Epoch 2, Batch 43: Loss = 197.016, a = 8.850 b = 14.780, c = -0.626, Joint = -0.799, Marginal = -4.247, MI = -36.637\n",
      "Pre-Train Epoch 2, Batch 44: Loss = 85.116, a = 16.840 b = 13.373, c = -1.649, Joint = -3.195, Marginal = -4.052, MI = -6.198\n",
      "Pre-Train Epoch 2, Batch 45: Loss = 176.885, a = 14.519 b = 17.430, c = -0.174, Joint = -1.061, Marginal = -4.830, MI = -5.609\n",
      "Pre-Train Epoch 2, Batch 46: Loss = 94.203, a = 10.205 b = 16.771, c = -0.998, Joint = -2.566, Marginal = -4.425, MI = -43.987\n",
      "Pre-Train Epoch 2, Batch 47: Loss = 212.239, a = 14.186 b = 17.573, c = -1.649, Joint = -2.228, Marginal = -4.929, MI = -25.083\n",
      "Pre-Train Epoch 2, Batch 48: Loss = 258.502, a = 14.134 b = 19.107, c = -1.649, Joint = -0.207, Marginal = -2.722, MI = -46.079\n",
      "Pre-Train Epoch 2, Batch 49: Loss = 64.114, a = 4.280 b = 12.032, c = -1.649, Joint = -1.836, Marginal = -3.201, MI = -9.573\n",
      "Pre-Train Epoch 2, Batch 50: Loss = 213.596, a = 10.943 b = 14.562, c = -1.649, Joint = 0.434, Marginal = -2.577, MI = -68.908\n",
      "Pre-Train Epoch 2, Batch 51: Loss = 89.286, a = 8.531 b = 13.049, c = -1.650, Joint = -1.367, Marginal = -1.135, MI = -77.550\n",
      "Pre-Train Epoch 2, Batch 52: Loss = 161.310, a = 12.565 b = 17.800, c = -1.650, Joint = -4.179, Marginal = -5.114, MI = -38.061\n",
      "Pre-Train Epoch 2, Batch 53: Loss = 744.086, a = 16.239 b = 17.918, c = -1.650, Joint = 1.649, Marginal = -0.461, MI = -52.873\n",
      "Pre-Train Epoch 2, Batch 54: Loss = 118.154, a = 12.479 b = 15.849, c = -1.650, Joint = -2.744, Marginal = -3.069, MI = -20.113\n",
      "Pre-Train Epoch 2, Batch 55: Loss = 282.618, a = 10.218 b = 18.795, c = 0.044, Joint = -0.166, Marginal = -5.504, MI = -27.967\n",
      "Pre-Train Epoch 2, Batch 56: Loss = 198.685, a = 13.234 b = 14.926, c = -1.650, Joint = -4.829, Marginal = -4.442, MI = -15.118\n",
      "Pre-Train Epoch 2, Batch 57: Loss = 1113.524, a = 14.121 b = 19.466, c = -0.905, Joint = 0.382, Marginal = -0.574, MI = -41.401\n",
      "Pre-Train Epoch 2, Batch 58: Loss = 211.359, a = 18.059 b = 20.199, c = -0.619, Joint = -2.024, Marginal = -5.321, MI = -14.723\n",
      "Pre-Train Epoch 2, Batch 59: Loss = 377.987, a = 18.234 b = 20.890, c = -0.659, Joint = -0.814, Marginal = -4.193, MI = -23.761\n",
      "Pre-Train Epoch 2, Batch 60: Loss = 624.305, a = 9.418 b = 19.110, c = -0.857, Joint = 1.966, Marginal = -3.944, MI = -39.062\n",
      "Pre-Train Epoch 2, Batch 61: Loss = 145.526, a = 13.594 b = 17.640, c = -0.420, Joint = -4.146, Marginal = -4.876, MI = -10.002\n",
      "Pre-Train Epoch 2, Batch 62: Loss = 137.015, a = 9.951 b = 15.933, c = -1.412, Joint = -1.763, Marginal = -2.424, MI = -38.744\n",
      "Pre-Train Epoch 2, Batch 63: Loss = 222.809, a = 18.518 b = 19.418, c = -1.275, Joint = -4.444, Marginal = -4.199, MI = -31.056\n",
      "Pre-Train Epoch 2, Batch 64: Loss = 416.376, a = 8.164 b = 16.754, c = -0.747, Joint = 0.858, Marginal = -3.396, MI = -37.267\n",
      "Pre-Train Epoch 2, Batch 65: Loss = 365.905, a = 17.653 b = 21.937, c = -1.628, Joint = -1.916, Marginal = -4.476, MI = -30.514\n",
      "Pre-Train Epoch 2, Batch 66: Loss = 500.698, a = 8.962 b = 20.666, c = 0.919, Joint = -1.115, Marginal = -3.424, MI = -56.700\n",
      "Pre-Train Epoch 2, Batch 67: Loss = 203.427, a = 25.916 b = 18.981, c = -1.652, Joint = -2.470, Marginal = -4.697, MI = -18.519\n",
      "Pre-Train Epoch 2, Batch 68: Loss = 163.432, a = 13.059 b = 17.839, c = -0.975, Joint = -0.808, Marginal = -2.873, MI = -16.415\n",
      "Pre-Train Epoch 2, Batch 69: Loss = 473.950, a = 14.039 b = 19.412, c = -1.652, Joint = 1.917, Marginal = -3.897, MI = -28.930\n",
      "Pre-Train Epoch 2, Batch 70: Loss = 133.498, a = 7.851 b = 14.724, c = -1.589, Joint = 0.737, Marginal = -2.905, MI = -44.367\n",
      "Pre-Train Epoch 2, Batch 71: Loss = 133.715, a = 17.963 b = 15.456, c = 1.002, Joint = -2.769, Marginal = -4.112, MI = -9.589\n",
      "Pre-Train Epoch 2, Batch 72: Loss = 342.331, a = 13.181 b = 16.134, c = -0.485, Joint = -0.975, Marginal = -2.100, MI = -49.348\n",
      "Pre-Train Epoch 2, Batch 73: Loss = 238.093, a = 14.063 b = 17.584, c = -1.653, Joint = 0.171, Marginal = -3.556, MI = -9.412\n",
      "Pre-Train Epoch 2, Batch 74: Loss = 77.011, a = 7.181 b = 12.391, c = -1.653, Joint = -2.752, Marginal = -3.269, MI = -17.235\n",
      "Pre-Train Epoch 2, Batch 75: Loss = 280.244, a = 15.488 b = 19.302, c = -0.798, Joint = -4.199, Marginal = -5.343, MI = -10.211\n",
      "Pre-Train Epoch 2, Batch 76: Loss = 354.182, a = 21.527 b = 20.042, c = -1.653, Joint = -2.589, Marginal = -5.406, MI = -32.503\n",
      "Pre-Train Epoch 2, Batch 77: Loss = 226.065, a = 19.092 b = 19.258, c = -1.653, Joint = -3.183, Marginal = -4.531, MI = -29.082\n",
      "Pre-Train Epoch 2, Batch 78: Loss = 127.796, a = 13.785 b = 16.692, c = -1.653, Joint = -1.946, Marginal = -4.071, MI = -8.379\n",
      "Pre-Train Epoch 2, Batch 79: Loss = 119.913, a = 18.632 b = 18.765, c = -0.219, Joint = -4.093, Marginal = -3.805, MI = -35.157\n",
      "Pre-Train Epoch 2, Batch 80: Loss = 177.465, a = 15.289 b = 19.438, c = -1.654, Joint = -1.175, Marginal = -5.135, MI = -13.406\n",
      "Pre-Train Epoch 2, Batch 81: Loss = 329.549, a = 11.598 b = 16.678, c = -1.654, Joint = -0.782, Marginal = -2.235, MI = -32.411\n",
      "Pre-Train Epoch 2, Batch 82: Loss = 183.723, a = 17.445 b = 17.384, c = -1.640, Joint = -2.380, Marginal = -5.189, MI = -18.232\n",
      "Pre-Train Epoch 2, Batch 83: Loss = 699.889, a = 17.154 b = 16.834, c = -0.436, Joint = 1.948, Marginal = -2.885, MI = -66.913\n",
      "Pre-Train Epoch 2, Batch 84: Loss = 269.405, a = 6.087 b = 17.882, c = -0.840, Joint = -0.623, Marginal = -6.243, MI = -12.238\n",
      "Pre-Train Epoch 2, Batch 85: Loss = 214.239, a = 13.414 b = 18.622, c = -1.235, Joint = -0.144, Marginal = -5.801, MI = -11.146\n",
      "Pre-Train Epoch 2, Batch 86: Loss = 128.517, a = 14.667 b = 19.099, c = -1.654, Joint = -2.665, Marginal = -4.388, MI = -25.418\n",
      "Pre-Train Epoch 2, Batch 87: Loss = 189.347, a = 12.895 b = 16.037, c = -0.019, Joint = -0.477, Marginal = -2.836, MI = -16.035\n",
      "Pre-Train Epoch 2, Batch 88: Loss = 102.470, a = 8.625 b = 14.633, c = -0.703, Joint = -3.502, Marginal = -3.115, MI = -8.695\n",
      "Pre-Train Epoch 2, Batch 89: Loss = 329.785, a = 15.898 b = 21.813, c = -1.655, Joint = -0.616, Marginal = -5.669, MI = -53.218\n",
      "Pre-Train Epoch 2, Batch 90: Loss = 128.898, a = 11.780 b = 18.475, c = -1.154, Joint = -5.215, Marginal = -4.464, MI = -33.356\n",
      "Pre-Train Epoch 2, Batch 91: Loss = 92.608, a = 6.251 b = 13.831, c = -1.459, Joint = -1.036, Marginal = -1.399, MI = -2.169\n",
      "Pre-Train Epoch 2, Batch 92: Loss = 441.495, a = 10.783 b = 20.272, c = -1.296, Joint = 0.416, Marginal = -5.436, MI = -11.628\n",
      "Pre-Train Epoch 2, Batch 93: Loss = 1592.620, a = 14.963 b = 19.848, c = 0.002, Joint = 1.612, Marginal = -2.954, MI = -18.222\n",
      "Pre-Train Epoch 2, Batch 94: Loss = 210.272, a = 15.267 b = 18.474, c = -1.163, Joint = -1.075, Marginal = -2.677, MI = -11.942\n",
      "Pre-Train Epoch 2, Batch 95: Loss = 206.371, a = 13.894 b = 19.683, c = -1.655, Joint = -4.909, Marginal = -4.696, MI = -59.474\n",
      "Pre-Train Epoch 2, Batch 96: Loss = 123.726, a = 5.663 b = 17.617, c = 0.164, Joint = -3.538, Marginal = -5.230, MI = -7.767\n",
      "Pre-Train Epoch 2, Batch 97: Loss = 98.536, a = 6.542 b = 14.419, c = -0.749, Joint = -2.046, Marginal = -3.958, MI = -4.777\n",
      "Pre-Train Epoch 2, Batch 98: Loss = 203.276, a = 7.181 b = 15.916, c = -1.325, Joint = -1.969, Marginal = -4.544, MI = -18.852\n",
      "Pre-Train Epoch 2, Batch 99: Loss = 238.902, a = 8.279 b = 17.215, c = 0.041, Joint = -0.190, Marginal = -4.557, MI = -17.600\n",
      "Pre-Train Epoch 2, Batch 100: Loss = 346.414, a = 9.514 b = 20.319, c = -0.849, Joint = 1.526, Marginal = -4.034, MI = -47.530\n",
      "Pre-Train Epoch 2, Batch 101: Loss = 99.335, a = 11.574 b = 14.223, c = 0.518, Joint = -3.655, Marginal = -4.422, MI = -5.351\n",
      "Pre-Train Epoch 2, Batch 102: Loss = 123.355, a = 10.917 b = 15.356, c = -1.656, Joint = -3.283, Marginal = -3.929, MI = -21.913\n",
      "Pre-Train Epoch 2, Batch 103: Loss = 247.256, a = 10.663 b = 19.323, c = -0.720, Joint = -0.542, Marginal = -3.095, MI = -79.031\n",
      "Pre-Train Epoch 2, Batch 104: Loss = 137.140, a = 15.773 b = 15.780, c = -1.229, Joint = -5.561, Marginal = -5.764, MI = -7.335\n",
      "Pre-Train Epoch 2, Batch 105: Loss = 118.617, a = 9.976 b = 14.833, c = -0.385, Joint = -3.916, Marginal = -4.545, MI = -30.491\n",
      "Pre-Train Epoch 2, Batch 106: Loss = 203.387, a = 17.386 b = 17.622, c = -0.049, Joint = -4.220, Marginal = -4.144, MI = -9.399\n",
      "Pre-Train Epoch 2, Batch 107: Loss = 321.748, a = 8.616 b = 15.557, c = -1.144, Joint = -1.517, Marginal = -5.105, MI = -13.995\n",
      "Pre-Train Epoch 2, Batch 108: Loss = 303.211, a = 15.208 b = 16.584, c = -1.657, Joint = -0.670, Marginal = -2.212, MI = -44.784\n",
      "Pre-Train Epoch 2, Batch 109: Loss = 83.000, a = 15.174 b = 14.588, c = 0.017, Joint = -4.220, Marginal = -3.971, MI = -9.661\n",
      "Pre-Train Epoch 2, Batch 110: Loss = 216.555, a = 10.094 b = 13.661, c = -1.108, Joint = -0.216, Marginal = -2.351, MI = -17.909\n",
      "Pre-Train Epoch 2, Batch 111: Loss = 341.429, a = 28.526 b = 17.241, c = -1.463, Joint = -5.459, Marginal = -6.125, MI = -13.488\n",
      "Pre-Train Epoch 2, Batch 112: Loss = 140.849, a = 9.668 b = 19.271, c = 0.365, Joint = -2.364, Marginal = -4.500, MI = -31.715\n",
      "Pre-Train Epoch 2, Batch 113: Loss = 128.115, a = 19.105 b = 16.084, c = -0.547, Joint = -3.745, Marginal = -3.877, MI = -11.273\n",
      "Pre-Train Epoch 2, Batch 114: Loss = 175.898, a = 10.981 b = 16.264, c = -0.960, Joint = -0.890, Marginal = -5.214, MI = -6.804\n",
      "Pre-Train Epoch 2, Batch 115: Loss = 177.960, a = 15.420 b = 20.326, c = 0.017, Joint = -1.934, Marginal = -0.640, MI = -80.079\n",
      "Pre-Train Epoch 2, Batch 116: Loss = 291.926, a = 10.997 b = 15.899, c = -1.657, Joint = -0.801, Marginal = -4.811, MI = -11.627\n",
      "Pre-Train Epoch 2, Batch 117: Loss = 300.150, a = 17.066 b = 18.692, c = -0.222, Joint = -0.979, Marginal = -5.813, MI = -12.218\n",
      "Pre-Train Epoch 2, Batch 118: Loss = 68.796, a = 10.944 b = 13.027, c = -1.657, Joint = -2.688, Marginal = -2.996, MI = -5.504\n",
      "Pre-Train Epoch 2, Batch 119: Loss = 182.630, a = 17.553 b = 17.017, c = -1.657, Joint = -2.107, Marginal = -4.498, MI = -58.057\n",
      "Pre-Train Epoch 2, Batch 120: Loss = 253.010, a = 20.061 b = 21.977, c = -1.657, Joint = 0.635, Marginal = -1.236, MI = -40.743\n",
      "Pre-Train Epoch 2, Batch 121: Loss = 250.029, a = 21.231 b = 20.863, c = 0.148, Joint = -1.868, Marginal = -5.771, MI = -20.143\n",
      "Pre-Train Epoch 2, Batch 122: Loss = 92.079, a = 10.232 b = 16.343, c = -1.547, Joint = -3.571, Marginal = -3.503, MI = -15.916\n",
      "Pre-Train Epoch 2, Batch 123: Loss = 353.901, a = 15.830 b = 21.672, c = -1.657, Joint = 0.006, Marginal = -4.031, MI = -33.921\n",
      "Pre-Train Epoch 2, Batch 124: Loss = 112.553, a = 17.493 b = 14.641, c = -1.657, Joint = -4.265, Marginal = -3.530, MI = -15.480\n",
      "Pre-Train Epoch 2, Batch 125: Loss = 189.168, a = 7.167 b = 19.067, c = -1.526, Joint = -0.064, Marginal = -5.923, MI = -11.692\n",
      "Pre-Train Epoch 2, Batch 126: Loss = 111.200, a = 13.850 b = 14.432, c = -0.654, Joint = -3.601, Marginal = -3.443, MI = -11.327\n",
      "Pre-Train Epoch 2, Batch 127: Loss = 252.058, a = 13.183 b = 20.509, c = -1.657, Joint = -3.769, Marginal = -6.638, MI = -19.385\n",
      "Pre-Train Epoch 2, Batch 128: Loss = 192.830, a = 12.646 b = 19.960, c = -1.205, Joint = 0.125, Marginal = -2.918, MI = -10.908\n",
      "Pre-Train Epoch 2, Batch 129: Loss = 176.484, a = 16.070 b = 17.258, c = -1.658, Joint = -0.500, Marginal = -4.544, MI = -5.591\n",
      "Pre-Train Epoch 2, Batch 130: Loss = 196.315, a = 17.649 b = 16.180, c = -1.658, Joint = -2.763, Marginal = -4.263, MI = -9.823\n",
      "Pre-Train Epoch 2, Batch 131: Loss = 96.786, a = 8.335 b = 12.583, c = -0.803, Joint = -1.479, Marginal = -3.124, MI = -5.137\n",
      "Pre-Train Epoch 2, Batch 132: Loss = 162.919, a = 10.367 b = 16.144, c = -1.658, Joint = -2.019, Marginal = -2.052, MI = -28.914\n",
      "Pre-Train Epoch 2, Batch 133: Loss = 123.168, a = 20.678 b = 18.714, c = -0.768, Joint = -3.226, Marginal = -4.885, MI = -8.955\n",
      "Pre-Train Epoch 2, Batch 134: Loss = 313.712, a = 12.446 b = 15.020, c = -1.658, Joint = -0.699, Marginal = -3.523, MI = -39.738\n",
      "Pre-Train Epoch 2, Batch 135: Loss = 308.528, a = 28.411 b = 20.728, c = -0.461, Joint = -2.318, Marginal = -3.491, MI = -27.225\n",
      "Pre-Train Epoch 2, Batch 136: Loss = 163.271, a = 13.711 b = 17.233, c = -1.472, Joint = -1.778, Marginal = -2.436, MI = -70.453\n",
      "Pre-Train Epoch 2, Batch 137: Loss = 357.963, a = 21.704 b = 21.763, c = -1.040, Joint = -0.028, Marginal = -3.806, MI = -13.000\n",
      "Pre-Train Epoch 2, Batch 138: Loss = 124.662, a = 21.063 b = 16.525, c = -1.658, Joint = -4.151, Marginal = -4.141, MI = -8.878\n",
      "Pre-Train Epoch 2, Batch 139: Loss = 77.283, a = 14.148 b = 15.987, c = -1.116, Joint = -2.034, Marginal = -3.494, MI = -42.701\n",
      "Pre-Train Epoch 2, Batch 140: Loss = 310.155, a = 35.602 b = 17.379, c = -1.508, Joint = -2.661, Marginal = -5.030, MI = -12.221\n",
      "Pre-Train Epoch 2, Batch 141: Loss = 111.246, a = 13.027 b = 16.158, c = -1.658, Joint = -3.188, Marginal = -2.668, MI = -13.217\n",
      "Pre-Train Epoch 2, Batch 142: Loss = 175.952, a = 16.928 b = 15.187, c = -1.441, Joint = -1.231, Marginal = -2.235, MI = -44.204\n",
      "Pre-Train Epoch 2, Batch 143: Loss = 217.061, a = 16.220 b = 14.967, c = -1.658, Joint = -1.650, Marginal = -2.158, MI = -42.086\n",
      "Pre-Train Epoch 2, Batch 144: Loss = 179.860, a = 12.016 b = 19.981, c = -1.611, Joint = -4.285, Marginal = -4.278, MI = -43.700\n",
      "Pre-Train Epoch 2, Batch 145: Loss = 101.762, a = 8.367 b = 12.652, c = -0.140, Joint = -3.212, Marginal = -3.643, MI = -10.141\n",
      "Pre-Train Epoch 2, Batch 146: Loss = 389.468, a = 13.353 b = 26.341, c = -0.965, Joint = 0.987, Marginal = -2.110, MI = -42.002\n",
      "Pre-Train Epoch 2, Batch 147: Loss = 528.846, a = 13.080 b = 19.214, c = -1.658, Joint = 0.244, Marginal = -3.446, MI = -73.763\n",
      "Pre-Train Epoch 2, Batch 148: Loss = 140.630, a = 13.314 b = 19.004, c = 1.603, Joint = -0.413, Marginal = -4.131, MI = -17.635\n",
      "Pre-Train Epoch 2, Batch 149: Loss = 140.809, a = 14.066 b = 14.448, c = -1.076, Joint = -2.231, Marginal = -3.130, MI = -34.516\n",
      "Pre-Train Epoch 2, Batch 150: Loss = 127.795, a = 12.562 b = 15.166, c = -0.796, Joint = -3.411, Marginal = -3.224, MI = -35.477\n",
      "Pre-Train Epoch 2, Batch 151: Loss = 166.543, a = 14.159 b = 19.418, c = -1.659, Joint = -3.658, Marginal = -6.391, MI = -26.955\n",
      "Pre-Train Epoch 2, Batch 152: Loss = 292.029, a = 29.932 b = 20.644, c = -0.195, Joint = -3.876, Marginal = -0.423, MI = -34.974\n",
      "Pre-Train Epoch 2, Batch 153: Loss = 71.921, a = 3.820 b = 12.621, c = 0.765, Joint = -2.487, Marginal = -2.136, MI = -39.383\n",
      "Pre-Train Epoch 2, Batch 154: Loss = 109.560, a = 4.908 b = 15.825, c = -0.657, Joint = -2.227, Marginal = -3.320, MI = -16.531\n",
      "Pre-Train Epoch 2, Batch 155: Loss = 130.016, a = 14.392 b = 14.118, c = -0.744, Joint = -3.282, Marginal = -4.914, MI = -16.767\n",
      "Pre-Train Epoch 2, Batch 156: Loss = 206.528, a = 24.759 b = 19.749, c = -1.659, Joint = -4.054, Marginal = -4.417, MI = -9.930\n",
      "Pre-Train Epoch 2, Batch 157: Loss = 303.493, a = 16.314 b = 21.026, c = -1.659, Joint = -0.149, Marginal = -6.551, MI = -26.475\n",
      "Pre-Train Epoch 2, Batch 158: Loss = 140.031, a = 12.851 b = 15.158, c = -1.292, Joint = -3.646, Marginal = -4.291, MI = -11.028\n",
      "Pre-Train Epoch 2, Batch 159: Loss = 152.316, a = 14.848 b = 17.825, c = -0.952, Joint = -2.576, Marginal = -4.187, MI = -4.944\n",
      "Pre-Train Epoch 2, Batch 160: Loss = 158.881, a = 16.084 b = 17.722, c = -1.659, Joint = -2.161, Marginal = -5.999, MI = -7.956\n",
      "Pre-Train Epoch 2, Batch 161: Loss = 84.454, a = 10.300 b = 14.786, c = -0.195, Joint = -3.361, Marginal = -4.168, MI = -13.776\n",
      "Pre-Train Epoch 2, Batch 162: Loss = 67.570, a = 13.494 b = 14.066, c = -1.002, Joint = -2.764, Marginal = -3.211, MI = -6.254\n",
      "Pre-Train Epoch 2, Batch 163: Loss = 99.000, a = 9.436 b = 14.128, c = -1.659, Joint = -0.176, Marginal = -4.535, MI = -3.493\n",
      "Pre-Train Epoch 2, Batch 164: Loss = 167.869, a = 7.027 b = 17.132, c = -1.042, Joint = -2.254, Marginal = -3.692, MI = -31.100\n",
      "Pre-Train Epoch 2, Batch 165: Loss = 251.702, a = 16.500 b = 17.880, c = -1.659, Joint = -3.578, Marginal = -4.299, MI = -41.709\n",
      "Pre-Train Epoch 2, Batch 166: Loss = 93.806, a = 19.255 b = 16.139, c = -1.200, Joint = -4.312, Marginal = -3.586, MI = -8.404\n",
      "Pre-Train Epoch 2, Batch 167: Loss = 112.056, a = 14.436 b = 11.357, c = 0.571, Joint = -2.832, Marginal = -2.472, MI = -6.828\n",
      "Pre-Train Epoch 2, Batch 168: Loss = 172.872, a = 4.846 b = 10.629, c = -1.096, Joint = 0.487, Marginal = -1.214, MI = -16.586\n",
      "Pre-Train Epoch 2, Batch 169: Loss = 78.256, a = 6.938 b = 16.045, c = -1.100, Joint = -2.263, Marginal = -3.739, MI = -9.861\n",
      "Pre-Train Epoch 2, Batch 170: Loss = 87.915, a = 9.840 b = 13.826, c = -1.659, Joint = -2.253, Marginal = -1.218, MI = -63.243\n",
      "Pre-Train Epoch 2, Batch 171: Loss = 148.334, a = 11.868 b = 16.355, c = -1.659, Joint = -0.720, Marginal = -3.051, MI = -49.620\n",
      "Pre-Train Epoch 2, Batch 172: Loss = 179.368, a = 14.957 b = 18.926, c = -1.250, Joint = -3.584, Marginal = -3.939, MI = -29.985\n",
      "Pre-Train Epoch 2, Batch 173: Loss = 241.268, a = 16.510 b = 18.520, c = -0.339, Joint = -0.205, Marginal = -1.309, MI = -80.480\n",
      "Pre-Train Epoch 2, Batch 174: Loss = 180.450, a = 5.792 b = 16.463, c = -0.354, Joint = -2.164, Marginal = -3.219, MI = -11.074\n",
      "Pre-Train Epoch 2, Batch 175: Loss = 64.341, a = 7.087 b = 14.856, c = -1.659, Joint = -2.785, Marginal = -3.160, MI = -19.806\n",
      "Pre-Train Epoch 2, Batch 176: Loss = 202.420, a = 13.250 b = 16.342, c = -1.659, Joint = 0.626, Marginal = -3.695, MI = -27.363\n",
      "Pre-Train Epoch 2, Batch 177: Loss = 106.131, a = 9.316 b = 17.315, c = 0.103, Joint = -4.711, Marginal = -4.453, MI = -6.363\n",
      "Pre-Train Epoch 2, Batch 178: Loss = 494.807, a = 23.052 b = 20.666, c = -1.659, Joint = -3.012, Marginal = -4.442, MI = -23.493\n",
      "Pre-Train Epoch 2, Batch 179: Loss = 427.080, a = 7.823 b = 20.763, c = -0.834, Joint = 1.148, Marginal = -5.563, MI = -12.500\n",
      "Pre-Train Epoch 2, Batch 180: Loss = 138.225, a = 7.785 b = 20.175, c = -1.659, Joint = -3.529, Marginal = -3.575, MI = -39.136\n",
      "Pre-Train Epoch 2, Batch 181: Loss = 139.995, a = 10.145 b = 15.409, c = -1.241, Joint = -1.105, Marginal = -2.379, MI = -14.127\n",
      "Pre-Train Epoch 2, Batch 182: Loss = 144.376, a = 11.065 b = 16.455, c = 1.042, Joint = -0.130, Marginal = -3.117, MI = -8.005\n",
      "Pre-Train Epoch 2, Batch 183: Loss = 109.140, a = 15.797 b = 13.745, c = -1.659, Joint = -1.950, Marginal = -2.449, MI = -12.053\n",
      "Pre-Train Epoch 2, Batch 184: Loss = 101.325, a = 7.363 b = 18.418, c = -1.138, Joint = -1.942, Marginal = -3.542, MI = -24.304\n",
      "Pre-Train Epoch 2, Batch 185: Loss = 168.269, a = 13.026 b = 15.095, c = -0.990, Joint = -2.792, Marginal = -1.829, MI = -50.670\n",
      "Pre-Train Epoch 2, Batch 186: Loss = 279.904, a = 7.710 b = 17.960, c = -0.522, Joint = -1.736, Marginal = -3.303, MI = -28.727\n",
      "Pre-Train Epoch 2, Batch 187: Loss = 300.740, a = 6.896 b = 13.432, c = -0.409, Joint = -1.504, Marginal = -2.687, MI = -37.792\n",
      "Pre-Train Epoch 2, Batch 188: Loss = 370.053, a = 4.218 b = 17.714, c = -1.660, Joint = 2.626, Marginal = -1.381, MI = -72.082\n",
      "Pre-Train Epoch 2, Batch 189: Loss = 105.005, a = 8.096 b = 16.960, c = -0.539, Joint = -0.516, Marginal = -2.793, MI = -28.163\n",
      "Pre-Train Epoch 2, Batch 190: Loss = 233.283, a = 27.287 b = 19.021, c = -1.660, Joint = -2.574, Marginal = -2.855, MI = -11.190\n",
      "Pre-Train Epoch 2, Batch 191: Loss = 299.805, a = 20.084 b = 17.504, c = -1.660, Joint = -2.176, Marginal = -0.741, MI = -30.114\n",
      "Pre-Train Epoch 2, Batch 192: Loss = 85.331, a = 9.413 b = 13.098, c = -1.046, Joint = -2.015, Marginal = -3.471, MI = -14.832\n",
      "Pre-Train Epoch 2, Batch 193: Loss = 148.270, a = 16.867 b = 16.944, c = -0.438, Joint = -2.528, Marginal = -3.157, MI = -15.215\n",
      "Pre-Train Epoch 2, Batch 194: Loss = 69.027, a = 4.158 b = 12.828, c = -1.660, Joint = -1.024, Marginal = -1.158, MI = -21.807\n",
      "Pre-Train Epoch 2, Batch 195: Loss = 189.223, a = 19.666 b = 15.742, c = -1.660, Joint = -1.523, Marginal = -3.781, MI = -11.167\n",
      "Pre-Train Epoch 2, Batch 196: Loss = 126.367, a = 11.335 b = 14.493, c = -0.618, Joint = -2.611, Marginal = -3.572, MI = -56.199\n",
      "Pre-Train Epoch 2, Batch 197: Loss = 85.022, a = 11.412 b = 11.060, c = -0.716, Joint = -2.289, Marginal = -1.349, MI = -43.595\n",
      "Pre-Train Epoch 2, Batch 198: Loss = 72.178, a = 4.371 b = 12.649, c = -1.660, Joint = -1.987, Marginal = -2.786, MI = -3.127\n",
      "Pre-Train Epoch 2, Batch 199: Loss = 101.978, a = 17.160 b = 15.420, c = -1.293, Joint = -2.260, Marginal = -2.599, MI = -36.268\n",
      "Pre-Train Epoch 2, Batch 200: Loss = 143.720, a = 13.561 b = 16.918, c = -0.772, Joint = -2.267, Marginal = -4.365, MI = -18.819\n",
      "Pre-Train Epoch 2, Batch 201: Loss = 448.102, a = 9.883 b = 18.488, c = -1.068, Joint = 0.941, Marginal = -3.908, MI = -30.695\n",
      "Pre-Train Epoch 2, Batch 202: Loss = 194.973, a = 4.748 b = 15.293, c = -1.607, Joint = -3.122, Marginal = -4.023, MI = -61.553\n",
      "Pre-Train Epoch 2, Batch 203: Loss = 167.142, a = 13.446 b = 18.496, c = -0.886, Joint = -3.123, Marginal = -4.769, MI = -40.526\n",
      "Pre-Train Epoch 2, Batch 204: Loss = 65.100, a = 7.849 b = 11.641, c = -1.660, Joint = -2.571, Marginal = -2.051, MI = -20.596\n",
      "Pre-Train Epoch 2, Batch 205: Loss = 92.314, a = 8.684 b = 14.932, c = -0.155, Joint = -1.899, Marginal = -2.872, MI = -17.052\n",
      "Pre-Train Epoch 2, Batch 206: Loss = 144.137, a = 13.020 b = 15.173, c = -1.020, Joint = -4.332, Marginal = -4.846, MI = -6.233\n",
      "Pre-Train Epoch 2, Batch 207: Loss = 70.635, a = 11.631 b = 12.474, c = -0.057, Joint = -2.243, Marginal = -2.349, MI = -26.478\n",
      "Pre-Train Epoch 2, Batch 208: Loss = 92.343, a = 11.402 b = 12.644, c = 0.718, Joint = -2.387, Marginal = -2.327, MI = -10.360\n",
      "Pre-Train Epoch 2, Batch 209: Loss = 81.856, a = 7.705 b = 14.742, c = -1.660, Joint = -1.119, Marginal = -3.324, MI = -9.943\n",
      "Pre-Train Epoch 2, Batch 210: Loss = 283.638, a = 4.171 b = 16.948, c = -0.642, Joint = -0.419, Marginal = -4.791, MI = -3.163\n",
      "Pre-Train Epoch 2, Batch 211: Loss = 107.174, a = 10.698 b = 14.419, c = -1.467, Joint = 0.194, Marginal = -2.330, MI = -5.915\n",
      "Pre-Train Epoch 2, Batch 212: Loss = 80.856, a = 6.717 b = 13.119, c = -1.661, Joint = -3.919, Marginal = -3.507, MI = -4.388\n",
      "Pre-Train Epoch 2, Batch 213: Loss = 178.981, a = 10.296 b = 16.674, c = -1.579, Joint = -1.602, Marginal = -2.944, MI = -30.563\n",
      "Pre-Train Epoch 2, Batch 214: Loss = 160.443, a = 8.530 b = 16.681, c = -1.654, Joint = -0.435, Marginal = -1.203, MI = -43.348\n",
      "Pre-Train Epoch 2, Batch 215: Loss = 193.888, a = 16.903 b = 16.452, c = -1.587, Joint = -3.531, Marginal = -3.260, MI = -76.988\n",
      "Pre-Train Epoch 2, Batch 216: Loss = 325.114, a = 9.294 b = 13.543, c = 0.298, Joint = -0.733, Marginal = -3.483, MI = -5.633\n",
      "Pre-Train Epoch 2, Batch 217: Loss = 282.337, a = 11.166 b = 18.655, c = -1.661, Joint = -1.144, Marginal = -2.607, MI = -41.529\n",
      "Pre-Train Epoch 2, Batch 218: Loss = 209.470, a = 5.592 b = 17.218, c = -1.661, Joint = -0.083, Marginal = -3.793, MI = -49.503\n",
      "Pre-Train Epoch 2, Batch 219: Loss = 104.530, a = 5.414 b = 17.088, c = -0.524, Joint = -1.650, Marginal = -3.133, MI = -25.321\n",
      "Pre-Train Epoch 2, Batch 220: Loss = 156.324, a = 8.227 b = 18.088, c = 0.088, Joint = -2.525, Marginal = -3.754, MI = -23.084\n",
      "Pre-Train Epoch 2, Batch 221: Loss = 84.798, a = 8.616 b = 14.163, c = -1.636, Joint = -2.245, Marginal = -2.801, MI = -23.093\n",
      "Pre-Train Epoch 2, Batch 222: Loss = 823.732, a = 9.162 b = 18.608, c = -1.661, Joint = 0.910, Marginal = -2.434, MI = -34.685\n",
      "Pre-Train Epoch 2, Batch 223: Loss = 250.991, a = 13.947 b = 17.557, c = 0.142, Joint = -0.662, Marginal = -3.495, MI = -21.750\n",
      "Pre-Train Epoch 2, Batch 224: Loss = 112.205, a = 16.376 b = 11.652, c = 0.664, Joint = -2.887, Marginal = -3.540, MI = -18.174\n",
      "Pre-Train Epoch 2, Batch 225: Loss = 171.548, a = 11.249 b = 19.703, c = -1.317, Joint = -2.237, Marginal = -4.554, MI = -15.316\n",
      "Pre-Train Epoch 2, Batch 226: Loss = 224.551, a = 13.823 b = 12.607, c = -1.662, Joint = -0.852, Marginal = -1.935, MI = -19.727\n",
      "Pre-Train Epoch 2, Batch 227: Loss = 103.423, a = 7.406 b = 13.008, c = -1.662, Joint = -1.551, Marginal = -3.063, MI = -5.570\n",
      "Pre-Train Epoch 2, Batch 228: Loss = 101.503, a = 3.933 b = 12.965, c = -1.050, Joint = -1.779, Marginal = -4.050, MI = -4.044\n",
      "Pre-Train Epoch 2, Batch 229: Loss = 91.365, a = 11.549 b = 16.441, c = -1.565, Joint = -3.468, Marginal = -4.739, MI = -7.642\n",
      "Pre-Train Epoch 2, Batch 230: Loss = 647.574, a = 7.486 b = 14.535, c = -1.662, Joint = -0.547, Marginal = -3.171, MI = -38.826\n",
      "Pre-Train Epoch 2, Batch 231: Loss = 166.623, a = 8.409 b = 19.520, c = -1.662, Joint = -4.887, Marginal = -5.690, MI = -18.517\n",
      "Pre-Train Epoch 2, Batch 232: Loss = 135.832, a = 8.420 b = 17.495, c = -0.582, Joint = -3.199, Marginal = -3.195, MI = -20.411\n",
      "Pre-Train Epoch 2, Batch 233: Loss = 201.997, a = 8.107 b = 16.587, c = -1.227, Joint = 0.550, Marginal = -2.225, MI = -64.110\n",
      "Pre-Train Epoch 2, Batch 234: Loss = 397.872, a = 15.522 b = 19.047, c = -1.201, Joint = 0.197, Marginal = -2.638, MI = -51.431\n",
      "Pre-Train Epoch 2, Batch 235: Loss = 173.839, a = 14.065 b = 12.260, c = -1.474, Joint = -3.358, Marginal = -2.729, MI = -7.680\n",
      "Pre-Train Epoch 2, Batch 236: Loss = 232.383, a = 8.315 b = 12.620, c = -1.663, Joint = -0.465, Marginal = -1.715, MI = -56.929\n",
      "Pre-Train Epoch 2, Batch 237: Loss = 91.422, a = 3.439 b = 14.224, c = -1.663, Joint = -1.266, Marginal = -2.545, MI = -22.701\n",
      "Pre-Train Epoch 2, Batch 238: Loss = 71.447, a = 7.952 b = 15.439, c = -1.663, Joint = -1.457, Marginal = -3.553, MI = -12.691\n",
      "Pre-Train Epoch 2, Batch 239: Loss = 168.049, a = 16.214 b = 16.854, c = -1.663, Joint = -2.011, Marginal = -4.376, MI = -8.239\n",
      "Pre-Train Epoch 2, Batch 240: Loss = 136.406, a = 9.953 b = 15.822, c = -0.284, Joint = -2.021, Marginal = -4.798, MI = -3.819\n",
      "Pre-Train Epoch 2, Batch 241: Loss = 146.288, a = 15.456 b = 17.661, c = -0.123, Joint = -3.187, Marginal = -3.443, MI = -41.208\n",
      "Pre-Train Epoch 2, Batch 242: Loss = 224.721, a = 10.956 b = 19.051, c = -1.663, Joint = -0.750, Marginal = -1.548, MI = -69.392\n",
      "Pre-Train Epoch 2, Batch 243: Loss = 260.034, a = 22.540 b = 16.067, c = -1.663, Joint = -0.693, Marginal = -3.922, MI = -3.401\n",
      "Pre-Train Epoch 2, Batch 244: Loss = 62.036, a = 7.005 b = 12.122, c = -1.663, Joint = -2.743, Marginal = -4.138, MI = -10.960\n",
      "Pre-Train Epoch 2, Batch 245: Loss = 81.578, a = 12.753 b = 15.895, c = -0.776, Joint = -3.321, Marginal = -2.376, MI = -15.684\n",
      "Pre-Train Epoch 2, Batch 246: Loss = 101.117, a = 6.029 b = 15.179, c = 0.610, Joint = -1.004, Marginal = -2.961, MI = -22.043\n",
      "Pre-Train Epoch 2, Batch 247: Loss = 71.824, a = 2.901 b = 14.249, c = 0.249, Joint = -1.847, Marginal = -2.619, MI = -55.718\n",
      "Pre-Train Epoch 2, Batch 248: Loss = 154.560, a = 9.781 b = 15.579, c = -1.664, Joint = -6.159, Marginal = -5.328, MI = -13.869\n",
      "Pre-Train Epoch 2, Batch 249: Loss = 109.839, a = 5.903 b = 14.718, c = -0.528, Joint = -2.822, Marginal = -3.446, MI = -8.725\n",
      "Pre-Train Epoch 2, Batch 250: Loss = 238.934, a = 23.794 b = 16.267, c = -0.906, Joint = -4.345, Marginal = -3.648, MI = -23.430\n",
      "Pre-Train Epoch 2, Batch 251: Loss = 47.726, a = 3.154 b = 10.114, c = -1.664, Joint = -3.066, Marginal = -2.099, MI = -8.418\n",
      "Pre-Train Epoch 2, Batch 252: Loss = 269.603, a = 12.830 b = 14.474, c = -1.664, Joint = -2.155, Marginal = -2.681, MI = -11.833\n",
      "Pre-Train Epoch 2, Batch 253: Loss = 104.820, a = 4.121 b = 12.787, c = -1.664, Joint = -1.657, Marginal = -2.193, MI = -45.706\n",
      "Pre-Train Epoch 2, Batch 254: Loss = 69.900, a = 9.025 b = 13.766, c = -1.664, Joint = -3.311, Marginal = -4.338, MI = -4.568\n",
      "Pre-Train Epoch 2, Batch 255: Loss = 533.805, a = 18.880 b = 22.739, c = -0.669, Joint = -1.072, Marginal = -4.446, MI = -46.960\n",
      "Pre-Train Epoch 2, Batch 256: Loss = 96.479, a = 8.118 b = 13.824, c = -1.664, Joint = -2.192, Marginal = -2.733, MI = -23.975\n",
      "Pre-Train Epoch 2, Batch 257: Loss = 154.507, a = 16.739 b = 16.698, c = -1.664, Joint = -3.234, Marginal = -5.205, MI = -6.173\n",
      "Pre-Train Epoch 2, Batch 258: Loss = 74.537, a = 10.059 b = 13.899, c = 0.368, Joint = -2.077, Marginal = -3.178, MI = -6.110\n",
      "Pre-Train Epoch 2, Batch 259: Loss = 117.020, a = 9.180 b = 13.268, c = 0.692, Joint = -2.382, Marginal = -3.724, MI = -11.005\n",
      "Pre-Train Epoch 2, Batch 260: Loss = 127.350, a = 9.673 b = 17.091, c = -1.664, Joint = -3.479, Marginal = -4.070, MI = -11.123\n",
      "Pre-Train Epoch 2, Batch 261: Loss = 99.570, a = 11.796 b = 17.971, c = -1.664, Joint = -2.820, Marginal = -2.018, MI = -23.636\n",
      "Pre-Train Epoch 2, Batch 262: Loss = 92.200, a = 9.269 b = 14.146, c = -1.664, Joint = -1.637, Marginal = -2.670, MI = -8.485\n",
      "Pre-Train Epoch 2, Batch 263: Loss = 274.239, a = 5.118 b = 15.230, c = -1.546, Joint = 0.750, Marginal = -2.306, MI = -24.462\n",
      "Pre-Train Epoch 2, Batch 264: Loss = 121.547, a = 4.956 b = 17.527, c = -0.752, Joint = -1.691, Marginal = -3.519, MI = -34.989\n",
      "Pre-Train Epoch 2, Batch 265: Loss = 161.496, a = 17.971 b = 14.418, c = -0.661, Joint = -4.537, Marginal = -4.692, MI = -7.826\n",
      "Pre-Train Epoch 2, Batch 266: Loss = 59.278, a = 12.241 b = 11.633, c = -1.664, Joint = -3.078, Marginal = -2.990, MI = -16.121\n",
      "Pre-Train Epoch 2, Batch 267: Loss = 146.186, a = 10.595 b = 15.543, c = -1.664, Joint = -1.264, Marginal = -2.235, MI = -24.162\n",
      "Pre-Train Epoch 2, Batch 268: Loss = 108.164, a = 12.393 b = 15.509, c = -1.664, Joint = -0.873, Marginal = -2.617, MI = -25.470\n",
      "Pre-Train Epoch 2, Batch 269: Loss = 262.274, a = 8.669 b = 16.704, c = -0.908, Joint = -0.890, Marginal = -1.591, MI = -69.820\n",
      "Pre-Train Epoch 2, Batch 270: Loss = 189.147, a = 14.427 b = 13.906, c = 0.602, Joint = -1.981, Marginal = -4.711, MI = -5.368\n",
      "Pre-Train Epoch 2, Batch 271: Loss = 412.428, a = 2.722 b = 14.822, c = -1.664, Joint = -0.034, Marginal = -2.393, MI = -49.611\n",
      "Pre-Train Epoch 2, Batch 272: Loss = 136.028, a = 14.572 b = 14.148, c = -1.664, Joint = -0.193, Marginal = -2.802, MI = -13.475\n",
      "Pre-Train Epoch 2, Batch 273: Loss = 153.987, a = 13.365 b = 13.685, c = -1.664, Joint = -1.753, Marginal = -3.965, MI = -15.338\n",
      "Pre-Train Epoch 2, Batch 274: Loss = 167.410, a = 5.621 b = 13.135, c = -1.348, Joint = -0.830, Marginal = -2.668, MI = -35.451\n",
      "Pre-Train Epoch 2, Batch 275: Loss = 62.495, a = 5.281 b = 11.151, c = -1.379, Joint = -3.646, Marginal = -2.560, MI = -4.305\n",
      "Pre-Train Epoch 2, Batch 276: Loss = 100.654, a = 4.826 b = 14.341, c = -1.094, Joint = -2.383, Marginal = -4.038, MI = -5.658\n",
      "Pre-Train Epoch 2, Batch 277: Loss = 502.972, a = 4.617 b = 19.827, c = -0.690, Joint = -1.042, Marginal = -3.511, MI = -60.477\n",
      "Pre-Train Epoch 2, Batch 278: Loss = 172.351, a = 18.732 b = 20.758, c = -1.664, Joint = -3.935, Marginal = -4.046, MI = -19.163\n",
      "Pre-Train Epoch 2, Batch 279: Loss = 101.751, a = 7.720 b = 12.381, c = -1.665, Joint = -3.748, Marginal = -2.990, MI = -9.602\n",
      "Pre-Train Epoch 2, Batch 280: Loss = 81.658, a = 4.351 b = 15.369, c = -1.665, Joint = -1.415, Marginal = -2.575, MI = -67.089\n",
      "Pre-Train Epoch 2, Batch 281: Loss = 151.386, a = 8.464 b = 19.167, c = -0.379, Joint = -3.921, Marginal = -3.659, MI = -56.586\n",
      "Pre-Train Epoch 2, Batch 282: Loss = 473.656, a = 11.948 b = 22.208, c = -0.354, Joint = 0.534, Marginal = -5.664, MI = -45.874\n",
      "Pre-Train Epoch 2, Batch 283: Loss = 103.786, a = 14.197 b = 15.045, c = -0.855, Joint = -4.129, Marginal = -4.389, MI = -9.081\n",
      "Pre-Train Epoch 2, Batch 284: Loss = 125.090, a = 5.064 b = 13.608, c = -1.665, Joint = -0.069, Marginal = -2.209, MI = -22.249\n",
      "Pre-Train Epoch 2, Batch 285: Loss = 104.044, a = 9.744 b = 17.312, c = 0.495, Joint = -2.368, Marginal = -3.932, MI = -10.262\n",
      "Pre-Train Epoch 2, Batch 286: Loss = 171.215, a = 7.917 b = 14.164, c = -0.437, Joint = 0.002, Marginal = -3.922, MI = -9.984\n",
      "Pre-Train Epoch 2, Batch 287: Loss = 717.429, a = 6.847 b = 15.714, c = -1.077, Joint = 3.813, Marginal = -2.490, MI = -74.293\n",
      "Pre-Train Epoch 2, Batch 288: Loss = 154.856, a = 21.427 b = 18.615, c = 0.563, Joint = -3.559, Marginal = -4.475, MI = -10.684\n",
      "Pre-Train Epoch 2, Batch 289: Loss = 93.009, a = 17.925 b = 12.501, c = -0.752, Joint = -2.513, Marginal = -2.480, MI = -11.274\n",
      "Pre-Train Epoch 2, Batch 290: Loss = 61.260, a = 6.586 b = 11.937, c = -1.538, Joint = -2.514, Marginal = -2.746, MI = -18.826\n",
      "Pre-Train Epoch 2, Batch 291: Loss = 298.624, a = 6.419 b = 13.520, c = -1.665, Joint = -1.266, Marginal = -3.032, MI = -31.533\n",
      "Pre-Train Epoch 2, Batch 292: Loss = 156.302, a = 26.008 b = 17.010, c = 0.332, Joint = -4.937, Marginal = -3.068, MI = -11.137\n",
      "Pre-Train Epoch 2, Batch 293: Loss = 86.481, a = 15.445 b = 16.888, c = -0.679, Joint = -3.391, Marginal = -4.200, MI = -7.604\n",
      "Pre-Train Epoch 2, Batch 294: Loss = 88.864, a = 5.476 b = 16.635, c = -0.912, Joint = -1.272, Marginal = -3.566, MI = -6.846\n",
      "Pre-Train Epoch 2, Batch 295: Loss = 63.597, a = 5.275 b = 12.217, c = -0.378, Joint = -1.572, Marginal = -2.954, MI = -6.445\n",
      "Pre-Train Epoch 2, Batch 296: Loss = 142.539, a = 9.053 b = 16.502, c = -0.330, Joint = -1.173, Marginal = -2.637, MI = -18.867\n",
      "Pre-Train Epoch 2, Batch 297: Loss = 172.753, a = 30.020 b = 20.411, c = -0.693, Joint = -3.474, Marginal = -7.028, MI = -7.579\n",
      "Pre-Train Epoch 2, Batch 298: Loss = 106.216, a = 12.457 b = 13.070, c = -1.586, Joint = -2.157, Marginal = -3.408, MI = -11.821\n",
      "Pre-Train Epoch 2, Batch 299: Loss = 229.645, a = 4.064 b = 17.373, c = -1.084, Joint = -0.899, Marginal = -4.333, MI = -34.075\n",
      "Pre-Train Epoch 2, Batch 300: Loss = 71.325, a = 6.818 b = 10.577, c = -1.260, Joint = -1.467, Marginal = -1.736, MI = -25.588\n",
      "Pre-Train Epoch 2, Batch 301: Loss = 106.730, a = 15.644 b = 18.512, c = -1.008, Joint = -2.757, Marginal = -4.029, MI = -7.345\n",
      "Pre-Train Epoch 2, Batch 302: Loss = 396.236, a = 9.302 b = 17.809, c = -1.666, Joint = -0.659, Marginal = -0.597, MI = -83.659\n",
      "Pre-Train Epoch 2, Batch 303: Loss = 153.627, a = 16.410 b = 16.715, c = 1.186, Joint = -1.688, Marginal = -2.698, MI = -58.698\n",
      "Pre-Train Epoch 2, Batch 304: Loss = 157.465, a = 3.213 b = 15.438, c = -1.666, Joint = 0.490, Marginal = -4.330, MI = -22.825\n",
      "Pre-Train Epoch 2, Batch 305: Loss = 91.633, a = 11.157 b = 11.703, c = -1.666, Joint = -1.562, Marginal = -2.220, MI = -10.848\n",
      "Pre-Train Epoch 2, Batch 306: Loss = 152.053, a = 7.469 b = 17.598, c = -0.034, Joint = -3.570, Marginal = -3.848, MI = -6.782\n",
      "\n",
      "Start of epoch 3\n",
      "Pre-Train Epoch 3, Batch 0: Loss = 84.228, a = 8.961 b = 14.505, c = -1.211, Joint = -2.216, Marginal = -4.637, MI = -7.693\n",
      "Pre-Train Epoch 3, Batch 1: Loss = 49.319, a = 3.248 b = 14.029, c = -1.330, Joint = -1.065, Marginal = -0.711, MI = -30.888\n",
      "Pre-Train Epoch 3, Batch 2: Loss = 98.472, a = 12.967 b = 19.568, c = -1.666, Joint = -4.343, Marginal = -6.667, MI = -13.890\n",
      "Pre-Train Epoch 3, Batch 3: Loss = 87.960, a = 9.235 b = 16.140, c = -0.082, Joint = -2.955, Marginal = -4.655, MI = -7.101\n",
      "Pre-Train Epoch 3, Batch 4: Loss = 57.090, a = 6.278 b = 12.885, c = -1.666, Joint = -2.492, Marginal = -3.791, MI = -12.120\n",
      "Pre-Train Epoch 3, Batch 5: Loss = 69.154, a = 9.144 b = 14.852, c = -0.382, Joint = -3.228, Marginal = -3.237, MI = -5.684\n",
      "Pre-Train Epoch 3, Batch 6: Loss = 130.532, a = 7.174 b = 15.457, c = -1.461, Joint = -1.535, Marginal = -2.565, MI = -30.264\n",
      "Pre-Train Epoch 3, Batch 7: Loss = 52.272, a = 4.466 b = 9.871, c = -1.508, Joint = -1.618, Marginal = -1.829, MI = -35.388\n",
      "Pre-Train Epoch 3, Batch 8: Loss = 227.263, a = 9.922 b = 15.780, c = -0.880, Joint = -2.112, Marginal = -3.182, MI = -20.254\n",
      "Pre-Train Epoch 3, Batch 9: Loss = 104.557, a = 10.307 b = 13.179, c = -1.323, Joint = -3.599, Marginal = -2.958, MI = -16.385\n",
      "Pre-Train Epoch 3, Batch 10: Loss = 161.378, a = 5.812 b = 15.228, c = -1.666, Joint = -0.203, Marginal = -1.465, MI = -20.360\n",
      "Pre-Train Epoch 3, Batch 11: Loss = 91.966, a = 8.633 b = 15.839, c = -1.317, Joint = -2.862, Marginal = -4.694, MI = -5.375\n",
      "Pre-Train Epoch 3, Batch 12: Loss = 53.078, a = 2.712 b = 10.838, c = -1.088, Joint = -0.654, Marginal = -1.739, MI = -10.176\n",
      "Pre-Train Epoch 3, Batch 13: Loss = 72.187, a = 2.559 b = 15.192, c = -0.003, Joint = -1.205, Marginal = -3.191, MI = -17.049\n",
      "Pre-Train Epoch 3, Batch 14: Loss = 124.300, a = 8.301 b = 22.139, c = -0.775, Joint = -2.422, Marginal = -3.058, MI = -46.326\n",
      "Pre-Train Epoch 3, Batch 15: Loss = 60.980, a = 8.608 b = 9.864, c = -1.666, Joint = -1.934, Marginal = -1.478, MI = -57.702\n",
      "Pre-Train Epoch 3, Batch 16: Loss = 122.753, a = 10.185 b = 11.932, c = -1.474, Joint = -0.658, Marginal = -2.899, MI = -6.212\n",
      "Pre-Train Epoch 3, Batch 17: Loss = 86.044, a = 4.693 b = 16.001, c = -0.940, Joint = -2.397, Marginal = -3.000, MI = -18.283\n",
      "Pre-Train Epoch 3, Batch 18: Loss = 123.651, a = 14.099 b = 13.052, c = -0.744, Joint = -2.098, Marginal = -2.523, MI = -13.737\n",
      "Pre-Train Epoch 3, Batch 19: Loss = 70.193, a = 14.098 b = 12.716, c = -0.740, Joint = -2.376, Marginal = -2.783, MI = -8.853\n",
      "Pre-Train Epoch 3, Batch 20: Loss = 89.514, a = 9.142 b = 16.123, c = -1.147, Joint = -3.453, Marginal = -4.057, MI = -21.093\n",
      "Pre-Train Epoch 3, Batch 21: Loss = 48.811, a = 3.145 b = 13.645, c = -0.860, Joint = -1.697, Marginal = -2.447, MI = -23.028\n",
      "Pre-Train Epoch 3, Batch 22: Loss = 77.354, a = 5.535 b = 12.064, c = 0.286, Joint = -1.830, Marginal = -1.909, MI = -17.831\n",
      "Pre-Train Epoch 3, Batch 23: Loss = 506.064, a = 8.979 b = 15.177, c = -1.322, Joint = -0.883, Marginal = -4.002, MI = -9.524\n",
      "Pre-Train Epoch 3, Batch 24: Loss = 172.914, a = 6.445 b = 17.383, c = -1.206, Joint = 1.060, Marginal = -1.819, MI = -61.966\n",
      "Pre-Train Epoch 3, Batch 25: Loss = 185.961, a = 9.887 b = 15.711, c = -1.666, Joint = -3.694, Marginal = -3.199, MI = -45.831\n",
      "Pre-Train Epoch 3, Batch 26: Loss = 80.189, a = 3.369 b = 13.482, c = 0.024, Joint = -0.067, Marginal = -1.378, MI = -16.022\n",
      "Pre-Train Epoch 3, Batch 27: Loss = 110.237, a = 8.831 b = 12.496, c = -1.016, Joint = -2.414, Marginal = -4.206, MI = -5.379\n",
      "Pre-Train Epoch 3, Batch 28: Loss = 148.576, a = 18.100 b = 15.768, c = -1.199, Joint = -2.201, Marginal = -1.694, MI = -49.502\n",
      "Pre-Train Epoch 3, Batch 29: Loss = 780.189, a = 8.890 b = 20.546, c = -1.666, Joint = 1.528, Marginal = -5.317, MI = -2.461\n",
      "Pre-Train Epoch 3, Batch 30: Loss = 148.908, a = 22.852 b = 19.996, c = -0.305, Joint = -3.177, Marginal = -2.917, MI = -51.295\n",
      "Pre-Train Epoch 3, Batch 31: Loss = 78.816, a = 7.158 b = 13.334, c = -1.667, Joint = -1.386, Marginal = -2.596, MI = -29.089\n",
      "Pre-Train Epoch 3, Batch 32: Loss = 86.494, a = 11.071 b = 15.766, c = -1.385, Joint = -2.373, Marginal = -4.889, MI = -6.832\n",
      "Pre-Train Epoch 3, Batch 33: Loss = 88.983, a = 8.251 b = 14.889, c = -1.667, Joint = -1.458, Marginal = -3.351, MI = -5.814\n",
      "Pre-Train Epoch 3, Batch 34: Loss = 91.903, a = 13.255 b = 14.631, c = -0.970, Joint = -2.421, Marginal = -4.528, MI = -6.584\n",
      "Pre-Train Epoch 3, Batch 35: Loss = 78.409, a = 9.228 b = 14.526, c = -0.433, Joint = -2.172, Marginal = -3.331, MI = -9.724\n",
      "Pre-Train Epoch 3, Batch 36: Loss = 62.092, a = 5.763 b = 13.256, c = -0.754, Joint = -2.058, Marginal = -3.147, MI = -12.690\n",
      "Pre-Train Epoch 3, Batch 37: Loss = 145.015, a = 11.262 b = 19.047, c = 0.659, Joint = -2.911, Marginal = -2.527, MI = -36.114\n",
      "Pre-Train Epoch 3, Batch 38: Loss = 144.349, a = 10.232 b = 15.725, c = -1.667, Joint = -2.906, Marginal = -4.018, MI = -5.537\n",
      "Pre-Train Epoch 3, Batch 39: Loss = 136.449, a = 8.402 b = 19.770, c = -1.667, Joint = -3.405, Marginal = -2.990, MI = -29.643\n",
      "Pre-Train Epoch 3, Batch 40: Loss = 109.033, a = 4.963 b = 13.073, c = -1.392, Joint = -1.528, Marginal = -2.888, MI = -8.437\n",
      "Pre-Train Epoch 3, Batch 41: Loss = 86.727, a = 7.773 b = 14.197, c = -1.667, Joint = -1.495, Marginal = -1.822, MI = -10.773\n",
      "Pre-Train Epoch 3, Batch 42: Loss = 100.184, a = 10.843 b = 15.709, c = 0.216, Joint = -3.052, Marginal = -3.860, MI = -8.365\n",
      "Pre-Train Epoch 3, Batch 43: Loss = 120.035, a = 4.283 b = 13.046, c = -0.763, Joint = -0.967, Marginal = -3.332, MI = -19.354\n",
      "Pre-Train Epoch 3, Batch 44: Loss = 60.711, a = 11.003 b = 11.687, c = -1.667, Joint = -2.799, Marginal = -2.387, MI = -12.442\n",
      "Pre-Train Epoch 3, Batch 45: Loss = 112.044, a = 7.669 b = 15.372, c = -0.456, Joint = -1.500, Marginal = -3.926, MI = -13.165\n",
      "Pre-Train Epoch 3, Batch 46: Loss = 61.258, a = 3.605 b = 14.781, c = -1.104, Joint = -2.678, Marginal = -3.925, MI = -5.210\n",
      "Pre-Train Epoch 3, Batch 47: Loss = 131.682, a = 7.065 b = 15.465, c = -1.667, Joint = -2.332, Marginal = -1.217, MI = -45.993\n",
      "Pre-Train Epoch 3, Batch 48: Loss = 122.236, a = 4.928 b = 16.790, c = -1.667, Joint = -0.951, Marginal = -3.614, MI = -9.402\n",
      "Pre-Train Epoch 3, Batch 49: Loss = 40.876, a = 2.014 b = 10.725, c = -1.667, Joint = -1.589, Marginal = -2.434, MI = -6.798\n",
      "Pre-Train Epoch 3, Batch 50: Loss = 122.204, a = 6.399 b = 12.936, c = -1.667, Joint = -0.098, Marginal = -2.054, MI = -18.478\n",
      "Pre-Train Epoch 3, Batch 51: Loss = 55.161, a = 3.596 b = 11.515, c = -1.667, Joint = -1.476, Marginal = -1.878, MI = -22.728\n",
      "Pre-Train Epoch 3, Batch 52: Loss = 107.489, a = 7.323 b = 15.694, c = -1.667, Joint = -3.775, Marginal = -4.341, MI = -7.865\n",
      "Pre-Train Epoch 3, Batch 53: Loss = 395.249, a = 8.251 b = 15.679, c = -1.667, Joint = 0.656, Marginal = -2.279, MI = -2.311\n",
      "Pre-Train Epoch 3, Batch 54: Loss = 79.910, a = 8.065 b = 14.013, c = -1.667, Joint = -2.516, Marginal = -3.621, MI = -4.498\n",
      "Pre-Train Epoch 3, Batch 55: Loss = 153.286, a = 3.843 b = 16.586, c = -0.150, Joint = -0.711, Marginal = -3.997, MI = -11.066\n",
      "Pre-Train Epoch 3, Batch 56: Loss = 131.081, a = 8.965 b = 13.130, c = -1.667, Joint = -4.005, Marginal = -2.819, MI = -11.635\n",
      "Pre-Train Epoch 3, Batch 57: Loss = 583.844, a = 6.410 b = 17.212, c = -0.997, Joint = -0.439, Marginal = -3.053, MI = -13.917\n",
      "Pre-Train Epoch 3, Batch 58: Loss = 132.454, a = 11.122 b = 17.725, c = -0.743, Joint = -2.116, Marginal = -3.904, MI = -10.373\n",
      "Pre-Train Epoch 3, Batch 59: Loss = 212.383, a = 8.684 b = 18.292, c = -0.707, Joint = -1.459, Marginal = -3.380, MI = -4.176\n",
      "Pre-Train Epoch 3, Batch 60: Loss = 336.427, a = 3.794 b = 17.025, c = -0.934, Joint = 0.843, Marginal = -2.610, MI = -15.162\n",
      "Pre-Train Epoch 3, Batch 61: Loss = 93.719, a = 7.391 b = 15.524, c = -0.566, Joint = -3.664, Marginal = -3.301, MI = -38.930\n",
      "Pre-Train Epoch 3, Batch 62: Loss = 78.625, a = 4.212 b = 14.056, c = -1.419, Joint = -1.942, Marginal = -3.493, MI = -11.247\n",
      "Pre-Train Epoch 3, Batch 63: Loss = 143.547, a = 11.971 b = 16.989, c = -1.361, Joint = -3.878, Marginal = -4.528, MI = -31.631\n",
      "Pre-Train Epoch 3, Batch 64: Loss = 233.448, a = 3.774 b = 14.900, c = -0.887, Joint = 0.158, Marginal = -0.844, MI = -10.438\n",
      "Pre-Train Epoch 3, Batch 65: Loss = 196.343, a = 10.020 b = 19.292, c = -1.543, Joint = -2.240, Marginal = -5.394, MI = -15.730\n",
      "Pre-Train Epoch 3, Batch 66: Loss = 271.872, a = 3.091 b = 18.297, c = 0.648, Joint = -1.387, Marginal = -4.240, MI = -36.079\n",
      "Pre-Train Epoch 3, Batch 67: Loss = 129.462, a = 20.552 b = 16.647, c = -1.668, Joint = -2.404, Marginal = -2.287, MI = -33.638\n",
      "Pre-Train Epoch 3, Batch 68: Loss = 101.892, a = 7.555 b = 15.836, c = -1.047, Joint = -1.123, Marginal = -1.686, MI = -12.108\n",
      "Pre-Train Epoch 3, Batch 69: Loss = 271.122, a = 7.793 b = 17.180, c = -1.669, Joint = 0.968, Marginal = -3.829, MI = -9.044\n",
      "Pre-Train Epoch 3, Batch 70: Loss = 76.900, a = 5.070 b = 13.221, c = -1.580, Joint = 0.304, Marginal = -1.598, MI = -44.295\n",
      "Pre-Train Epoch 3, Batch 71: Loss = 83.800, a = 12.569 b = 13.575, c = 0.703, Joint = -2.502, Marginal = -3.775, MI = -6.317\n",
      "Pre-Train Epoch 3, Batch 72: Loss = 216.816, a = 8.010 b = 14.316, c = -0.608, Joint = -1.235, Marginal = -1.703, MI = -28.170\n",
      "Pre-Train Epoch 3, Batch 73: Loss = 133.839, a = 7.547 b = 15.580, c = -1.669, Joint = -0.353, Marginal = -2.792, MI = -21.029\n",
      "Pre-Train Epoch 3, Batch 74: Loss = 51.249, a = 4.837 b = 11.006, c = -1.669, Joint = -2.377, Marginal = -2.830, MI = -5.884\n",
      "Pre-Train Epoch 3, Batch 75: Loss = 186.320, a = 7.536 b = 16.970, c = -0.876, Joint = -3.868, Marginal = -4.227, MI = -40.989\n",
      "Pre-Train Epoch 3, Batch 76: Loss = 215.223, a = 11.666 b = 17.594, c = -1.669, Joint = -2.551, Marginal = -4.229, MI = -15.258\n",
      "Pre-Train Epoch 3, Batch 77: Loss = 132.865, a = 13.715 b = 17.011, c = -1.669, Joint = -2.997, Marginal = -4.573, MI = -11.157\n",
      "Pre-Train Epoch 3, Batch 78: Loss = 78.977, a = 6.915 b = 14.698, c = -1.669, Joint = -1.980, Marginal = -2.957, MI = -36.680\n",
      "Pre-Train Epoch 3, Batch 79: Loss = 89.639, a = 10.604 b = 16.452, c = -0.375, Joint = -3.798, Marginal = -3.233, MI = -35.024\n",
      "Pre-Train Epoch 3, Batch 80: Loss = 104.862, a = 8.737 b = 17.172, c = -1.669, Joint = -1.598, Marginal = -4.456, MI = -10.211\n",
      "Pre-Train Epoch 3, Batch 81: Loss = 200.988, a = 7.239 b = 14.853, c = -1.669, Joint = -0.903, Marginal = -1.889, MI = -42.709\n",
      "Pre-Train Epoch 3, Batch 82: Loss = 117.267, a = 8.739 b = 15.221, c = -1.666, Joint = -2.435, Marginal = -4.327, MI = -9.476\n",
      "Pre-Train Epoch 3, Batch 83: Loss = 361.047, a = 10.863 b = 14.849, c = -0.536, Joint = 1.006, Marginal = -3.936, MI = -15.784\n",
      "Pre-Train Epoch 3, Batch 84: Loss = 150.707, a = 1.111 b = 15.913, c = -0.923, Joint = -1.087, Marginal = -4.297, MI = -8.123\n",
      "Pre-Train Epoch 3, Batch 85: Loss = 115.840, a = 6.357 b = 16.461, c = -1.311, Joint = -0.802, Marginal = -4.446, MI = -18.850\n",
      "Pre-Train Epoch 3, Batch 86: Loss = 99.230, a = 7.329 b = 16.926, c = -1.670, Joint = -2.802, Marginal = -1.020, MI = -63.830\n",
      "Pre-Train Epoch 3, Batch 87: Loss = 115.393, a = 7.997 b = 14.226, c = -0.202, Joint = -0.786, Marginal = -2.992, MI = -14.418\n",
      "Pre-Train Epoch 3, Batch 88: Loss = 73.310, a = 3.271 b = 12.926, c = -0.858, Joint = -3.290, Marginal = -3.631, MI = -10.881\n",
      "Pre-Train Epoch 3, Batch 89: Loss = 200.939, a = 8.736 b = 19.297, c = -1.670, Joint = -1.230, Marginal = -1.738, MI = -48.945\n",
      "Pre-Train Epoch 3, Batch 90: Loss = 97.463, a = 5.408 b = 16.337, c = -1.223, Joint = -4.742, Marginal = -4.820, MI = -21.946\n",
      "Pre-Train Epoch 3, Batch 91: Loss = 52.344, a = 2.736 b = 12.345, c = -1.427, Joint = -1.219, Marginal = -2.032, MI = -3.691\n",
      "Pre-Train Epoch 3, Batch 92: Loss = 255.084, a = 4.071 b = 17.953, c = -1.393, Joint = -0.226, Marginal = -5.123, MI = -12.914\n",
      "Pre-Train Epoch 3, Batch 93: Loss = 922.271, a = 9.184 b = 17.621, c = -0.166, Joint = 0.701, Marginal = -3.594, MI = -9.991\n",
      "Pre-Train Epoch 3, Batch 94: Loss = 122.405, a = 8.950 b = 16.345, c = -1.252, Joint = -1.459, Marginal = -3.704, MI = -12.786\n",
      "Pre-Train Epoch 3, Batch 95: Loss = 143.860, a = 5.173 b = 17.287, c = -1.670, Joint = -4.688, Marginal = -3.983, MI = -46.089\n",
      "Pre-Train Epoch 3, Batch 96: Loss = 82.312, a = 1.643 b = 15.679, c = -0.052, Joint = -3.345, Marginal = -3.780, MI = -11.784\n",
      "Pre-Train Epoch 3, Batch 97: Loss = 65.375, a = 2.881 b = 12.867, c = -0.828, Joint = -2.027, Marginal = -2.653, MI = -19.225\n",
      "Pre-Train Epoch 3, Batch 98: Loss = 130.988, a = 1.853 b = 14.107, c = -1.317, Joint = -2.180, Marginal = -3.695, MI = -12.148\n",
      "Pre-Train Epoch 3, Batch 99: Loss = 128.313, a = 3.364 b = 15.307, c = -0.153, Joint = -0.705, Marginal = -3.287, MI = -33.146\n",
      "Pre-Train Epoch 3, Batch 100: Loss = 163.209, a = 2.538 b = 18.056, c = -0.944, Joint = 0.325, Marginal = -4.390, MI = -7.580\n",
      "Pre-Train Epoch 3, Batch 101: Loss = 68.123, a = 6.689 b = 12.510, c = 0.306, Joint = -3.251, Marginal = -3.158, MI = -3.654\n",
      "Pre-Train Epoch 3, Batch 102: Loss = 77.133, a = 5.368 b = 13.482, c = -1.671, Joint = -2.989, Marginal = -4.222, MI = -5.473\n",
      "Pre-Train Epoch 3, Batch 103: Loss = 148.256, a = 5.735 b = 17.200, c = -0.813, Joint = -0.929, Marginal = -2.348, MI = -28.232\n",
      "Pre-Train Epoch 3, Batch 104: Loss = 98.463, a = 10.852 b = 13.832, c = -1.260, Joint = -4.690, Marginal = -4.431, MI = -12.650\n",
      "Pre-Train Epoch 3, Batch 105: Loss = 84.569, a = 5.585 b = 13.063, c = -0.566, Joint = -3.435, Marginal = -4.038, MI = -16.510\n",
      "Pre-Train Epoch 3, Batch 106: Loss = 144.531, a = 11.799 b = 15.494, c = -0.254, Joint = -3.903, Marginal = -4.763, MI = -6.572\n",
      "Pre-Train Epoch 3, Batch 107: Loss = 186.886, a = 4.832 b = 13.819, c = -1.198, Joint = -1.596, Marginal = -3.480, MI = -23.056\n",
      "Pre-Train Epoch 3, Batch 108: Loss = 174.638, a = 10.199 b = 14.642, c = -1.655, Joint = -0.845, Marginal = -1.901, MI = -30.533\n",
      "Pre-Train Epoch 3, Batch 109: Loss = 62.474, a = 11.968 b = 12.845, c = -0.209, Joint = -3.526, Marginal = -3.200, MI = -9.892\n",
      "Pre-Train Epoch 3, Batch 110: Loss = 126.601, a = 6.906 b = 12.137, c = -1.176, Joint = -0.494, Marginal = -1.704, MI = -28.733\n",
      "Pre-Train Epoch 3, Batch 111: Loss = 235.724, a = 26.070 b = 15.093, c = -1.523, Joint = -4.633, Marginal = -4.587, MI = -16.375\n",
      "Pre-Train Epoch 3, Batch 112: Loss = 94.901, a = 5.181 b = 17.125, c = 0.151, Joint = -2.367, Marginal = -3.687, MI = -6.029\n",
      "Pre-Train Epoch 3, Batch 113: Loss = 93.040, a = 15.172 b = 14.169, c = -0.632, Joint = -3.320, Marginal = -3.085, MI = -9.155\n",
      "Pre-Train Epoch 3, Batch 114: Loss = 102.498, a = 6.303 b = 14.357, c = -1.064, Joint = -1.091, Marginal = -3.352, MI = -15.966\n",
      "Pre-Train Epoch 3, Batch 115: Loss = 110.293, a = 10.679 b = 18.083, c = -0.151, Joint = -2.141, Marginal = -2.230, MI = -44.887\n",
      "Pre-Train Epoch 3, Batch 116: Loss = 175.862, a = 7.015 b = 14.048, c = -1.671, Joint = -1.003, Marginal = -2.983, MI = -16.649\n",
      "Pre-Train Epoch 3, Batch 117: Loss = 195.513, a = 11.752 b = 16.509, c = -0.366, Joint = -1.133, Marginal = -1.788, MI = -65.952\n",
      "Pre-Train Epoch 3, Batch 118: Loss = 47.154, a = 7.449 b = 11.527, c = -1.671, Joint = -2.466, Marginal = -2.743, MI = -4.191\n",
      "Pre-Train Epoch 3, Batch 119: Loss = 103.510, a = 11.228 b = 14.930, c = -1.671, Joint = -2.226, Marginal = -3.074, MI = -11.361\n",
      "Pre-Train Epoch 3, Batch 120: Loss = 152.940, a = 13.167 b = 19.437, c = -1.671, Joint = -0.133, Marginal = -3.163, MI = -22.671\n",
      "Pre-Train Epoch 3, Batch 121: Loss = 148.036, a = 17.799 b = 18.459, c = -0.051, Joint = -1.904, Marginal = -3.447, MI = -14.469\n",
      "Pre-Train Epoch 3, Batch 122: Loss = 64.667, a = 6.599 b = 14.491, c = -1.572, Joint = -3.185, Marginal = -3.030, MI = -14.279\n",
      "Pre-Train Epoch 3, Batch 123: Loss = 218.108, a = 9.350 b = 19.156, c = -1.671, Joint = -0.642, Marginal = -1.842, MI = -55.155\n",
      "Pre-Train Epoch 3, Batch 124: Loss = 78.923, a = 13.712 b = 12.898, c = -1.671, Joint = -3.555, Marginal = -3.923, MI = -3.613\n",
      "Pre-Train Epoch 3, Batch 125: Loss = 93.887, a = 2.857 b = 16.959, c = -1.590, Joint = -0.672, Marginal = -4.628, MI = -14.440\n",
      "Pre-Train Epoch 3, Batch 126: Loss = 86.586, a = 9.576 b = 12.775, c = -0.749, Joint = -3.305, Marginal = -3.514, MI = -3.725\n",
      "Pre-Train Epoch 3, Batch 127: Loss = 157.841, a = 7.426 b = 18.088, c = -1.671, Joint = -3.523, Marginal = -4.253, MI = -44.918\n",
      "Pre-Train Epoch 3, Batch 128: Loss = 91.624, a = 5.552 b = 17.604, c = -1.240, Joint = -0.666, Marginal = -3.840, MI = -40.021\n",
      "Pre-Train Epoch 3, Batch 129: Loss = 100.670, a = 11.104 b = 15.273, c = -1.671, Joint = -0.964, Marginal = -2.727, MI = -10.872\n",
      "Pre-Train Epoch 3, Batch 130: Loss = 132.851, a = 13.512 b = 14.295, c = -1.671, Joint = -2.531, Marginal = -2.563, MI = -10.067\n",
      "Pre-Train Epoch 3, Batch 131: Loss = 61.339, a = 6.555 b = 11.274, c = -0.892, Joint = -1.432, Marginal = -3.034, MI = -2.819\n",
      "Pre-Train Epoch 3, Batch 132: Loss = 107.754, a = 7.395 b = 14.379, c = -1.671, Joint = -1.918, Marginal = -1.814, MI = -15.007\n",
      "Pre-Train Epoch 3, Batch 133: Loss = 87.429, a = 13.875 b = 16.413, c = -0.810, Joint = -3.185, Marginal = -3.893, MI = -6.880\n",
      "Pre-Train Epoch 3, Batch 134: Loss = 204.903, a = 8.456 b = 13.351, c = -1.671, Joint = -0.872, Marginal = -1.551, MI = -58.919\n",
      "Pre-Train Epoch 3, Batch 135: Loss = 198.534, a = 21.858 b = 18.269, c = -0.635, Joint = -2.478, Marginal = -3.240, MI = -23.177\n",
      "Pre-Train Epoch 3, Batch 136: Loss = 99.145, a = 9.722 b = 15.285, c = -1.529, Joint = -1.797, Marginal = -3.315, MI = -11.695\n",
      "Pre-Train Epoch 3, Batch 137: Loss = 179.575, a = 14.591 b = 19.202, c = -1.085, Joint = -0.724, Marginal = -2.558, MI = -19.285\n",
      "Pre-Train Epoch 3, Batch 138: Loss = 83.299, a = 15.040 b = 14.446, c = -1.671, Joint = -3.725, Marginal = -4.242, MI = -9.107\n",
      "Pre-Train Epoch 3, Batch 139: Loss = 52.687, a = 8.873 b = 14.101, c = -1.216, Joint = -2.257, Marginal = -3.621, MI = -5.677\n",
      "Pre-Train Epoch 3, Batch 140: Loss = 189.140, a = 28.581 b = 15.178, c = -1.547, Joint = -2.506, Marginal = -2.964, MI = -33.366\n",
      "Pre-Train Epoch 3, Batch 141: Loss = 79.077, a = 9.437 b = 14.377, c = -1.664, Joint = -2.909, Marginal = -2.582, MI = -8.261\n",
      "Pre-Train Epoch 3, Batch 142: Loss = 113.510, a = 12.045 b = 13.486, c = -1.490, Joint = -1.363, Marginal = -1.298, MI = -31.174\n",
      "Pre-Train Epoch 3, Batch 143: Loss = 132.403, a = 11.406 b = 13.238, c = -1.671, Joint = -1.604, Marginal = -3.260, MI = -5.094\n",
      "Pre-Train Epoch 3, Batch 144: Loss = 120.516, a = 7.465 b = 17.662, c = -1.618, Joint = -3.959, Marginal = -3.929, MI = -30.934\n",
      "Pre-Train Epoch 3, Batch 145: Loss = 69.101, a = 6.215 b = 11.301, c = -0.268, Joint = -2.705, Marginal = -2.545, MI = -8.906\n",
      "Pre-Train Epoch 3, Batch 146: Loss = 218.955, a = 7.850 b = 23.469, c = -1.043, Joint = -0.112, Marginal = -2.771, MI = -19.894\n",
      "Pre-Train Epoch 3, Batch 147: Loss = 315.304, a = 9.728 b = 17.160, c = -1.671, Joint = -0.280, Marginal = -1.460, MI = -45.191\n",
      "Pre-Train Epoch 3, Batch 148: Loss = 89.895, a = 8.596 b = 16.949, c = 1.272, Joint = -0.944, Marginal = -1.564, MI = -18.191\n",
      "Pre-Train Epoch 3, Batch 149: Loss = 89.431, a = 10.862 b = 12.832, c = -1.108, Joint = -2.139, Marginal = -3.094, MI = -6.186\n",
      "Pre-Train Epoch 3, Batch 150: Loss = 86.550, a = 9.435 b = 13.431, c = -0.862, Joint = -2.993, Marginal = -3.343, MI = -20.686\n",
      "Pre-Train Epoch 3, Batch 151: Loss = 112.904, a = 7.912 b = 17.093, c = -1.671, Joint = -3.543, Marginal = -5.849, MI = -5.816\n",
      "Pre-Train Epoch 3, Batch 152: Loss = 204.988, a = 22.767 b = 18.180, c = -0.335, Joint = -3.814, Marginal = -2.099, MI = -50.042\n",
      "Pre-Train Epoch 3, Batch 153: Loss = 51.468, a = 1.570 b = 11.290, c = 0.602, Joint = -2.341, Marginal = -2.855, MI = -8.110\n",
      "Pre-Train Epoch 3, Batch 154: Loss = 71.794, a = 3.252 b = 14.176, c = -0.770, Joint = -2.065, Marginal = -2.368, MI = -35.254\n",
      "Pre-Train Epoch 3, Batch 155: Loss = 90.036, a = 10.941 b = 12.468, c = -0.821, Joint = -2.836, Marginal = -2.798, MI = -28.297\n",
      "Pre-Train Epoch 3, Batch 156: Loss = 144.759, a = 17.291 b = 17.397, c = -1.671, Joint = -3.992, Marginal = -3.824, MI = -7.758\n",
      "Pre-Train Epoch 3, Batch 157: Loss = 183.169, a = 9.987 b = 18.612, c = -1.671, Joint = -0.831, Marginal = -3.119, MI = -32.498\n",
      "Pre-Train Epoch 3, Batch 158: Loss = 100.574, a = 8.230 b = 13.384, c = -1.372, Joint = -3.302, Marginal = -3.763, MI = -4.073\n",
      "Pre-Train Epoch 3, Batch 159: Loss = 94.548, a = 8.585 b = 15.717, c = -1.057, Joint = -2.770, Marginal = -2.979, MI = -19.760\n",
      "Pre-Train Epoch 3, Batch 160: Loss = 101.354, a = 9.815 b = 15.569, c = -1.671, Joint = -2.345, Marginal = -3.873, MI = -3.824\n",
      "Pre-Train Epoch 3, Batch 161: Loss = 61.420, a = 7.390 b = 13.090, c = -0.372, Joint = -3.058, Marginal = -3.301, MI = -7.080\n",
      "Pre-Train Epoch 3, Batch 162: Loss = 51.626, a = 9.565 b = 12.487, c = -1.082, Joint = -2.584, Marginal = -2.490, MI = -7.069\n",
      "Pre-Train Epoch 3, Batch 163: Loss = 61.115, a = 5.411 b = 12.555, c = -1.671, Joint = -0.644, Marginal = -1.774, MI = -27.869\n",
      "Pre-Train Epoch 3, Batch 164: Loss = 93.178, a = 3.804 b = 15.214, c = -1.113, Joint = -2.421, Marginal = -4.082, MI = -14.378\n",
      "Pre-Train Epoch 3, Batch 165: Loss = 165.323, a = 13.051 b = 15.820, c = -1.671, Joint = -3.285, Marginal = -2.491, MI = -20.720\n",
      "Pre-Train Epoch 3, Batch 166: Loss = 70.961, a = 15.898 b = 14.249, c = -1.270, Joint = -3.709, Marginal = -3.192, MI = -10.605\n",
      "Pre-Train Epoch 3, Batch 167: Loss = 82.009, a = 11.860 b = 10.063, c = 0.376, Joint = -2.496, Marginal = -3.292, MI = -5.187\n",
      "Pre-Train Epoch 3, Batch 168: Loss = 106.928, a = 3.488 b = 9.553, c = -1.171, Joint = 0.171, Marginal = -1.828, MI = -13.385\n",
      "Pre-Train Epoch 3, Batch 169: Loss = 51.768, a = 4.109 b = 14.268, c = -1.156, Joint = -2.290, Marginal = -3.366, MI = -12.020\n",
      "Pre-Train Epoch 3, Batch 170: Loss = 63.499, a = 7.689 b = 12.347, c = -1.670, Joint = -1.957, Marginal = -1.017, MI = -18.588\n",
      "Pre-Train Epoch 3, Batch 171: Loss = 100.725, a = 8.434 b = 14.479, c = -1.670, Joint = -0.910, Marginal = -3.814, MI = -5.220\n",
      "Pre-Train Epoch 3, Batch 172: Loss = 117.811, a = 12.341 b = 16.764, c = -1.315, Joint = -3.243, Marginal = -3.899, MI = -30.407\n",
      "Pre-Train Epoch 3, Batch 173: Loss = 150.846, a = 12.853 b = 16.390, c = -0.460, Joint = -0.678, Marginal = -2.669, MI = -40.306\n",
      "Pre-Train Epoch 3, Batch 174: Loss = 113.517, a = 3.937 b = 14.628, c = -0.477, Joint = -2.099, Marginal = -3.394, MI = -22.085\n",
      "Pre-Train Epoch 3, Batch 175: Loss = 45.934, a = 5.266 b = 13.193, c = -1.670, Joint = -2.560, Marginal = -2.783, MI = -6.577\n",
      "Pre-Train Epoch 3, Batch 176: Loss = 111.759, a = 9.553 b = 14.420, c = -1.670, Joint = 0.137, Marginal = -3.397, MI = -4.555\n",
      "Pre-Train Epoch 3, Batch 177: Loss = 80.075, a = 6.373 b = 15.343, c = -0.100, Joint = -4.287, Marginal = -3.964, MI = -18.857\n",
      "Pre-Train Epoch 3, Batch 178: Loss = 308.811, a = 20.167 b = 18.316, c = -1.670, Joint = -2.910, Marginal = -3.651, MI = -25.997\n",
      "Pre-Train Epoch 3, Batch 179: Loss = 230.069, a = 4.115 b = 18.398, c = -0.919, Joint = 0.179, Marginal = -3.659, MI = -17.788\n",
      "Pre-Train Epoch 3, Batch 180: Loss = 104.212, a = 4.748 b = 17.832, c = -1.670, Joint = -3.327, Marginal = -2.524, MI = -54.176\n",
      "Pre-Train Epoch 3, Batch 181: Loss = 80.790, a = 7.894 b = 13.724, c = -1.317, Joint = -1.299, Marginal = -3.084, MI = -12.992\n",
      "Pre-Train Epoch 3, Batch 182: Loss = 76.270, a = 8.202 b = 14.581, c = 0.741, Joint = -0.517, Marginal = -4.279, MI = -8.032\n",
      "Pre-Train Epoch 3, Batch 183: Loss = 77.846, a = 13.142 b = 12.208, c = -1.670, Joint = -1.820, Marginal = -1.626, MI = -14.043\n",
      "Pre-Train Epoch 3, Batch 184: Loss = 69.107, a = 5.054 b = 16.368, c = -1.171, Joint = -2.085, Marginal = -4.240, MI = -19.966\n",
      "Pre-Train Epoch 3, Batch 185: Loss = 112.351, a = 11.170 b = 13.453, c = -1.071, Joint = -2.570, Marginal = -3.111, MI = -17.964\n",
      "Pre-Train Epoch 3, Batch 186: Loss = 180.833, a = 5.042 b = 15.953, c = -0.620, Joint = -1.743, Marginal = -3.387, MI = -24.245\n",
      "Pre-Train Epoch 3, Batch 187: Loss = 193.186, a = 5.032 b = 11.944, c = -0.536, Joint = -1.501, Marginal = -3.016, MI = -2.409\n",
      "Pre-Train Epoch 3, Batch 188: Loss = 231.376, a = 3.133 b = 15.971, c = -1.670, Joint = 1.699, Marginal = -2.617, MI = -7.498\n",
      "Pre-Train Epoch 3, Batch 189: Loss = 67.228, a = 6.258 b = 15.166, c = -0.658, Joint = -0.785, Marginal = -1.731, MI = -41.708\n",
      "Pre-Train Epoch 3, Batch 190: Loss = 160.685, a = 23.183 b = 16.813, c = -1.670, Joint = -2.445, Marginal = -1.085, MI = -19.748\n",
      "Pre-Train Epoch 3, Batch 191: Loss = 196.890, a = 18.857 b = 15.595, c = -1.670, Joint = -2.105, Marginal = -3.038, MI = -6.507\n",
      "Pre-Train Epoch 3, Batch 192: Loss = 58.053, a = 7.165 b = 11.603, c = -1.117, Joint = -1.871, Marginal = -2.579, MI = -4.486\n",
      "Pre-Train Epoch 3, Batch 193: Loss = 104.244, a = 13.337 b = 15.009, c = -0.596, Joint = -2.365, Marginal = -2.865, MI = -16.120\n",
      "Pre-Train Epoch 3, Batch 194: Loss = 47.402, a = 3.091 b = 11.518, c = -1.670, Joint = -1.127, Marginal = -2.059, MI = -11.326\n",
      "Pre-Train Epoch 3, Batch 195: Loss = 121.091, a = 16.105 b = 13.890, c = -1.670, Joint = -1.562, Marginal = -1.705, MI = -43.599\n",
      "Pre-Train Epoch 3, Batch 196: Loss = 77.153, a = 9.432 b = 12.885, c = -0.759, Joint = -2.360, Marginal = -2.801, MI = -5.961\n",
      "Pre-Train Epoch 3, Batch 197: Loss = 61.368, a = 9.529 b = 9.873, c = -0.815, Joint = -2.008, Marginal = -1.849, MI = -11.579\n",
      "Pre-Train Epoch 3, Batch 198: Loss = 53.777, a = 3.130 b = 11.337, c = -1.670, Joint = -1.957, Marginal = -1.593, MI = -20.081\n",
      "Pre-Train Epoch 3, Batch 199: Loss = 73.164, a = 13.557 b = 13.687, c = -1.368, Joint = -2.153, Marginal = -2.389, MI = -36.766\n",
      "Pre-Train Epoch 3, Batch 200: Loss = 92.953, a = 9.485 b = 15.025, c = -0.875, Joint = -2.343, Marginal = -3.876, MI = -7.357\n",
      "Pre-Train Epoch 3, Batch 201: Loss = 248.378, a = 6.841 b = 16.508, c = -1.126, Joint = 0.184, Marginal = -2.331, MI = -52.315\n",
      "Pre-Train Epoch 3, Batch 202: Loss = 124.058, a = 2.813 b = 13.598, c = -1.603, Joint = -2.899, Marginal = -4.455, MI = -10.745\n",
      "Pre-Train Epoch 3, Batch 203: Loss = 111.909, a = 8.791 b = 16.299, c = -1.005, Joint = -2.987, Marginal = -4.238, MI = -48.090\n",
      "Pre-Train Epoch 3, Batch 204: Loss = 49.473, a = 5.344 b = 10.384, c = -1.670, Joint = -2.378, Marginal = -2.351, MI = -7.343\n",
      "Pre-Train Epoch 3, Batch 205: Loss = 64.394, a = 6.806 b = 13.396, c = -0.279, Joint = -1.903, Marginal = -2.760, MI = -11.581\n",
      "Pre-Train Epoch 3, Batch 206: Loss = 101.402, a = 9.953 b = 13.423, c = -1.079, Joint = -3.780, Marginal = -4.239, MI = -5.528\n",
      "Pre-Train Epoch 3, Batch 207: Loss = 50.157, a = 8.230 b = 11.052, c = -0.142, Joint = -2.165, Marginal = -3.012, MI = -3.879\n",
      "Pre-Train Epoch 3, Batch 208: Loss = 69.071, a = 7.604 b = 11.156, c = 0.497, Joint = -2.364, Marginal = -2.908, MI = -6.812\n",
      "Pre-Train Epoch 3, Batch 209: Loss = 54.859, a = 5.551 b = 13.249, c = -1.670, Joint = -1.389, Marginal = -3.413, MI = -7.642\n",
      "Pre-Train Epoch 3, Batch 210: Loss = 160.440, a = 1.944 b = 15.111, c = -0.754, Joint = -0.858, Marginal = -4.000, MI = -37.141\n",
      "Pre-Train Epoch 3, Batch 211: Loss = 70.243, a = 7.831 b = 12.964, c = -1.489, Joint = -0.325, Marginal = -1.791, MI = -2.982\n",
      "Pre-Train Epoch 3, Batch 212: Loss = 58.335, a = 5.126 b = 11.693, c = -1.670, Joint = -3.456, Marginal = -2.502, MI = -12.391\n",
      "Pre-Train Epoch 3, Batch 213: Loss = 111.242, a = 7.451 b = 14.815, c = -1.583, Joint = -1.789, Marginal = -2.597, MI = -13.487\n",
      "Pre-Train Epoch 3, Batch 214: Loss = 100.406, a = 6.388 b = 14.950, c = -1.623, Joint = -0.773, Marginal = -1.402, MI = -18.813\n",
      "Pre-Train Epoch 3, Batch 215: Loss = 125.882, a = 13.618 b = 14.492, c = -1.621, Joint = -3.214, Marginal = -3.998, MI = -7.968\n",
      "Pre-Train Epoch 3, Batch 216: Loss = 182.853, a = 6.338 b = 12.005, c = 0.134, Joint = -1.064, Marginal = -1.201, MI = -74.829\n",
      "Pre-Train Epoch 3, Batch 217: Loss = 176.190, a = 9.276 b = 16.657, c = -1.670, Joint = -1.283, Marginal = -2.509, MI = -14.335\n",
      "Pre-Train Epoch 3, Batch 218: Loss = 150.870, a = 3.750 b = 15.424, c = -1.670, Joint = -0.554, Marginal = -0.295, MI = -67.975\n",
      "Pre-Train Epoch 3, Batch 219: Loss = 69.912, a = 4.262 b = 15.265, c = -0.626, Joint = -1.740, Marginal = -3.860, MI = -2.832\n",
      "Pre-Train Epoch 3, Batch 220: Loss = 108.742, a = 6.631 b = 16.172, c = -0.078, Joint = -2.487, Marginal = -2.910, MI = -32.489\n",
      "Pre-Train Epoch 3, Batch 221: Loss = 56.653, a = 6.396 b = 12.559, c = -1.654, Joint = -2.165, Marginal = -3.871, MI = -6.999\n",
      "Pre-Train Epoch 3, Batch 222: Loss = 448.647, a = 6.766 b = 16.527, c = -1.670, Joint = 0.240, Marginal = -4.979, MI = -4.770\n",
      "Pre-Train Epoch 3, Batch 223: Loss = 153.920, a = 11.054 b = 15.576, c = -0.047, Joint = -0.858, Marginal = -2.711, MI = -9.406\n",
      "Pre-Train Epoch 3, Batch 224: Loss = 76.714, a = 14.920 b = 10.323, c = 0.435, Joint = -2.462, Marginal = -2.397, MI = -8.459\n",
      "Pre-Train Epoch 3, Batch 225: Loss = 112.672, a = 9.731 b = 17.547, c = -1.334, Joint = -2.178, Marginal = -4.265, MI = -32.096\n",
      "Pre-Train Epoch 3, Batch 226: Loss = 126.699, a = 11.582 b = 11.166, c = -1.671, Joint = -0.993, Marginal = -2.689, MI = -2.317\n",
      "Pre-Train Epoch 3, Batch 227: Loss = 55.507, a = 4.955 b = 11.579, c = -1.671, Joint = -1.686, Marginal = -3.232, MI = -5.039\n",
      "Pre-Train Epoch 3, Batch 228: Loss = 65.856, a = 2.423 b = 11.597, c = -1.099, Joint = -1.818, Marginal = -3.400, MI = -4.835\n",
      "Pre-Train Epoch 3, Batch 229: Loss = 70.421, a = 7.825 b = 14.553, c = -1.580, Joint = -3.302, Marginal = -3.506, MI = -5.807\n",
      "Pre-Train Epoch 3, Batch 230: Loss = 395.082, a = 5.431 b = 13.015, c = -1.671, Joint = -0.870, Marginal = -1.497, MI = -38.023\n",
      "Pre-Train Epoch 3, Batch 231: Loss = 117.069, a = 4.948 b = 17.255, c = -1.671, Joint = -4.582, Marginal = -5.728, MI = -5.179\n",
      "Pre-Train Epoch 3, Batch 232: Loss = 89.900, a = 5.661 b = 15.535, c = -0.666, Joint = -3.098, Marginal = -4.071, MI = -16.414\n",
      "Pre-Train Epoch 3, Batch 233: Loss = 127.090, a = 6.151 b = 14.878, c = -1.251, Joint = 0.107, Marginal = -1.323, MI = -28.030\n",
      "Pre-Train Epoch 3, Batch 234: Loss = 239.986, a = 11.830 b = 16.946, c = -1.235, Joint = -0.296, Marginal = -3.037, MI = -22.134\n",
      "Pre-Train Epoch 3, Batch 235: Loss = 124.088, a = 12.819 b = 10.987, c = -1.497, Joint = -2.942, Marginal = -2.759, MI = -2.952\n",
      "Pre-Train Epoch 3, Batch 236: Loss = 141.323, a = 6.090 b = 11.241, c = -1.671, Joint = -0.693, Marginal = -3.432, MI = -24.536\n",
      "Pre-Train Epoch 3, Batch 237: Loss = 62.442, a = 2.914 b = 12.839, c = -1.671, Joint = -1.356, Marginal = -2.361, MI = -16.719\n",
      "Pre-Train Epoch 3, Batch 238: Loss = 47.537, a = 5.962 b = 13.859, c = -1.672, Joint = -1.549, Marginal = -3.258, MI = -4.489\n",
      "Pre-Train Epoch 3, Batch 239: Loss = 118.297, a = 13.605 b = 15.013, c = -1.672, Joint = -1.942, Marginal = -3.146, MI = -8.687\n",
      "Pre-Train Epoch 3, Batch 240: Loss = 90.339, a = 6.764 b = 14.042, c = -0.413, Joint = -2.042, Marginal = -2.726, MI = -56.658\n",
      "Pre-Train Epoch 3, Batch 241: Loss = 105.456, a = 11.470 b = 15.714, c = -0.277, Joint = -3.079, Marginal = -2.747, MI = -12.927\n",
      "Pre-Train Epoch 3, Batch 242: Loss = 133.177, a = 6.897 b = 16.928, c = -1.672, Joint = -1.268, Marginal = -3.874, MI = -13.572\n",
      "Pre-Train Epoch 3, Batch 243: Loss = 157.701, a = 18.450 b = 14.265, c = -1.672, Joint = -0.862, Marginal = -1.998, MI = -13.575\n",
      "Pre-Train Epoch 3, Batch 244: Loss = 44.820, a = 4.595 b = 10.772, c = -1.672, Joint = -2.499, Marginal = -2.618, MI = -12.585\n",
      "Pre-Train Epoch 3, Batch 245: Loss = 61.680, a = 9.848 b = 14.169, c = -0.874, Joint = -3.109, Marginal = -3.419, MI = -6.374\n",
      "Pre-Train Epoch 3, Batch 246: Loss = 65.696, a = 4.450 b = 13.669, c = 0.377, Joint = -1.155, Marginal = -2.114, MI = -19.041\n",
      "Pre-Train Epoch 3, Batch 247: Loss = 51.210, a = 2.596 b = 12.840, c = 0.078, Joint = -1.767, Marginal = -3.016, MI = -19.334\n",
      "Pre-Train Epoch 3, Batch 248: Loss = 114.899, a = 7.480 b = 13.748, c = -1.672, Joint = -5.144, Marginal = -5.986, MI = -7.589\n",
      "Pre-Train Epoch 3, Batch 249: Loss = 70.931, a = 4.344 b = 13.174, c = -0.644, Joint = -2.660, Marginal = -3.023, MI = -4.214\n",
      "Pre-Train Epoch 3, Batch 250: Loss = 163.507, a = 21.157 b = 14.376, c = -0.969, Joint = -3.811, Marginal = -3.747, MI = -16.254\n",
      "Pre-Train Epoch 3, Batch 251: Loss = 35.006, a = 3.176 b = 9.158, c = -1.672, Joint = -2.544, Marginal = -2.007, MI = -8.069\n",
      "Pre-Train Epoch 3, Batch 252: Loss = 176.854, a = 9.461 b = 12.903, c = -1.672, Joint = -2.131, Marginal = -1.604, MI = -25.616\n",
      "Pre-Train Epoch 3, Batch 253: Loss = 63.625, a = 2.948 b = 11.482, c = -1.672, Joint = -1.643, Marginal = -3.711, MI = -6.859\n",
      "Pre-Train Epoch 3, Batch 254: Loss = 53.583, a = 7.533 b = 12.287, c = -1.672, Joint = -2.928, Marginal = -3.750, MI = -3.829\n",
      "Pre-Train Epoch 3, Batch 255: Loss = 315.377, a = 13.886 b = 20.119, c = -0.751, Joint = -1.551, Marginal = -5.090, MI = -21.056\n",
      "Pre-Train Epoch 3, Batch 256: Loss = 69.850, a = 6.430 b = 12.393, c = -1.672, Joint = -2.017, Marginal = -2.601, MI = -5.797\n",
      "Pre-Train Epoch 3, Batch 257: Loss = 101.858, a = 13.554 b = 14.796, c = -1.672, Joint = -2.866, Marginal = -3.450, MI = -10.952\n",
      "Pre-Train Epoch 3, Batch 258: Loss = 56.207, a = 7.566 b = 12.433, c = 0.184, Joint = -2.039, Marginal = -2.912, MI = -8.916\n",
      "Pre-Train Epoch 3, Batch 259: Loss = 78.169, a = 7.274 b = 11.818, c = 0.479, Joint = -2.056, Marginal = -2.435, MI = -28.343\n",
      "Pre-Train Epoch 3, Batch 260: Loss = 88.785, a = 8.861 b = 15.366, c = -1.671, Joint = -3.117, Marginal = -3.204, MI = -7.380\n",
      "Pre-Train Epoch 3, Batch 261: Loss = 71.698, a = 8.914 b = 15.995, c = -1.671, Joint = -2.748, Marginal = -3.481, MI = -9.488\n",
      "Pre-Train Epoch 3, Batch 262: Loss = 63.774, a = 7.592 b = 12.677, c = -1.671, Joint = -1.500, Marginal = -2.441, MI = -6.902\n",
      "Pre-Train Epoch 3, Batch 263: Loss = 151.056, a = 3.304 b = 13.661, c = -1.526, Joint = 0.119, Marginal = -3.161, MI = -13.436\n",
      "Pre-Train Epoch 3, Batch 264: Loss = 85.182, a = 4.992 b = 15.801, c = -0.849, Joint = -1.643, Marginal = -2.520, MI = -17.623\n",
      "Pre-Train Epoch 3, Batch 265: Loss = 117.235, a = 15.179 b = 12.783, c = -0.710, Joint = -3.905, Marginal = -3.150, MI = -27.106\n",
      "Pre-Train Epoch 3, Batch 266: Loss = 45.108, a = 11.525 b = 10.400, c = -1.671, Joint = -2.586, Marginal = -2.213, MI = -10.381\n",
      "Pre-Train Epoch 3, Batch 267: Loss = 95.673, a = 8.539 b = 13.871, c = -1.671, Joint = -1.263, Marginal = -1.746, MI = -34.507\n",
      "Pre-Train Epoch 3, Batch 268: Loss = 73.697, a = 10.074 b = 13.887, c = -1.671, Joint = -1.080, Marginal = -1.718, MI = -11.700\n",
      "Pre-Train Epoch 3, Batch 269: Loss = 161.100, a = 7.531 b = 14.939, c = -0.958, Joint = -1.005, Marginal = -2.329, MI = -20.965\n",
      "Pre-Train Epoch 3, Batch 270: Loss = 118.615, a = 13.168 b = 12.378, c = 0.459, Joint = -1.810, Marginal = -3.055, MI = -14.169\n",
      "Pre-Train Epoch 3, Batch 271: Loss = 222.259, a = 2.016 b = 13.254, c = -1.671, Joint = -0.477, Marginal = -3.333, MI = -15.405\n",
      "Pre-Train Epoch 3, Batch 272: Loss = 90.930, a = 12.571 b = 12.687, c = -1.671, Joint = -0.476, Marginal = -0.857, MI = -18.865\n",
      "Pre-Train Epoch 3, Batch 273: Loss = 102.961, a = 11.296 b = 12.211, c = -1.671, Joint = -1.599, Marginal = -2.623, MI = -9.538\n",
      "Pre-Train Epoch 3, Batch 274: Loss = 109.967, a = 4.758 b = 11.830, c = -1.411, Joint = -0.930, Marginal = -2.377, MI = -35.744\n",
      "Pre-Train Epoch 3, Batch 275: Loss = 48.389, a = 4.358 b = 9.980, c = -1.424, Joint = -3.194, Marginal = -3.181, MI = -3.735\n",
      "Pre-Train Epoch 3, Batch 276: Loss = 68.431, a = 3.487 b = 12.871, c = -1.159, Joint = -2.308, Marginal = -2.766, MI = -5.317\n",
      "Pre-Train Epoch 3, Batch 277: Loss = 316.745, a = 3.534 b = 17.740, c = -0.781, Joint = -1.327, Marginal = -4.105, MI = -36.182\n",
      "Pre-Train Epoch 3, Batch 278: Loss = 124.601, a = 16.166 b = 18.533, c = -1.672, Joint = -3.676, Marginal = -3.919, MI = -12.812\n",
      "Pre-Train Epoch 3, Batch 279: Loss = 75.628, a = 6.583 b = 11.078, c = -1.672, Joint = -3.303, Marginal = -2.905, MI = -6.700\n",
      "Pre-Train Epoch 3, Batch 280: Loss = 50.981, a = 2.747 b = 13.761, c = -1.672, Joint = -1.676, Marginal = -2.490, MI = -17.464\n",
      "Pre-Train Epoch 3, Batch 281: Loss = 105.050, a = 5.592 b = 17.028, c = -0.494, Joint = -3.823, Marginal = -4.606, MI = -4.350\n",
      "Pre-Train Epoch 3, Batch 282: Loss = 281.550, a = 8.343 b = 19.764, c = -0.437, Joint = -0.221, Marginal = -2.198, MI = -10.929\n",
      "Pre-Train Epoch 3, Batch 283: Loss = 78.841, a = 11.409 b = 13.388, c = -0.928, Joint = -3.626, Marginal = -2.946, MI = -6.632\n",
      "Pre-Train Epoch 3, Batch 284: Loss = 80.416, a = 4.113 b = 12.308, c = -1.672, Joint = -0.390, Marginal = -2.067, MI = -9.043\n",
      "Pre-Train Epoch 3, Batch 285: Loss = 75.895, a = 8.439 b = 15.531, c = 0.347, Joint = -2.267, Marginal = -2.609, MI = -23.426\n",
      "Pre-Train Epoch 3, Batch 286: Loss = 96.812, a = 5.800 b = 12.691, c = -0.528, Joint = -0.474, Marginal = -3.706, MI = -3.375\n",
      "Pre-Train Epoch 3, Batch 287: Loss = 465.772, a = 5.302 b = 14.196, c = -1.162, Joint = 2.837, Marginal = -3.035, MI = -0.862\n",
      "Pre-Train Epoch 3, Batch 288: Loss = 114.599, a = 18.441 b = 16.616, c = 0.356, Joint = -3.292, Marginal = -2.424, MI = -18.829\n",
      "Pre-Train Epoch 3, Batch 289: Loss = 67.720, a = 15.582 b = 11.191, c = -0.820, Joint = -2.184, Marginal = -1.368, MI = -11.977\n",
      "Pre-Train Epoch 3, Batch 290: Loss = 46.339, a = 5.303 b = 10.661, c = -1.563, Joint = -2.218, Marginal = -3.226, MI = -9.578\n",
      "Pre-Train Epoch 3, Batch 291: Loss = 202.250, a = 5.173 b = 12.140, c = -1.672, Joint = -1.276, Marginal = -1.992, MI = -74.997\n",
      "Pre-Train Epoch 3, Batch 292: Loss = 112.648, a = 22.074 b = 15.134, c = 0.115, Joint = -4.364, Marginal = -2.644, MI = -11.544\n",
      "Pre-Train Epoch 3, Batch 293: Loss = 63.490, a = 12.868 b = 15.054, c = -0.760, Joint = -3.070, Marginal = -3.536, MI = -3.741\n",
      "Pre-Train Epoch 3, Batch 294: Loss = 62.362, a = 3.746 b = 14.957, c = -0.952, Joint = -1.513, Marginal = -2.557, MI = -12.936\n",
      "Pre-Train Epoch 3, Batch 295: Loss = 46.228, a = 3.978 b = 10.987, c = -0.487, Joint = -1.501, Marginal = -1.614, MI = -19.323\n",
      "Pre-Train Epoch 3, Batch 296: Loss = 93.129, a = 7.134 b = 14.761, c = -0.435, Joint = -1.257, Marginal = -4.073, MI = -15.915\n",
      "Pre-Train Epoch 3, Batch 297: Loss = 123.422, a = 25.179 b = 18.094, c = -0.769, Joint = -3.241, Marginal = -2.863, MI = -27.198\n",
      "Pre-Train Epoch 3, Batch 298: Loss = 74.807, a = 10.287 b = 11.674, c = -1.586, Joint = -1.999, Marginal = -1.878, MI = -9.036\n",
      "Pre-Train Epoch 3, Batch 299: Loss = 151.254, a = 3.177 b = 15.611, c = -1.142, Joint = -1.101, Marginal = -3.547, MI = -18.166\n",
      "Pre-Train Epoch 3, Batch 300: Loss = 51.530, a = 5.326 b = 9.492, c = -1.278, Joint = -1.332, Marginal = -2.141, MI = -4.617\n",
      "Pre-Train Epoch 3, Batch 301: Loss = 78.996, a = 11.757 b = 16.506, c = -1.056, Joint = -2.793, Marginal = -3.690, MI = -8.689\n",
      "Pre-Train Epoch 3, Batch 302: Loss = 244.620, a = 6.651 b = 15.937, c = -1.672, Joint = -0.994, Marginal = -3.494, MI = -44.497\n",
      "Pre-Train Epoch 3, Batch 303: Loss = 105.528, a = 13.573 b = 14.955, c = 0.945, Joint = -1.784, Marginal = -2.734, MI = -6.571\n",
      "Pre-Train Epoch 3, Batch 304: Loss = 100.561, a = 2.480 b = 13.958, c = -1.672, Joint = 0.003, Marginal = -2.240, MI = -14.924\n",
      "Pre-Train Epoch 3, Batch 305: Loss = 65.736, a = 9.211 b = 10.572, c = -1.672, Joint = -1.542, Marginal = -1.738, MI = -9.284\n",
      "Pre-Train Epoch 3, Batch 306: Loss = 105.544, a = 6.002 b = 15.770, c = -0.193, Joint = -3.296, Marginal = -3.358, MI = -6.178\n",
      "\n",
      "Start of epoch 4\n",
      "Pre-Train Epoch 4, Batch 0: Loss = 58.256, a = 6.676 b = 12.931, c = -1.237, Joint = -2.150, Marginal = -3.633, MI = -9.540\n",
      "Pre-Train Epoch 4, Batch 1: Loss = 35.261, a = 2.316 b = 12.685, c = -1.358, Joint = -1.337, Marginal = -3.574, MI = -5.134\n",
      "Pre-Train Epoch 4, Batch 2: Loss = 79.240, a = 9.342 b = 17.332, c = -1.672, Joint = -4.061, Marginal = -3.868, MI = -28.392\n",
      "Pre-Train Epoch 4, Batch 3: Loss = 67.238, a = 7.066 b = 14.394, c = -0.254, Joint = -2.886, Marginal = -3.551, MI = -6.371\n",
      "Pre-Train Epoch 4, Batch 4: Loss = 41.424, a = 4.989 b = 11.543, c = -1.672, Joint = -2.303, Marginal = -3.070, MI = -2.793\n",
      "Pre-Train Epoch 4, Batch 5: Loss = 55.438, a = 7.763 b = 13.365, c = -0.496, Joint = -2.969, Marginal = -2.700, MI = -6.469\n",
      "Pre-Train Epoch 4, Batch 6: Loss = 81.551, a = 5.928 b = 13.901, c = -1.473, Joint = -1.613, Marginal = -2.365, MI = -23.704\n",
      "Pre-Train Epoch 4, Batch 7: Loss = 37.955, a = 3.358 b = 8.879, c = -1.525, Joint = -1.548, Marginal = -1.869, MI = -12.363\n",
      "Pre-Train Epoch 4, Batch 8: Loss = 164.754, a = 8.249 b = 14.231, c = -0.930, Joint = -2.125, Marginal = -1.689, MI = -34.458\n",
      "Pre-Train Epoch 4, Batch 9: Loss = 73.716, a = 8.471 b = 11.747, c = -1.386, Joint = -3.237, Marginal = -2.225, MI = -7.949\n",
      "Pre-Train Epoch 4, Batch 10: Loss = 102.019, a = 4.926 b = 13.782, c = -1.672, Joint = -0.548, Marginal = -1.178, MI = -16.063\n",
      "Pre-Train Epoch 4, Batch 11: Loss = 64.167, a = 7.282 b = 14.188, c = -1.361, Joint = -2.730, Marginal = -3.268, MI = -4.580\n",
      "Pre-Train Epoch 4, Batch 12: Loss = 38.309, a = 2.976 b = 9.903, c = -1.114, Joint = -0.645, Marginal = -0.471, MI = -22.593\n",
      "Pre-Train Epoch 4, Batch 13: Loss = 50.320, a = 2.943 b = 13.795, c = -0.132, Joint = -1.330, Marginal = -2.453, MI = -14.462\n",
      "Pre-Train Epoch 4, Batch 14: Loss = 87.210, a = 7.056 b = 19.829, c = -0.863, Joint = -2.528, Marginal = -4.428, MI = -16.654\n",
      "Pre-Train Epoch 4, Batch 15: Loss = 44.203, a = 7.709 b = 8.871, c = -1.672, Joint = -1.672, Marginal = -2.218, MI = -21.955\n",
      "Pre-Train Epoch 4, Batch 16: Loss = 81.017, a = 7.913 b = 10.661, c = -1.449, Joint = -0.763, Marginal = -2.359, MI = -7.581\n",
      "Pre-Train Epoch 4, Batch 17: Loss = 61.916, a = 4.010 b = 14.451, c = -0.991, Joint = -2.393, Marginal = -2.571, MI = -8.008\n",
      "Pre-Train Epoch 4, Batch 18: Loss = 87.566, a = 12.250 b = 11.658, c = -0.812, Joint = -1.860, Marginal = -1.665, MI = -17.118\n",
      "Pre-Train Epoch 4, Batch 19: Loss = 54.749, a = 11.914 b = 11.349, c = -0.802, Joint = -2.137, Marginal = -1.405, MI = -6.833\n",
      "Pre-Train Epoch 4, Batch 20: Loss = 68.550, a = 8.203 b = 14.420, c = -1.190, Joint = -3.095, Marginal = -2.530, MI = -23.566\n",
      "Pre-Train Epoch 4, Batch 21: Loss = 36.045, a = 3.171 b = 12.365, c = -0.922, Joint = -1.687, Marginal = -2.215, MI = -10.335\n",
      "Pre-Train Epoch 4, Batch 22: Loss = 53.253, a = 4.621 b = 10.827, c = 0.136, Joint = -1.690, Marginal = -2.413, MI = -13.126\n",
      "Pre-Train Epoch 4, Batch 23: Loss = 314.097, a = 7.462 b = 13.535, c = -1.378, Joint = -1.028, Marginal = -2.981, MI = -37.942\n",
      "Pre-Train Epoch 4, Batch 24: Loss = 91.566, a = 4.079 b = 15.552, c = -1.231, Joint = 0.184, Marginal = -3.461, MI = -5.695\n",
      "Pre-Train Epoch 4, Batch 25: Loss = 135.855, a = 7.452 b = 13.926, c = -1.672, Joint = -3.345, Marginal = -2.030, MI = -50.581\n",
      "Pre-Train Epoch 4, Batch 26: Loss = 52.016, a = 2.902 b = 12.188, c = -0.101, Joint = -0.396, Marginal = -2.344, MI = -4.123\n",
      "Pre-Train Epoch 4, Batch 27: Loss = 78.892, a = 8.095 b = 11.278, c = -1.045, Joint = -2.158, Marginal = -1.667, MI = -5.306\n",
      "Pre-Train Epoch 4, Batch 28: Loss = 105.799, a = 16.115 b = 14.179, c = -1.211, Joint = -2.075, Marginal = -1.865, MI = -17.555\n",
      "Pre-Train Epoch 4, Batch 29: Loss = 465.982, a = 6.882 b = 18.396, c = -1.672, Joint = 0.582, Marginal = -3.951, MI = -8.282\n",
      "Pre-Train Epoch 4, Batch 30: Loss = 111.180, a = 19.842 b = 17.752, c = -0.391, Joint = -2.994, Marginal = -4.418, MI = -7.296\n",
      "Pre-Train Epoch 4, Batch 31: Loss = 55.670, a = 5.942 b = 12.059, c = -1.672, Joint = -1.505, Marginal = -1.871, MI = -9.433\n",
      "Pre-Train Epoch 4, Batch 32: Loss = 62.376, a = 9.099 b = 14.103, c = -1.427, Joint = -2.307, Marginal = -2.897, MI = -8.619\n",
      "Pre-Train Epoch 4, Batch 33: Loss = 63.738, a = 6.923 b = 13.389, c = -1.672, Joint = -1.568, Marginal = -1.694, MI = -13.413\n",
      "Pre-Train Epoch 4, Batch 34: Loss = 68.444, a = 12.023 b = 13.146, c = -1.043, Joint = -2.248, Marginal = -3.393, MI = -4.876\n",
      "Pre-Train Epoch 4, Batch 35: Loss = 54.631, a = 6.461 b = 12.935, c = -0.526, Joint = -2.176, Marginal = -3.724, MI = -4.196\n",
      "Pre-Train Epoch 4, Batch 36: Loss = 44.048, a = 4.936 b = 11.941, c = -0.805, Joint = -1.987, Marginal = -2.842, MI = -7.355\n",
      "Pre-Train Epoch 4, Batch 37: Loss = 102.806, a = 8.508 b = 17.036, c = 0.455, Joint = -2.949, Marginal = -3.296, MI = -16.615\n",
      "Pre-Train Epoch 4, Batch 38: Loss = 91.009, a = 6.963 b = 13.947, c = -1.672, Joint = -2.913, Marginal = -3.896, MI = -3.817\n",
      "Pre-Train Epoch 4, Batch 39: Loss = 102.602, a = 6.313 b = 17.755, c = -1.672, Joint = -3.329, Marginal = -3.766, MI = -17.503\n",
      "Pre-Train Epoch 4, Batch 40: Loss = 77.399, a = 3.877 b = 11.751, c = -1.390, Joint = -1.531, Marginal = -1.701, MI = -15.437\n",
      "Pre-Train Epoch 4, Batch 41: Loss = 60.014, a = 6.597 b = 12.828, c = -1.672, Joint = -1.548, Marginal = -1.788, MI = -9.483\n",
      "Pre-Train Epoch 4, Batch 42: Loss = 72.233, a = 9.151 b = 14.096, c = 0.066, Joint = -2.811, Marginal = -2.045, MI = -21.624\n",
      "Pre-Train Epoch 4, Batch 43: Loss = 83.452, a = 4.045 b = 11.743, c = -0.842, Joint = -0.997, Marginal = -2.322, MI = -22.152\n",
      "Pre-Train Epoch 4, Batch 44: Loss = 47.996, a = 8.985 b = 10.429, c = -1.672, Joint = -2.499, Marginal = -2.945, MI = -2.174\n",
      "Pre-Train Epoch 4, Batch 45: Loss = 82.407, a = 6.399 b = 13.820, c = -0.618, Joint = -1.652, Marginal = -2.535, MI = -19.137\n",
      "Pre-Train Epoch 4, Batch 46: Loss = 47.694, a = 3.010 b = 13.281, c = -1.161, Joint = -2.597, Marginal = -2.908, MI = -3.859\n",
      "Pre-Train Epoch 4, Batch 47: Loss = 93.300, a = 5.579 b = 13.876, c = -1.672, Joint = -2.277, Marginal = -2.489, MI = -10.944\n",
      "Pre-Train Epoch 4, Batch 48: Loss = 74.495, a = 3.378 b = 15.046, c = -1.672, Joint = -1.341, Marginal = -3.451, MI = -28.900\n",
      "Pre-Train Epoch 4, Batch 49: Loss = 30.752, a = 2.269 b = 9.742, c = -1.672, Joint = -1.471, Marginal = -1.910, MI = -9.933\n",
      "Pre-Train Epoch 4, Batch 50: Loss = 81.712, a = 5.619 b = 11.713, c = -1.672, Joint = -0.353, Marginal = -0.391, MI = -32.562\n",
      "Pre-Train Epoch 4, Batch 51: Loss = 40.179, a = 2.802 b = 10.363, c = -1.672, Joint = -1.503, Marginal = -2.367, MI = -6.924\n",
      "Pre-Train Epoch 4, Batch 52: Loss = 80.362, a = 6.880 b = 14.100, c = -1.672, Joint = -3.389, Marginal = -2.394, MI = -17.322\n",
      "Pre-Train Epoch 4, Batch 53: Loss = 242.104, a = 6.338 b = 13.995, c = -1.672, Joint = 0.125, Marginal = -3.462, MI = -2.324\n",
      "Pre-Train Epoch 4, Batch 54: Loss = 60.823, a = 7.148 b = 12.628, c = -1.672, Joint = -2.333, Marginal = -2.573, MI = -10.226\n",
      "Pre-Train Epoch 4, Batch 55: Loss = 98.940, a = 3.366 b = 14.913, c = -0.270, Joint = -0.958, Marginal = -3.328, MI = -28.434\n",
      "Pre-Train Epoch 4, Batch 56: Loss = 94.225, a = 7.979 b = 11.779, c = -1.672, Joint = -3.408, Marginal = -2.476, MI = -18.832\n",
      "Pre-Train Epoch 4, Batch 57: Loss = 352.474, a = 4.964 b = 15.508, c = -1.049, Joint = -0.871, Marginal = -3.063, MI = -23.277\n",
      "Pre-Train Epoch 4, Batch 58: Loss = 94.816, a = 9.280 b = 15.852, c = -0.817, Joint = -2.108, Marginal = -3.641, MI = -7.521\n",
      "Pre-Train Epoch 4, Batch 59: Loss = 140.625, a = 6.424 b = 16.327, c = -0.741, Joint = -1.747, Marginal = -3.227, MI = -12.874\n",
      "Pre-Train Epoch 4, Batch 60: Loss = 209.677, a = 3.436 b = 15.438, c = -0.978, Joint = 0.239, Marginal = -0.952, MI = -50.153\n",
      "Pre-Train Epoch 4, Batch 61: Loss = 69.504, a = 6.183 b = 13.923, c = -0.654, Joint = -3.285, Marginal = -2.843, MI = -23.683\n",
      "Pre-Train Epoch 4, Batch 62: Loss = 54.145, a = 3.647 b = 12.635, c = -1.419, Joint = -1.903, Marginal = -3.086, MI = -2.841\n",
      "Pre-Train Epoch 4, Batch 63: Loss = 103.109, a = 10.194 b = 15.147, c = -1.404, Joint = -3.388, Marginal = -3.246, MI = -10.937\n",
      "Pre-Train Epoch 4, Batch 64: Loss = 150.280, a = 3.614 b = 13.487, c = -0.966, Joint = -0.161, Marginal = -1.647, MI = -38.077\n",
      "Pre-Train Epoch 4, Batch 65: Loss = 124.988, a = 8.742 b = 17.278, c = -1.492, Joint = -2.287, Marginal = -2.674, MI = -29.458\n",
      "Pre-Train Epoch 4, Batch 66: Loss = 171.113, a = 3.493 b = 16.492, c = 0.472, Joint = -1.426, Marginal = -3.085, MI = -20.979\n",
      "Pre-Train Epoch 4, Batch 67: Loss = 92.511, a = 17.874 b = 14.873, c = -1.672, Joint = -2.254, Marginal = -2.485, MI = -14.134\n",
      "Pre-Train Epoch 4, Batch 68: Loss = 73.498, a = 6.239 b = 14.311, c = -1.087, Joint = -1.249, Marginal = -1.779, MI = -14.220\n",
      "Pre-Train Epoch 4, Batch 69: Loss = 177.886, a = 6.754 b = 15.476, c = -1.672, Joint = 0.448, Marginal = -2.955, MI = -27.473\n",
      "Pre-Train Epoch 4, Batch 70: Loss = 52.866, a = 4.642 b = 12.079, c = -1.569, Joint = 0.025, Marginal = -0.613, MI = -50.178\n",
      "Pre-Train Epoch 4, Batch 71: Loss = 60.116, a = 10.159 b = 12.148, c = 0.512, Joint = -2.320, Marginal = -1.932, MI = -6.991\n",
      "Pre-Train Epoch 4, Batch 72: Loss = 152.888, a = 6.524 b = 12.928, c = -0.682, Joint = -1.360, Marginal = -1.158, MI = -28.165\n",
      "Pre-Train Epoch 4, Batch 73: Loss = 88.503, a = 5.583 b = 14.052, c = -1.673, Joint = -0.707, Marginal = -1.479, MI = -32.273\n",
      "Pre-Train Epoch 4, Batch 74: Loss = 38.983, a = 4.362 b = 9.950, c = -1.673, Joint = -2.169, Marginal = -2.019, MI = -6.008\n",
      "Pre-Train Epoch 4, Batch 75: Loss = 136.371, a = 5.906 b = 15.193, c = -0.922, Joint = -3.560, Marginal = -2.912, MI = -12.650\n",
      "Pre-Train Epoch 4, Batch 76: Loss = 147.982, a = 8.791 b = 15.734, c = -1.673, Joint = -2.449, Marginal = -3.535, MI = -39.280\n",
      "Pre-Train Epoch 4, Batch 77: Loss = 91.510, a = 12.071 b = 15.292, c = -1.673, Joint = -2.834, Marginal = -3.337, MI = -17.603\n",
      "Pre-Train Epoch 4, Batch 78: Loss = 56.953, a = 5.162 b = 13.184, c = -1.673, Joint = -1.977, Marginal = -2.807, MI = -10.763\n",
      "Pre-Train Epoch 4, Batch 79: Loss = 72.470, a = 8.023 b = 14.686, c = -0.471, Joint = -3.548, Marginal = -3.867, MI = -3.869\n",
      "Pre-Train Epoch 4, Batch 80: Loss = 73.318, a = 7.060 b = 15.437, c = -1.673, Joint = -1.843, Marginal = -3.373, MI = -5.142\n",
      "Pre-Train Epoch 4, Batch 81: Loss = 136.982, a = 6.174 b = 13.457, c = -1.673, Joint = -1.001, Marginal = -2.753, MI = -7.173\n",
      "Pre-Train Epoch 4, Batch 82: Loss = 83.794, a = 6.019 b = 13.574, c = -1.671, Joint = -2.362, Marginal = -3.305, MI = -12.496\n",
      "Pre-Train Epoch 4, Batch 83: Loss = 214.935, a = 8.757 b = 13.331, c = -0.600, Joint = 0.446, Marginal = -2.910, MI = -1.380\n",
      "Pre-Train Epoch 4, Batch 84: Loss = 98.221, a = 1.459 b = 14.403, c = -0.970, Joint = -1.280, Marginal = -2.948, MI = -19.863\n",
      "Pre-Train Epoch 4, Batch 85: Loss = 75.254, a = 5.096 b = 14.806, c = -1.348, Joint = -1.075, Marginal = -2.959, MI = -11.056\n",
      "Pre-Train Epoch 4, Batch 86: Loss = 82.771, a = 5.848 b = 15.264, c = -1.673, Joint = -2.707, Marginal = -2.167, MI = -34.582\n",
      "Pre-Train Epoch 4, Batch 87: Loss = 80.170, a = 6.587 b = 12.837, c = -0.316, Joint = -0.891, Marginal = -3.028, MI = -4.706\n",
      "Pre-Train Epoch 4, Batch 88: Loss = 57.482, a = 2.840 b = 11.623, c = -0.945, Joint = -2.989, Marginal = -3.490, MI = -3.445\n",
      "Pre-Train Epoch 4, Batch 89: Loss = 140.901, a = 7.800 b = 17.362, c = -1.673, Joint = -1.408, Marginal = -4.745, MI = -7.672\n",
      "Pre-Train Epoch 4, Batch 90: Loss = 78.483, a = 4.687 b = 14.697, c = -1.259, Joint = -4.258, Marginal = -3.503, MI = -11.133\n",
      "Pre-Train Epoch 4, Batch 91: Loss = 36.837, a = 2.819 b = 11.209, c = -1.406, Joint = -1.237, Marginal = -2.366, MI = -2.562\n",
      "Pre-Train Epoch 4, Batch 92: Loss = 168.392, a = 3.560 b = 16.169, c = -1.441, Joint = -0.502, Marginal = -2.357, MI = -33.228\n",
      "Pre-Train Epoch 4, Batch 93: Loss = 597.538, a = 8.155 b = 15.908, c = -0.273, Joint = 0.275, Marginal = -1.996, MI = -42.070\n",
      "Pre-Train Epoch 4, Batch 94: Loss = 83.745, a = 7.197 b = 14.706, c = -1.298, Joint = -1.561, Marginal = -3.056, MI = -6.356\n",
      "Pre-Train Epoch 4, Batch 95: Loss = 109.812, a = 4.130 b = 15.449, c = -1.673, Joint = -4.231, Marginal = -2.418, MI = -45.423\n",
      "Pre-Train Epoch 4, Batch 96: Loss = 61.882, a = 2.522 b = 14.184, c = -0.185, Joint = -2.988, Marginal = -3.962, MI = -6.907\n",
      "Pre-Train Epoch 4, Batch 97: Loss = 49.373, a = 2.894 b = 11.672, c = -0.875, Joint = -1.901, Marginal = -2.449, MI = -5.417\n",
      "Pre-Train Epoch 4, Batch 98: Loss = 93.873, a = 1.924 b = 12.715, c = -1.311, Joint = -2.092, Marginal = -3.825, MI = -2.773\n",
      "Pre-Train Epoch 4, Batch 99: Loss = 81.908, a = 3.371 b = 13.838, c = -0.273, Joint = -0.853, Marginal = -2.211, MI = -25.995\n",
      "Pre-Train Epoch 4, Batch 100: Loss = 94.665, a = 2.427 b = 16.309, c = -0.999, Joint = -0.208, Marginal = -1.908, MI = -36.010\n",
      "Pre-Train Epoch 4, Batch 101: Loss = 51.769, a = 5.547 b = 11.198, c = 0.169, Joint = -2.822, Marginal = -2.876, MI = -6.563\n",
      "Pre-Train Epoch 4, Batch 102: Loss = 56.224, a = 4.666 b = 12.043, c = -1.673, Joint = -2.630, Marginal = -4.322, MI = -3.237\n",
      "Pre-Train Epoch 4, Batch 103: Loss = 101.734, a = 5.401 b = 15.561, c = -0.869, Joint = -1.022, Marginal = -3.113, MI = -13.999\n",
      "Pre-Train Epoch 4, Batch 104: Loss = 75.766, a = 9.187 b = 12.333, c = -1.275, Joint = -4.004, Marginal = -3.401, MI = -6.437\n",
      "Pre-Train Epoch 4, Batch 105: Loss = 65.380, a = 5.041 b = 11.700, c = -0.672, Joint = -2.968, Marginal = -3.292, MI = -21.218\n",
      "Pre-Train Epoch 4, Batch 106: Loss = 109.867, a = 10.263 b = 13.855, c = -0.378, Joint = -3.483, Marginal = -3.171, MI = -10.868\n",
      "Pre-Train Epoch 4, Batch 107: Loss = 123.672, a = 4.599 b = 12.473, c = -1.226, Joint = -1.558, Marginal = -3.651, MI = -8.753\n",
      "Pre-Train Epoch 4, Batch 108: Loss = 114.877, a = 8.363 b = 13.148, c = -1.637, Joint = -0.876, Marginal = -0.884, MI = -66.713\n",
      "Pre-Train Epoch 4, Batch 109: Loss = 50.428, a = 10.718 b = 11.502, c = -0.345, Joint = -3.001, Marginal = -2.716, MI = -5.215\n",
      "Pre-Train Epoch 4, Batch 110: Loss = 84.647, a = 6.273 b = 10.959, c = -1.212, Joint = -0.576, Marginal = -1.934, MI = -4.547\n",
      "Pre-Train Epoch 4, Batch 111: Loss = 172.201, a = 24.387 b = 13.435, c = -1.548, Joint = -3.952, Marginal = -3.326, MI = -7.726\n",
      "Pre-Train Epoch 4, Batch 112: Loss = 71.659, a = 5.269 b = 15.465, c = 0.014, Joint = -2.249, Marginal = -2.856, MI = -8.079\n",
      "Pre-Train Epoch 4, Batch 113: Loss = 72.728, a = 13.150 b = 12.690, c = -0.686, Joint = -2.937, Marginal = -2.886, MI = -8.853\n",
      "Pre-Train Epoch 4, Batch 114: Loss = 69.415, a = 5.500 b = 12.882, c = -1.121, Joint = -1.126, Marginal = -2.116, MI = -17.341\n",
      "Pre-Train Epoch 4, Batch 115: Loss = 79.359, a = 9.686 b = 16.345, c = -0.257, Joint = -2.152, Marginal = -0.850, MI = -37.772\n",
      "Pre-Train Epoch 4, Batch 116: Loss = 119.441, a = 6.327 b = 12.613, c = -1.673, Joint = -1.054, Marginal = -3.411, MI = -18.667\n",
      "Pre-Train Epoch 4, Batch 117: Loss = 140.299, a = 10.371 b = 14.823, c = -0.457, Joint = -1.140, Marginal = -3.325, MI = -10.641\n",
      "Pre-Train Epoch 4, Batch 118: Loss = 36.708, a = 6.521 b = 10.371, c = -1.673, Joint = -2.231, Marginal = -1.941, MI = -7.694\n",
      "Pre-Train Epoch 4, Batch 119: Loss = 70.526, a = 9.057 b = 13.320, c = -1.673, Joint = -2.223, Marginal = -1.914, MI = -23.752\n",
      "Pre-Train Epoch 4, Batch 120: Loss = 106.677, a = 10.757 b = 17.466, c = -1.673, Joint = -0.552, Marginal = -3.199, MI = -10.715\n",
      "Pre-Train Epoch 4, Batch 121: Loss = 100.428, a = 17.007 b = 16.597, c = -0.176, Joint = -1.816, Marginal = -3.947, MI = -10.358\n",
      "Pre-Train Epoch 4, Batch 122: Loss = 51.039, a = 6.167 b = 13.056, c = -1.578, Joint = -2.894, Marginal = -3.216, MI = -7.353\n",
      "Pre-Train Epoch 4, Batch 123: Loss = 152.937, a = 8.172 b = 17.201, c = -1.673, Joint = -0.927, Marginal = -4.673, MI = -28.787\n",
      "Pre-Train Epoch 4, Batch 124: Loss = 60.232, a = 12.014 b = 11.554, c = -1.673, Joint = -3.054, Marginal = -2.498, MI = -11.480\n",
      "Pre-Train Epoch 4, Batch 125: Loss = 58.644, a = 3.115 b = 15.322, c = -1.616, Joint = -0.972, Marginal = -3.175, MI = -14.975\n",
      "Pre-Train Epoch 4, Batch 126: Loss = 69.881, a = 8.036 b = 11.491, c = -0.806, Joint = -3.037, Marginal = -1.766, MI = -4.475\n",
      "Pre-Train Epoch 4, Batch 127: Loss = 111.928, a = 6.957 b = 16.211, c = -1.673, Joint = -3.234, Marginal = -4.040, MI = -24.918\n",
      "Pre-Train Epoch 4, Batch 128: Loss = 55.608, a = 4.514 b = 15.776, c = -1.257, Joint = -1.025, Marginal = -4.157, MI = -2.108\n",
      "Pre-Train Epoch 4, Batch 129: Loss = 67.692, a = 9.354 b = 13.731, c = -1.673, Joint = -1.195, Marginal = -2.974, MI = -8.556\n",
      "Pre-Train Epoch 4, Batch 130: Loss = 98.192, a = 11.743 b = 12.833, c = -1.673, Joint = -2.316, Marginal = -2.107, MI = -62.187\n",
      "Pre-Train Epoch 4, Batch 131: Loss = 44.239, a = 6.218 b = 10.255, c = -0.943, Joint = -1.390, Marginal = -1.674, MI = -11.755\n",
      "Pre-Train Epoch 4, Batch 132: Loss = 79.109, a = 6.833 b = 13.008, c = -1.673, Joint = -1.790, Marginal = -1.826, MI = -13.009\n",
      "Pre-Train Epoch 4, Batch 133: Loss = 68.734, a = 11.458 b = 14.631, c = -0.839, Joint = -3.046, Marginal = -3.577, MI = -5.734\n",
      "Pre-Train Epoch 4, Batch 134: Loss = 146.536, a = 6.988 b = 12.057, c = -1.673, Joint = -0.990, Marginal = -0.836, MI = -6.356\n",
      "Pre-Train Epoch 4, Batch 135: Loss = 141.848, a = 18.286 b = 16.355, c = -0.737, Joint = -2.455, Marginal = -3.065, MI = -6.415\n",
      "Pre-Train Epoch 4, Batch 136: Loss = 69.359, a = 8.646 b = 13.771, c = -1.553, Joint = -1.743, Marginal = -2.862, MI = -12.557\n",
      "Pre-Train Epoch 4, Batch 137: Loss = 108.144, a = 11.816 b = 17.206, c = -1.110, Joint = -1.079, Marginal = -3.707, MI = -9.634\n",
      "Pre-Train Epoch 4, Batch 138: Loss = 62.939, a = 12.302 b = 12.833, c = -1.673, Joint = -3.357, Marginal = -3.589, MI = -4.267\n",
      "Pre-Train Epoch 4, Batch 139: Loss = 41.557, a = 7.197 b = 12.636, c = -1.268, Joint = -2.272, Marginal = -2.194, MI = -12.902\n",
      "Pre-Train Epoch 4, Batch 140: Loss = 128.124, a = 23.542 b = 13.470, c = -1.562, Joint = -2.340, Marginal = -2.534, MI = -5.147\n",
      "Pre-Train Epoch 4, Batch 141: Loss = 60.880, a = 8.199 b = 12.989, c = -1.650, Joint = -2.651, Marginal = -2.188, MI = -14.696\n",
      "Pre-Train Epoch 4, Batch 142: Loss = 81.319, a = 9.655 b = 12.162, c = -1.511, Joint = -1.385, Marginal = -2.503, MI = -8.141\n",
      "Pre-Train Epoch 4, Batch 143: Loss = 90.348, a = 8.894 b = 11.892, c = -1.672, Joint = -1.576, Marginal = -2.394, MI = -12.493\n",
      "Pre-Train Epoch 4, Batch 144: Loss = 90.199, a = 7.128 b = 15.854, c = -1.614, Joint = -3.611, Marginal = -3.023, MI = -8.694\n",
      "Pre-Train Epoch 4, Batch 145: Loss = 51.352, a = 5.594 b = 10.248, c = -0.352, Joint = -2.381, Marginal = -1.886, MI = -9.949\n",
      "Pre-Train Epoch 4, Batch 146: Loss = 142.667, a = 7.466 b = 21.217, c = -1.087, Joint = -0.690, Marginal = -3.248, MI = -47.555\n",
      "Pre-Train Epoch 4, Batch 147: Loss = 210.804, a = 8.985 b = 15.551, c = -1.672, Joint = -0.576, Marginal = -1.831, MI = -58.019\n",
      "Pre-Train Epoch 4, Batch 148: Loss = 65.875, a = 7.392 b = 15.343, c = 1.054, Joint = -1.186, Marginal = -2.126, MI = -12.704\n",
      "Pre-Train Epoch 4, Batch 149: Loss = 63.667, a = 9.620 b = 11.573, c = -1.125, Joint = -2.025, Marginal = -1.992, MI = -12.309\n",
      "Pre-Train Epoch 4, Batch 150: Loss = 63.793, a = 8.677 b = 12.076, c = -0.902, Joint = -2.641, Marginal = -2.812, MI = -8.280\n",
      "Pre-Train Epoch 4, Batch 151: Loss = 86.076, a = 6.731 b = 15.276, c = -1.672, Joint = -3.299, Marginal = -4.294, MI = -13.460\n",
      "Pre-Train Epoch 4, Batch 152: Loss = 153.565, a = 19.179 b = 16.256, c = -0.424, Joint = -3.636, Marginal = -3.223, MI = -30.808\n",
      "Pre-Train Epoch 4, Batch 153: Loss = 40.944, a = 2.016 b = 10.257, c = 0.485, Joint = -2.147, Marginal = -2.908, MI = -4.565\n",
      "Pre-Train Epoch 4, Batch 154: Loss = 53.135, a = 3.859 b = 12.884, c = -0.836, Joint = -1.902, Marginal = -2.771, MI = -3.369\n",
      "Pre-Train Epoch 4, Batch 155: Loss = 67.989, a = 9.808 b = 11.183, c = -0.868, Joint = -2.493, Marginal = -2.793, MI = -22.915\n",
      "Pre-Train Epoch 4, Batch 156: Loss = 109.744, a = 13.988 b = 15.558, c = -1.672, Joint = -3.756, Marginal = -2.804, MI = -12.746\n",
      "Pre-Train Epoch 4, Batch 157: Loss = 126.990, a = 8.433 b = 16.721, c = -1.672, Joint = -1.096, Marginal = -3.784, MI = -27.346\n",
      "Pre-Train Epoch 4, Batch 158: Loss = 77.082, a = 6.697 b = 11.997, c = -1.411, Joint = -2.987, Marginal = -3.264, MI = -7.050\n",
      "Pre-Train Epoch 4, Batch 159: Loss = 68.773, a = 6.672 b = 14.069, c = -1.115, Joint = -2.728, Marginal = -3.823, MI = -3.209\n",
      "Pre-Train Epoch 4, Batch 160: Loss = 74.873, a = 7.797 b = 13.886, c = -1.672, Joint = -2.319, Marginal = -4.587, MI = -3.838\n",
      "Pre-Train Epoch 4, Batch 161: Loss = 48.930, a = 6.918 b = 11.764, c = -0.479, Joint = -2.744, Marginal = -2.757, MI = -13.933\n",
      "Pre-Train Epoch 4, Batch 162: Loss = 42.137, a = 7.717 b = 11.250, c = -1.126, Joint = -2.412, Marginal = -1.504, MI = -4.137\n",
      "Pre-Train Epoch 4, Batch 163: Loss = 44.202, a = 4.274 b = 11.322, c = -1.672, Joint = -0.862, Marginal = -2.745, MI = -6.725\n",
      "Pre-Train Epoch 4, Batch 164: Loss = 64.046, a = 3.850 b = 13.710, c = -1.152, Joint = -2.409, Marginal = -2.841, MI = -6.052\n",
      "Pre-Train Epoch 4, Batch 165: Loss = 119.640, a = 11.990 b = 14.207, c = -1.672, Joint = -2.998, Marginal = -2.707, MI = -5.302\n",
      "Pre-Train Epoch 4, Batch 166: Loss = 57.147, a = 13.755 b = 12.771, c = -1.305, Joint = -3.266, Marginal = -2.413, MI = -7.642\n",
      "Pre-Train Epoch 4, Batch 167: Loss = 64.074, a = 10.003 b = 9.052, c = 0.246, Joint = -2.257, Marginal = -1.832, MI = -11.257\n",
      "Pre-Train Epoch 4, Batch 168: Loss = 73.823, a = 3.393 b = 8.710, c = -1.210, Joint = 0.007, Marginal = -1.155, MI = -17.309\n",
      "Pre-Train Epoch 4, Batch 169: Loss = 40.871, a = 4.076 b = 12.872, c = -1.186, Joint = -2.217, Marginal = -2.514, MI = -34.372\n",
      "Pre-Train Epoch 4, Batch 170: Loss = 49.538, a = 7.067 b = 11.192, c = -1.671, Joint = -1.733, Marginal = -2.118, MI = -12.453\n",
      "Pre-Train Epoch 4, Batch 171: Loss = 75.541, a = 7.277 b = 13.008, c = -1.671, Joint = -0.975, Marginal = -2.769, MI = -3.591\n",
      "Pre-Train Epoch 4, Batch 172: Loss = 86.852, a = 11.597 b = 15.066, c = -1.347, Joint = -2.916, Marginal = -1.558, MI = -58.435\n",
      "Pre-Train Epoch 4, Batch 173: Loss = 107.513, a = 11.518 b = 14.719, c = -0.535, Joint = -0.878, Marginal = -2.778, MI = -27.291\n",
      "Pre-Train Epoch 4, Batch 174: Loss = 80.232, a = 4.447 b = 13.188, c = -0.555, Joint = -1.964, Marginal = -3.228, MI = -4.599\n",
      "Pre-Train Epoch 4, Batch 175: Loss = 37.030, a = 5.358 b = 11.886, c = -1.671, Joint = -2.326, Marginal = -3.177, MI = -3.450\n",
      "Pre-Train Epoch 4, Batch 176: Loss = 71.743, a = 8.218 b = 12.918, c = -1.671, Joint = -0.134, Marginal = -3.454, MI = -1.797\n",
      "Pre-Train Epoch 4, Batch 177: Loss = 65.178, a = 6.000 b = 13.793, c = -0.227, Joint = -3.921, Marginal = -2.642, MI = -13.398\n",
      "Pre-Train Epoch 4, Batch 178: Loss = 212.762, a = 17.945 b = 16.465, c = -1.671, Joint = -2.768, Marginal = -3.809, MI = -13.356\n",
      "Pre-Train Epoch 4, Batch 179: Loss = 143.831, a = 4.288 b = 16.533, c = -0.968, Joint = -0.308, Marginal = -4.294, MI = -19.861\n",
      "Pre-Train Epoch 4, Batch 180: Loss = 83.523, a = 5.121 b = 15.992, c = -1.671, Joint = -3.069, Marginal = -4.155, MI = -24.695\n",
      "Pre-Train Epoch 4, Batch 181: Loss = 55.356, a = 7.332 b = 12.399, c = -1.355, Joint = -1.343, Marginal = -2.431, MI = -7.019\n",
      "Pre-Train Epoch 4, Batch 182: Loss = 49.278, a = 7.425 b = 13.110, c = 0.548, Joint = -0.709, Marginal = -3.119, MI = -1.495\n",
      "Pre-Train Epoch 4, Batch 183: Loss = 59.078, a = 11.081 b = 11.004, c = -1.671, Joint = -1.725, Marginal = -1.876, MI = -4.116\n",
      "Pre-Train Epoch 4, Batch 184: Loss = 54.834, a = 5.037 b = 14.749, c = -1.188, Joint = -2.115, Marginal = -2.976, MI = -11.616\n",
      "Pre-Train Epoch 4, Batch 185: Loss = 82.132, a = 9.908 b = 12.161, c = -1.115, Joint = -2.441, Marginal = -1.679, MI = -9.629\n",
      "Pre-Train Epoch 4, Batch 186: Loss = 127.914, a = 5.008 b = 14.375, c = -0.682, Joint = -1.684, Marginal = -2.652, MI = -26.235\n",
      "Pre-Train Epoch 4, Batch 187: Loss = 135.651, a = 4.639 b = 10.774, c = -0.613, Joint = -1.469, Marginal = -2.183, MI = -13.958\n",
      "Pre-Train Epoch 4, Batch 188: Loss = 159.849, a = 3.693 b = 14.591, c = -1.671, Joint = 1.120, Marginal = 0.286, MI = -45.939\n",
      "Pre-Train Epoch 4, Batch 189: Loss = 48.973, a = 6.049 b = 13.751, c = -0.731, Joint = -0.930, Marginal = -2.239, MI = -8.689\n",
      "Pre-Train Epoch 4, Batch 190: Loss = 119.313, a = 19.675 b = 15.075, c = -1.671, Joint = -2.288, Marginal = -2.275, MI = -11.162\n",
      "Pre-Train Epoch 4, Batch 191: Loss = 140.036, a = 17.767 b = 14.088, c = -1.671, Joint = -1.987, Marginal = -1.562, MI = -20.285\n",
      "Pre-Train Epoch 4, Batch 192: Loss = 43.686, a = 6.409 b = 10.428, c = -1.156, Joint = -1.715, Marginal = -2.057, MI = -7.569\n",
      "Pre-Train Epoch 4, Batch 193: Loss = 78.550, a = 11.489 b = 13.485, c = -0.690, Joint = -2.155, Marginal = -1.909, MI = -37.438\n",
      "Pre-Train Epoch 4, Batch 194: Loss = 36.547, a = 3.273 b = 10.488, c = -1.671, Joint = -1.162, Marginal = -1.658, MI = -8.625\n",
      "Pre-Train Epoch 4, Batch 195: Loss = 85.807, a = 14.011 b = 12.431, c = -1.671, Joint = -1.509, Marginal = -2.770, MI = -2.325\n",
      "Pre-Train Epoch 4, Batch 196: Loss = 54.403, a = 8.575 b = 11.615, c = -0.840, Joint = -2.157, Marginal = -2.719, MI = -6.278\n",
      "Pre-Train Epoch 4, Batch 197: Loss = 47.437, a = 8.207 b = 8.942, c = -0.874, Joint = -1.809, Marginal = -2.339, MI = -4.219\n",
      "Pre-Train Epoch 4, Batch 198: Loss = 43.737, a = 3.150 b = 10.300, c = -1.671, Joint = -1.910, Marginal = -2.101, MI = -6.508\n",
      "Pre-Train Epoch 4, Batch 199: Loss = 56.833, a = 11.767 b = 12.321, c = -1.405, Joint = -1.996, Marginal = -2.481, MI = -11.687\n",
      "Pre-Train Epoch 4, Batch 200: Loss = 68.017, a = 7.890 b = 13.529, c = -0.935, Joint = -2.267, Marginal = -2.778, MI = -11.319\n",
      "Pre-Train Epoch 4, Batch 201: Loss = 156.909, a = 6.119 b = 14.938, c = -1.157, Joint = -0.260, Marginal = -1.729, MI = -43.968\n",
      "Pre-Train Epoch 4, Batch 202: Loss = 89.220, a = 3.260 b = 12.260, c = -1.592, Joint = -2.650, Marginal = -3.536, MI = -38.509\n",
      "Pre-Train Epoch 4, Batch 203: Loss = 82.590, a = 7.498 b = 14.565, c = -1.071, Joint = -2.767, Marginal = -2.487, MI = -49.584\n",
      "Pre-Train Epoch 4, Batch 204: Loss = 39.786, a = 4.667 b = 9.396, c = -1.670, Joint = -2.154, Marginal = -1.283, MI = -33.710\n",
      "Pre-Train Epoch 4, Batch 205: Loss = 49.725, a = 6.343 b = 12.178, c = -0.359, Joint = -1.846, Marginal = -2.028, MI = -7.099\n",
      "Pre-Train Epoch 4, Batch 206: Loss = 76.637, a = 8.854 b = 12.041, c = -1.111, Joint = -3.331, Marginal = -2.473, MI = -3.764\n",
      "Pre-Train Epoch 4, Batch 207: Loss = 38.863, a = 6.882 b = 9.936, c = -0.205, Joint = -1.992, Marginal = -2.379, MI = -5.516\n",
      "Pre-Train Epoch 4, Batch 208: Loss = 54.713, a = 6.308 b = 9.985, c = 0.351, Joint = -2.189, Marginal = -2.140, MI = -8.435\n",
      "Pre-Train Epoch 4, Batch 209: Loss = 41.669, a = 5.018 b = 12.066, c = -1.670, Joint = -1.460, Marginal = -1.935, MI = -6.320\n",
      "Pre-Train Epoch 4, Batch 210: Loss = 103.198, a = 2.613 b = 13.656, c = -0.820, Joint = -0.989, Marginal = -1.879, MI = -16.729\n",
      "Pre-Train Epoch 4, Batch 211: Loss = 51.841, a = 6.621 b = 11.810, c = -1.495, Joint = -0.571, Marginal = -1.647, MI = -5.363\n",
      "Pre-Train Epoch 4, Batch 212: Loss = 45.595, a = 4.996 b = 10.566, c = -1.670, Joint = -3.048, Marginal = -2.488, MI = -5.059\n",
      "Pre-Train Epoch 4, Batch 213: Loss = 78.296, a = 6.831 b = 13.341, c = -1.577, Joint = -1.771, Marginal = -2.639, MI = -5.465\n",
      "Pre-Train Epoch 4, Batch 214: Loss = 70.725, a = 5.964 b = 13.575, c = -1.598, Joint = -0.902, Marginal = -1.776, MI = -8.789\n",
      "Pre-Train Epoch 4, Batch 215: Loss = 90.162, a = 12.080 b = 12.944, c = -1.631, Joint = -2.874, Marginal = -2.553, MI = -42.812\n",
      "Pre-Train Epoch 4, Batch 216: Loss = 115.159, a = 5.422 b = 10.789, c = 0.025, Joint = -1.161, Marginal = -3.124, MI = -2.622\n",
      "Pre-Train Epoch 4, Batch 217: Loss = 121.356, a = 8.619 b = 15.067, c = -1.670, Joint = -1.336, Marginal = -2.677, MI = -20.815\n",
      "Pre-Train Epoch 4, Batch 218: Loss = 116.359, a = 4.089 b = 14.002, c = -1.670, Joint = -0.739, Marginal = -3.130, MI = -7.746\n",
      "Pre-Train Epoch 4, Batch 219: Loss = 52.837, a = 4.632 b = 13.815, c = -0.689, Joint = -1.732, Marginal = -2.228, MI = -3.595\n",
      "Pre-Train Epoch 4, Batch 220: Loss = 82.056, a = 6.674 b = 14.647, c = -0.185, Joint = -2.368, Marginal = -3.056, MI = -4.238\n",
      "Pre-Train Epoch 4, Batch 221: Loss = 42.924, a = 5.922 b = 11.290, c = -1.655, Joint = -2.027, Marginal = -2.456, MI = -5.043\n",
      "Pre-Train Epoch 4, Batch 222: Loss = 275.197, a = 6.419 b = 14.876, c = -1.670, Joint = -0.174, Marginal = -3.008, MI = -1.066\n",
      "Pre-Train Epoch 4, Batch 223: Loss = 104.525, a = 9.645 b = 14.000, c = -0.167, Joint = -0.970, Marginal = -3.519, MI = -6.915\n",
      "Pre-Train Epoch 4, Batch 224: Loss = 56.364, a = 13.534 b = 9.274, c = 0.286, Joint = -2.187, Marginal = -2.344, MI = -3.672\n",
      "Pre-Train Epoch 4, Batch 225: Loss = 81.531, a = 9.413 b = 15.828, c = -1.340, Joint = -2.115, Marginal = -2.367, MI = -23.118\n",
      "Pre-Train Epoch 4, Batch 226: Loss = 81.151, a = 10.175 b = 10.026, c = -1.670, Joint = -1.082, Marginal = -2.285, MI = -4.203\n",
      "Pre-Train Epoch 4, Batch 227: Loss = 37.029, a = 4.037 b = 10.445, c = -1.670, Joint = -1.760, Marginal = -2.376, MI = -3.997\n",
      "Pre-Train Epoch 4, Batch 228: Loss = 48.142, a = 2.372 b = 10.513, c = -1.126, Joint = -1.829, Marginal = -2.547, MI = -2.524\n",
      "Pre-Train Epoch 4, Batch 229: Loss = 57.688, a = 6.229 b = 13.052, c = -1.580, Joint = -3.161, Marginal = -3.226, MI = -6.199\n",
      "Pre-Train Epoch 4, Batch 230: Loss = 264.437, a = 4.331 b = 11.803, c = -1.670, Joint = -1.124, Marginal = -2.634, MI = -20.860\n",
      "Pre-Train Epoch 4, Batch 231: Loss = 90.180, a = 4.174 b = 15.449, c = -1.671, Joint = -4.339, Marginal = -4.772, MI = -4.254\n",
      "Pre-Train Epoch 4, Batch 232: Loss = 67.135, a = 4.860 b = 13.974, c = -0.719, Joint = -2.992, Marginal = -3.144, MI = -7.276\n",
      "Pre-Train Epoch 4, Batch 233: Loss = 88.729, a = 5.409 b = 13.515, c = -1.261, Joint = -0.195, Marginal = -1.284, MI = -17.193\n",
      "Pre-Train Epoch 4, Batch 234: Loss = 160.822, a = 9.983 b = 15.268, c = -1.252, Joint = -0.551, Marginal = -2.495, MI = -25.499\n",
      "Pre-Train Epoch 4, Batch 235: Loss = 92.995, a = 11.525 b = 9.971, c = -1.504, Joint = -2.626, Marginal = -1.419, MI = -12.526\n",
      "Pre-Train Epoch 4, Batch 236: Loss = 94.927, a = 5.306 b = 10.145, c = -1.671, Joint = -0.782, Marginal = -2.445, MI = -10.044\n",
      "Pre-Train Epoch 4, Batch 237: Loss = 47.532, a = 3.313 b = 11.730, c = -1.671, Joint = -1.375, Marginal = -2.645, MI = -4.588\n",
      "Pre-Train Epoch 4, Batch 238: Loss = 36.189, a = 5.493 b = 12.596, c = -1.671, Joint = -1.551, Marginal = -2.534, MI = -7.376\n",
      "Pre-Train Epoch 4, Batch 239: Loss = 89.084, a = 12.059 b = 13.541, c = -1.671, Joint = -1.826, Marginal = -3.031, MI = -9.987\n",
      "Pre-Train Epoch 4, Batch 240: Loss = 66.088, a = 5.725 b = 12.620, c = -0.494, Joint = -1.973, Marginal = -3.122, MI = -4.418\n",
      "Pre-Train Epoch 4, Batch 241: Loss = 81.647, a = 9.316 b = 14.154, c = -0.374, Joint = -2.902, Marginal = -3.050, MI = -4.823\n",
      "Pre-Train Epoch 4, Batch 242: Loss = 90.438, a = 5.867 b = 15.229, c = -1.671, Joint = -1.473, Marginal = -3.007, MI = -24.733\n",
      "Pre-Train Epoch 4, Batch 243: Loss = 105.895, a = 15.336 b = 12.827, c = -1.671, Joint = -0.919, Marginal = -2.124, MI = -7.446\n",
      "Pre-Train Epoch 4, Batch 244: Loss = 35.479, a = 4.149 b = 9.699, c = -1.671, Joint = -2.248, Marginal = -2.732, MI = -9.505\n",
      "Pre-Train Epoch 4, Batch 245: Loss = 50.461, a = 8.684 b = 12.788, c = -0.931, Joint = -2.855, Marginal = -2.893, MI = -4.371\n",
      "Pre-Train Epoch 4, Batch 246: Loss = 48.413, a = 4.241 b = 12.457, c = 0.227, Joint = -1.211, Marginal = -1.677, MI = -7.381\n",
      "Pre-Train Epoch 4, Batch 247: Loss = 40.576, a = 3.258 b = 11.707, c = -0.034, Joint = -1.677, Marginal = -2.022, MI = -16.309\n",
      "Pre-Train Epoch 4, Batch 248: Loss = 89.273, a = 7.146 b = 12.287, c = -1.671, Joint = -4.396, Marginal = -4.558, MI = -4.987\n",
      "Pre-Train Epoch 4, Batch 249: Loss = 51.788, a = 4.307 b = 11.935, c = -0.714, Joint = -2.466, Marginal = -2.456, MI = -3.630\n",
      "Pre-Train Epoch 4, Batch 250: Loss = 118.997, a = 19.311 b = 12.867, c = -1.006, Joint = -3.336, Marginal = -2.937, MI = -8.996\n",
      "Pre-Train Epoch 4, Batch 251: Loss = 27.991, a = 3.617 b = 8.395, c = -1.671, Joint = -2.185, Marginal = -1.437, MI = -4.315\n",
      "Pre-Train Epoch 4, Batch 252: Loss = 125.320, a = 7.769 b = 11.646, c = -1.666, Joint = -2.030, Marginal = -2.780, MI = -2.819\n",
      "Pre-Train Epoch 4, Batch 253: Loss = 44.418, a = 3.157 b = 10.436, c = -1.670, Joint = -1.581, Marginal = -2.960, MI = -1.335\n",
      "Pre-Train Epoch 4, Batch 254: Loss = 43.795, a = 7.109 b = 11.100, c = -1.670, Joint = -2.596, Marginal = -2.121, MI = -36.711\n",
      "Pre-Train Epoch 4, Batch 255: Loss = 207.740, a = 12.065 b = 18.017, c = -0.801, Joint = -1.715, Marginal = -2.950, MI = -11.818\n",
      "Pre-Train Epoch 4, Batch 256: Loss = 54.297, a = 5.860 b = 11.247, c = -1.670, Joint = -1.858, Marginal = -1.715, MI = -27.431\n",
      "Pre-Train Epoch 4, Batch 257: Loss = 74.015, a = 11.830 b = 13.274, c = -1.670, Joint = -2.576, Marginal = -2.797, MI = -6.774\n",
      "Pre-Train Epoch 4, Batch 258: Loss = 45.350, a = 6.351 b = 11.258, c = 0.065, Joint = -1.974, Marginal = -1.944, MI = -6.105\n",
      "Pre-Train Epoch 4, Batch 259: Loss = 57.235, a = 6.402 b = 10.658, c = 0.338, Joint = -1.837, Marginal = -2.211, MI = -3.509\n",
      "Pre-Train Epoch 4, Batch 260: Loss = 67.485, a = 8.730 b = 13.976, c = -1.670, Joint = -2.838, Marginal = -1.602, MI = -11.753\n",
      "Pre-Train Epoch 4, Batch 261: Loss = 56.697, a = 7.793 b = 14.405, c = -1.670, Joint = -2.632, Marginal = -3.290, MI = -5.522\n",
      "Pre-Train Epoch 4, Batch 262: Loss = 48.059, a = 6.726 b = 11.503, c = -1.670, Joint = -1.444, Marginal = -1.965, MI = -6.831\n",
      "Pre-Train Epoch 4, Batch 263: Loss = 95.722, a = 3.290 b = 12.402, c = -1.509, Joint = -0.250, Marginal = -1.811, MI = -15.374\n",
      "Pre-Train Epoch 4, Batch 264: Loss = 65.203, a = 5.535 b = 14.409, c = -0.905, Joint = -1.633, Marginal = -2.651, MI = -15.814\n",
      "Pre-Train Epoch 4, Batch 265: Loss = 89.700, a = 13.261 b = 11.476, c = -0.743, Joint = -3.431, Marginal = -3.184, MI = -7.636\n",
      "Pre-Train Epoch 4, Batch 266: Loss = 36.516, a = 10.712 b = 9.412, c = -1.670, Joint = -2.260, Marginal = -1.785, MI = -4.522\n",
      "Pre-Train Epoch 4, Batch 267: Loss = 69.066, a = 7.362 b = 12.528, c = -1.670, Joint = -1.302, Marginal = -2.905, MI = -15.597\n",
      "Pre-Train Epoch 4, Batch 268: Loss = 54.974, a = 8.813 b = 12.583, c = -1.670, Joint = -1.190, Marginal = -2.358, MI = -5.932\n",
      "Pre-Train Epoch 4, Batch 269: Loss = 109.641, a = 7.033 b = 13.518, c = -0.987, Joint = -1.120, Marginal = -2.484, MI = -12.512\n",
      "Pre-Train Epoch 4, Batch 270: Loss = 82.314, a = 12.260 b = 11.154, c = 0.357, Joint = -1.679, Marginal = -2.651, MI = -4.207\n",
      "Pre-Train Epoch 4, Batch 271: Loss = 135.604, a = 2.399 b = 11.994, c = -1.670, Joint = -0.757, Marginal = -3.420, MI = -0.742\n",
      "Pre-Train Epoch 4, Batch 272: Loss = 66.578, a = 10.893 b = 11.513, c = -1.670, Joint = -0.675, Marginal = -1.839, MI = -2.388\n",
      "Pre-Train Epoch 4, Batch 273: Loss = 74.700, a = 9.586 b = 11.028, c = -1.670, Joint = -1.516, Marginal = -2.128, MI = -24.206\n",
      "Pre-Train Epoch 4, Batch 274: Loss = 79.201, a = 4.323 b = 10.779, c = -1.440, Joint = -1.056, Marginal = -2.634, MI = -8.855\n",
      "Pre-Train Epoch 4, Batch 275: Loss = 39.458, a = 3.963 b = 9.039, c = -1.442, Joint = -2.897, Marginal = -2.250, MI = -4.435\n",
      "Pre-Train Epoch 4, Batch 276: Loss = 51.045, a = 3.258 b = 11.688, c = -1.193, Joint = -2.238, Marginal = -1.871, MI = -7.957\n",
      "Pre-Train Epoch 4, Batch 277: Loss = 217.638, a = 3.910 b = 16.051, c = -0.834, Joint = -1.477, Marginal = -3.757, MI = -8.607\n",
      "Pre-Train Epoch 4, Batch 278: Loss = 96.463, a = 14.429 b = 16.733, c = -1.670, Joint = -3.420, Marginal = -2.856, MI = -10.721\n",
      "Pre-Train Epoch 4, Batch 279: Loss = 59.241, a = 6.217 b = 10.028, c = -1.670, Joint = -2.934, Marginal = -2.109, MI = -3.042\n",
      "Pre-Train Epoch 4, Batch 280: Loss = 38.184, a = 2.687 b = 12.463, c = -1.670, Joint = -1.776, Marginal = -3.146, MI = -2.427\n",
      "Pre-Train Epoch 4, Batch 281: Loss = 79.787, a = 4.955 b = 15.300, c = -0.566, Joint = -3.618, Marginal = -3.815, MI = -4.107\n",
      "Pre-Train Epoch 4, Batch 282: Loss = 185.674, a = 7.206 b = 17.791, c = -0.492, Joint = -0.625, Marginal = -4.333, MI = -3.921\n",
      "Pre-Train Epoch 4, Batch 283: Loss = 62.348, a = 9.423 b = 12.052, c = -0.970, Joint = -3.259, Marginal = -2.781, MI = -7.832\n",
      "Pre-Train Epoch 4, Batch 284: Loss = 57.566, a = 3.948 b = 11.256, c = -1.670, Joint = -0.566, Marginal = -1.605, MI = -10.266\n",
      "Pre-Train Epoch 4, Batch 285: Loss = 59.903, a = 8.041 b = 14.090, c = 0.244, Joint = -2.131, Marginal = -2.132, MI = -10.837\n",
      "Pre-Train Epoch 4, Batch 286: Loss = 63.129, a = 5.010 b = 11.502, c = -0.587, Joint = -0.706, Marginal = -1.816, MI = -4.885\n",
      "Pre-Train Epoch 4, Batch 287: Loss = 326.198, a = 4.873 b = 12.963, c = -1.206, Joint = 2.201, Marginal = 0.331, MI = -71.509\n",
      "Pre-Train Epoch 4, Batch 288: Loss = 88.931, a = 16.154 b = 14.997, c = 0.221, Joint = -3.031, Marginal = -1.586, MI = -25.752\n",
      "Pre-Train Epoch 4, Batch 289: Loss = 51.931, a = 13.371 b = 10.136, c = -0.861, Joint = -1.930, Marginal = -1.216, MI = -8.498\n",
      "Pre-Train Epoch 4, Batch 290: Loss = 37.554, a = 4.827 b = 9.633, c = -1.569, Joint = -2.005, Marginal = -2.255, MI = -3.277\n",
      "Pre-Train Epoch 4, Batch 291: Loss = 147.289, a = 4.900 b = 11.022, c = -1.670, Joint = -1.225, Marginal = -2.338, MI = -10.391\n",
      "Pre-Train Epoch 4, Batch 292: Loss = 86.200, a = 18.889 b = 13.620, c = -0.022, Joint = -3.887, Marginal = -2.466, MI = -7.576\n",
      "Pre-Train Epoch 4, Batch 293: Loss = 50.176, a = 11.207 b = 13.568, c = -0.809, Joint = -2.783, Marginal = -2.790, MI = -9.078\n",
      "Pre-Train Epoch 4, Batch 294: Loss = 48.676, a = 3.572 b = 13.597, c = -0.976, Joint = -1.601, Marginal = -2.679, MI = -7.534\n",
      "Pre-Train Epoch 4, Batch 295: Loss = 36.234, a = 3.567 b = 9.994, c = -0.556, Joint = -1.439, Marginal = -1.887, MI = -16.352\n",
      "Pre-Train Epoch 4, Batch 296: Loss = 67.590, a = 6.523 b = 13.352, c = -0.502, Joint = -1.261, Marginal = -2.854, MI = -37.599\n",
      "Pre-Train Epoch 4, Batch 297: Loss = 94.339, a = 21.179 b = 16.218, c = -0.816, Joint = -3.002, Marginal = -3.332, MI = -8.804\n",
      "Pre-Train Epoch 4, Batch 298: Loss = 56.714, a = 8.856 b = 10.545, c = -1.578, Joint = -1.837, Marginal = -2.209, MI = -2.944\n",
      "Pre-Train Epoch 4, Batch 299: Loss = 109.054, a = 3.492 b = 14.179, c = -1.173, Joint = -1.185, Marginal = -2.091, MI = -34.412\n",
      "Pre-Train Epoch 4, Batch 300: Loss = 39.603, a = 4.578 b = 8.619, c = -1.285, Joint = -1.240, Marginal = -1.699, MI = -2.597\n",
      "Pre-Train Epoch 4, Batch 301: Loss = 63.281, a = 9.711 b = 14.881, c = -1.082, Joint = -2.712, Marginal = -3.208, MI = -5.115\n",
      "Pre-Train Epoch 4, Batch 302: Loss = 165.738, a = 5.693 b = 14.421, c = -1.670, Joint = -1.156, Marginal = -1.413, MI = -27.151\n",
      "Pre-Train Epoch 4, Batch 303: Loss = 78.474, a = 11.772 b = 13.532, c = 0.782, Joint = -1.757, Marginal = -1.715, MI = -17.849\n",
      "Pre-Train Epoch 4, Batch 304: Loss = 71.365, a = 2.654 b = 12.752, c = -1.670, Joint = -0.296, Marginal = -1.902, MI = -17.390\n",
      "Pre-Train Epoch 4, Batch 305: Loss = 50.609, a = 7.862 b = 9.660, c = -1.670, Joint = -1.476, Marginal = -1.614, MI = -7.168\n",
      "Pre-Train Epoch 4, Batch 306: Loss = 79.091, a = 5.650 b = 14.282, c = -0.293, Joint = -3.045, Marginal = -3.275, MI = -5.450\n",
      "Epoch 1/1000\n",
      "20/20 - 0s - loss: 37.9102 - MI: -3.7414e+01\n",
      "Epoch 2/1000\n",
      "20/20 - 0s - loss: 22.4759 - MI: -2.2253e+01\n",
      "Epoch 3/1000\n",
      "20/20 - 0s - loss: 18.9999 - MI: -1.8786e+01\n",
      "Epoch 4/1000\n",
      "20/20 - 0s - loss: 16.1240 - MI: -1.5931e+01\n",
      "Epoch 5/1000\n",
      "20/20 - 0s - loss: 13.9555 - MI: -1.3988e+01\n",
      "Epoch 6/1000\n",
      "20/20 - 0s - loss: 11.1669 - MI: -1.1115e+01\n",
      "Epoch 7/1000\n",
      "20/20 - 0s - loss: 9.6667 - MI: -9.6322e+00\n",
      "Epoch 8/1000\n",
      "20/20 - 0s - loss: 9.0968 - MI: -9.0337e+00\n",
      "Epoch 9/1000\n",
      "20/20 - 0s - loss: 9.0133 - MI: -8.9351e+00\n",
      "Epoch 10/1000\n",
      "20/20 - 0s - loss: 8.0528 - MI: -7.9896e+00\n",
      "Epoch 11/1000\n",
      "20/20 - 0s - loss: 7.1158 - MI: -7.0785e+00\n",
      "Epoch 12/1000\n",
      "20/20 - 0s - loss: 6.9683 - MI: -6.9124e+00\n",
      "Epoch 13/1000\n",
      "20/20 - 0s - loss: 5.9018 - MI: -5.8662e+00\n",
      "Epoch 14/1000\n",
      "20/20 - 0s - loss: 5.5345 - MI: -5.4968e+00\n",
      "Epoch 15/1000\n",
      "20/20 - 0s - loss: 5.6221 - MI: -5.5894e+00\n",
      "Epoch 16/1000\n",
      "20/20 - 0s - loss: 5.3165 - MI: -5.2848e+00\n",
      "Epoch 17/1000\n",
      "20/20 - 0s - loss: 4.6888 - MI: -4.6640e+00\n",
      "Epoch 18/1000\n",
      "20/20 - 0s - loss: 3.7075 - MI: -3.6919e+00\n",
      "Epoch 19/1000\n",
      "20/20 - 0s - loss: 3.9133 - MI: -3.9992e+00\n",
      "Epoch 20/1000\n",
      "20/20 - 0s - loss: 3.5976 - MI: -3.5808e+00\n",
      "Epoch 21/1000\n",
      "20/20 - 0s - loss: 3.0867 - MI: -3.0791e+00\n",
      "Epoch 22/1000\n",
      "20/20 - 0s - loss: 3.3618 - MI: -3.3365e+00\n",
      "Epoch 23/1000\n",
      "20/20 - 0s - loss: 3.1016 - MI: -3.0868e+00\n",
      "Epoch 24/1000\n",
      "20/20 - 0s - loss: 2.4655 - MI: -2.4600e+00\n",
      "Epoch 25/1000\n",
      "20/20 - 0s - loss: 2.3517 - MI: -2.3467e+00\n",
      "Epoch 26/1000\n",
      "20/20 - 0s - loss: 2.8123 - MI: -2.7963e+00\n",
      "Epoch 27/1000\n",
      "20/20 - 0s - loss: 2.2042 - MI: -2.1999e+00\n",
      "Epoch 28/1000\n",
      "20/20 - 0s - loss: 2.0430 - MI: -2.0372e+00\n",
      "Epoch 29/1000\n",
      "20/20 - 0s - loss: 1.9469 - MI: -1.9468e+00\n",
      "Epoch 30/1000\n",
      "20/20 - 0s - loss: 1.8469 - MI: -1.8470e+00\n",
      "Epoch 31/1000\n",
      "20/20 - 0s - loss: 2.1103 - MI: -2.1038e+00\n",
      "Epoch 32/1000\n",
      "20/20 - 0s - loss: 1.8293 - MI: -1.8263e+00\n",
      "Epoch 33/1000\n",
      "20/20 - 0s - loss: 2.1745 - MI: -2.1634e+00\n",
      "Epoch 34/1000\n",
      "20/20 - 0s - loss: 2.4662 - MI: -2.4458e+00\n",
      "Epoch 35/1000\n",
      "20/20 - 0s - loss: 1.8071 - MI: -1.8030e+00\n",
      "Epoch 36/1000\n",
      "20/20 - 0s - loss: 1.6574 - MI: -1.6586e+00\n",
      "Epoch 37/1000\n",
      "20/20 - 0s - loss: 1.6160 - MI: -1.6159e+00\n",
      "Epoch 38/1000\n",
      "20/20 - 0s - loss: 1.7708 - MI: -1.7612e+00\n",
      "Epoch 39/1000\n",
      "20/20 - 0s - loss: 1.6670 - MI: -1.6735e+00\n",
      "Epoch 40/1000\n",
      "20/20 - 0s - loss: 1.7799 - MI: -1.7738e+00\n",
      "Epoch 41/1000\n",
      "20/20 - 0s - loss: 1.4620 - MI: -1.4647e+00\n",
      "Epoch 42/1000\n",
      "20/20 - 0s - loss: 1.4287 - MI: -1.4255e+00\n",
      "Epoch 43/1000\n",
      "20/20 - 0s - loss: 1.5427 - MI: -1.5392e+00\n",
      "Epoch 44/1000\n",
      "20/20 - 0s - loss: 1.4473 - MI: -1.4491e+00\n",
      "Epoch 45/1000\n",
      "20/20 - 0s - loss: 1.3452 - MI: -1.3449e+00\n",
      "Epoch 46/1000\n",
      "20/20 - 0s - loss: 1.2781 - MI: -1.2742e+00\n",
      "Epoch 47/1000\n",
      "20/20 - 0s - loss: 1.2722 - MI: -1.2704e+00\n",
      "Epoch 48/1000\n",
      "20/20 - 0s - loss: 1.7394 - MI: -1.7311e+00\n",
      "Epoch 49/1000\n",
      "20/20 - 0s - loss: 1.2519 - MI: -1.2480e+00\n",
      "Epoch 50/1000\n",
      "20/20 - 0s - loss: 1.1739 - MI: -1.1731e+00\n",
      "Epoch 51/1000\n",
      "20/20 - 0s - loss: 1.1391 - MI: -1.1364e+00\n",
      "Epoch 52/1000\n",
      "20/20 - 0s - loss: 1.1159 - MI: -1.1154e+00\n",
      "Epoch 53/1000\n",
      "20/20 - 0s - loss: 1.1539 - MI: -1.1536e+00\n",
      "Epoch 54/1000\n",
      "20/20 - 0s - loss: 1.0470 - MI: -1.0469e+00\n",
      "Epoch 55/1000\n",
      "20/20 - 0s - loss: 1.0905 - MI: -1.0869e+00\n",
      "Epoch 56/1000\n",
      "20/20 - 0s - loss: 0.9961 - MI: -9.9715e-01\n",
      "Epoch 57/1000\n",
      "20/20 - 0s - loss: 1.0156 - MI: -1.0146e+00\n",
      "Epoch 58/1000\n",
      "20/20 - 0s - loss: 1.3071 - MI: -1.3030e+00\n",
      "Epoch 59/1000\n",
      "20/20 - 0s - loss: 1.0376 - MI: -1.0345e+00\n",
      "Epoch 60/1000\n",
      "20/20 - 0s - loss: 1.0419 - MI: -1.0399e+00\n",
      "Epoch 61/1000\n",
      "20/20 - 0s - loss: 0.9270 - MI: -9.2459e-01\n",
      "Epoch 62/1000\n",
      "20/20 - 0s - loss: 0.8719 - MI: -8.7002e-01\n",
      "Epoch 63/1000\n",
      "20/20 - 0s - loss: 0.8505 - MI: -8.4944e-01\n",
      "Epoch 64/1000\n",
      "20/20 - 0s - loss: 0.9887 - MI: -9.8601e-01\n",
      "Epoch 65/1000\n",
      "20/20 - 0s - loss: 0.8107 - MI: -8.1295e-01\n",
      "Epoch 66/1000\n",
      "20/20 - 0s - loss: 0.8633 - MI: -8.6154e-01\n",
      "Epoch 67/1000\n",
      "20/20 - 0s - loss: 0.8037 - MI: -8.0188e-01\n",
      "Epoch 68/1000\n",
      "20/20 - 0s - loss: 0.8322 - MI: -8.3115e-01\n",
      "Epoch 69/1000\n",
      "20/20 - 0s - loss: 0.7821 - MI: -7.7992e-01\n",
      "Epoch 70/1000\n",
      "20/20 - 0s - loss: 0.7197 - MI: -7.1749e-01\n",
      "Epoch 71/1000\n",
      "20/20 - 0s - loss: 0.6886 - MI: -6.8847e-01\n",
      "Epoch 72/1000\n",
      "20/20 - 0s - loss: 0.8051 - MI: -8.0050e-01\n",
      "Epoch 73/1000\n",
      "20/20 - 0s - loss: 0.7153 - MI: -7.2334e-01\n",
      "Epoch 74/1000\n",
      "20/20 - 0s - loss: 0.6890 - MI: -6.8875e-01\n",
      "Epoch 75/1000\n",
      "20/20 - 0s - loss: 0.6649 - MI: -6.6250e-01\n",
      "Epoch 76/1000\n",
      "20/20 - 0s - loss: 0.6367 - MI: -6.3607e-01\n",
      "Epoch 77/1000\n",
      "20/20 - 0s - loss: 0.6015 - MI: -6.0073e-01\n",
      "Epoch 78/1000\n",
      "20/20 - 0s - loss: 0.6040 - MI: -6.0346e-01\n",
      "Epoch 79/1000\n",
      "20/20 - 0s - loss: 0.5853 - MI: -5.8457e-01\n",
      "Epoch 80/1000\n",
      "20/20 - 0s - loss: 0.7496 - MI: -7.4695e-01\n",
      "Epoch 81/1000\n",
      "20/20 - 0s - loss: 0.6441 - MI: -6.4131e-01\n",
      "Epoch 82/1000\n",
      "20/20 - 0s - loss: 0.6067 - MI: -6.0489e-01\n",
      "Epoch 83/1000\n",
      "20/20 - 0s - loss: 0.5661 - MI: -5.6496e-01\n",
      "Epoch 84/1000\n",
      "20/20 - 0s - loss: 0.5305 - MI: -5.2946e-01\n",
      "Epoch 85/1000\n",
      "20/20 - 0s - loss: 0.5337 - MI: -5.3426e-01\n",
      "Epoch 86/1000\n",
      "20/20 - 0s - loss: 0.5716 - MI: -5.6967e-01\n",
      "Epoch 87/1000\n",
      "20/20 - 0s - loss: 0.5282 - MI: -5.2698e-01\n",
      "Epoch 88/1000\n",
      "20/20 - 0s - loss: 0.4717 - MI: -4.7109e-01\n",
      "Epoch 89/1000\n",
      "20/20 - 0s - loss: 0.4652 - MI: -4.6565e-01\n",
      "Epoch 90/1000\n",
      "20/20 - 0s - loss: 0.4665 - MI: -4.6531e-01\n",
      "Epoch 91/1000\n",
      "20/20 - 0s - loss: 0.4444 - MI: -4.4352e-01\n",
      "Epoch 92/1000\n",
      "20/20 - 0s - loss: 0.4489 - MI: -4.4805e-01\n",
      "Epoch 93/1000\n",
      "20/20 - 0s - loss: 0.4215 - MI: -4.2102e-01\n",
      "Epoch 94/1000\n",
      "20/20 - 0s - loss: 0.5125 - MI: -5.1169e-01\n",
      "Epoch 95/1000\n",
      "20/20 - 0s - loss: 0.4273 - MI: -4.2726e-01\n",
      "Epoch 96/1000\n",
      "20/20 - 0s - loss: 0.4040 - MI: -4.0337e-01\n",
      "Epoch 97/1000\n",
      "20/20 - 0s - loss: 0.3805 - MI: -3.8007e-01\n",
      "Epoch 98/1000\n",
      "20/20 - 0s - loss: 0.3778 - MI: -3.7721e-01\n",
      "Epoch 99/1000\n",
      "20/20 - 0s - loss: 0.3566 - MI: -3.5666e-01\n",
      "Epoch 100/1000\n",
      "20/20 - 0s - loss: 0.3640 - MI: -3.6319e-01\n",
      "Epoch 101/1000\n",
      "20/20 - 0s - loss: 0.3571 - MI: -3.6093e-01\n",
      "Epoch 102/1000\n",
      "20/20 - 0s - loss: 0.3434 - MI: -3.4258e-01\n",
      "Epoch 103/1000\n",
      "20/20 - 0s - loss: 0.3345 - MI: -3.3402e-01\n",
      "Epoch 104/1000\n",
      "20/20 - 0s - loss: 0.3325 - MI: -3.3280e-01\n",
      "Epoch 105/1000\n",
      "20/20 - 0s - loss: 0.3305 - MI: -3.3009e-01\n",
      "Epoch 106/1000\n",
      "20/20 - 0s - loss: 0.3181 - MI: -3.2477e-01\n",
      "Epoch 107/1000\n",
      "20/20 - 0s - loss: 0.3085 - MI: -3.0786e-01\n",
      "Epoch 108/1000\n",
      "20/20 - 0s - loss: 0.3249 - MI: -3.2398e-01\n",
      "Epoch 109/1000\n",
      "20/20 - 0s - loss: 0.3197 - MI: -3.1899e-01\n",
      "Epoch 110/1000\n",
      "20/20 - 0s - loss: 0.2947 - MI: -2.9504e-01\n",
      "Epoch 111/1000\n",
      "20/20 - 0s - loss: 0.2752 - MI: -2.7482e-01\n",
      "Epoch 112/1000\n",
      "20/20 - 0s - loss: 0.2743 - MI: -2.7881e-01\n",
      "Epoch 113/1000\n",
      "20/20 - 0s - loss: 0.2899 - MI: -2.8880e-01\n",
      "Epoch 114/1000\n",
      "20/20 - 0s - loss: 0.2804 - MI: -2.7968e-01\n",
      "Epoch 115/1000\n",
      "20/20 - 0s - loss: 0.2493 - MI: -2.4895e-01\n",
      "Epoch 116/1000\n",
      "20/20 - 0s - loss: 0.2583 - MI: -2.5756e-01\n",
      "Epoch 117/1000\n",
      "20/20 - 0s - loss: 0.2821 - MI: -2.8220e-01\n",
      "Epoch 118/1000\n",
      "20/20 - 0s - loss: 0.2486 - MI: -2.5189e-01\n",
      "Epoch 119/1000\n",
      "20/20 - 0s - loss: 0.2430 - MI: -2.4255e-01\n",
      "Epoch 120/1000\n",
      "20/20 - 0s - loss: 0.2217 - MI: -2.2144e-01\n",
      "Epoch 121/1000\n",
      "20/20 - 0s - loss: 0.2339 - MI: -2.3348e-01\n",
      "Epoch 122/1000\n",
      "20/20 - 0s - loss: 0.2197 - MI: -2.2234e-01\n",
      "Epoch 123/1000\n",
      "20/20 - 0s - loss: 0.2124 - MI: -2.1397e-01\n",
      "Epoch 124/1000\n",
      "20/20 - 0s - loss: 0.2136 - MI: -2.1325e-01\n",
      "Epoch 125/1000\n",
      "20/20 - 0s - loss: 0.2164 - MI: -2.1578e-01\n",
      "Epoch 126/1000\n",
      "20/20 - 0s - loss: 0.2170 - MI: -2.1630e-01\n",
      "Epoch 127/1000\n",
      "20/20 - 0s - loss: 0.1957 - MI: -1.9542e-01\n",
      "Epoch 128/1000\n",
      "20/20 - 0s - loss: 0.1791 - MI: -1.7873e-01\n",
      "Epoch 129/1000\n",
      "20/20 - 0s - loss: 0.1860 - MI: -1.8570e-01\n",
      "Epoch 130/1000\n",
      "20/20 - 0s - loss: 0.1819 - MI: -1.8175e-01\n",
      "Epoch 131/1000\n",
      "20/20 - 0s - loss: 0.1799 - MI: -1.8008e-01\n",
      "Epoch 132/1000\n",
      "20/20 - 0s - loss: 0.1868 - MI: -1.8590e-01\n",
      "Epoch 133/1000\n",
      "20/20 - 0s - loss: 0.1692 - MI: -1.6970e-01\n",
      "Epoch 134/1000\n",
      "20/20 - 0s - loss: 0.1814 - MI: -1.8053e-01\n",
      "Epoch 135/1000\n",
      "20/20 - 0s - loss: 0.1644 - MI: -1.6426e-01\n",
      "Epoch 136/1000\n",
      "20/20 - 0s - loss: 0.1606 - MI: -1.6054e-01\n",
      "Epoch 137/1000\n",
      "20/20 - 0s - loss: 0.1559 - MI: -1.5502e-01\n",
      "Epoch 138/1000\n",
      "20/20 - 0s - loss: 0.1564 - MI: -1.5603e-01\n",
      "Epoch 139/1000\n",
      "20/20 - 0s - loss: 0.1446 - MI: -1.4389e-01\n",
      "Epoch 140/1000\n",
      "20/20 - 0s - loss: 0.1387 - MI: -1.3775e-01\n",
      "Epoch 141/1000\n",
      "20/20 - 0s - loss: 0.1390 - MI: -1.3855e-01\n",
      "Epoch 142/1000\n",
      "20/20 - 0s - loss: 0.1363 - MI: -1.3642e-01\n",
      "Epoch 143/1000\n",
      "20/20 - 0s - loss: 0.1265 - MI: -1.2630e-01\n",
      "Epoch 144/1000\n",
      "20/20 - 0s - loss: 0.1154 - MI: -1.1520e-01\n",
      "Epoch 145/1000\n",
      "20/20 - 0s - loss: 0.1209 - MI: -1.1994e-01\n",
      "Epoch 146/1000\n",
      "20/20 - 0s - loss: 0.1138 - MI: -1.1411e-01\n",
      "Epoch 147/1000\n",
      "20/20 - 0s - loss: 0.1091 - MI: -1.0878e-01\n",
      "Epoch 148/1000\n",
      "20/20 - 0s - loss: 0.1053 - MI: -1.0442e-01\n",
      "Epoch 149/1000\n",
      "20/20 - 0s - loss: 0.1173 - MI: -1.1593e-01\n",
      "Epoch 150/1000\n",
      "20/20 - 0s - loss: 0.1057 - MI: -1.0438e-01\n",
      "Epoch 151/1000\n",
      "20/20 - 0s - loss: 0.1050 - MI: -1.1040e-01\n",
      "Epoch 152/1000\n",
      "20/20 - 0s - loss: 0.1007 - MI: -1.0061e-01\n",
      "Epoch 153/1000\n",
      "20/20 - 0s - loss: 0.0994 - MI: -9.9288e-02\n",
      "Epoch 154/1000\n",
      "20/20 - 0s - loss: 0.0838 - MI: -8.3898e-02\n",
      "Epoch 155/1000\n",
      "20/20 - 0s - loss: 0.0916 - MI: -9.0574e-02\n",
      "Epoch 156/1000\n",
      "20/20 - 0s - loss: 0.0823 - MI: -8.3031e-02\n",
      "Epoch 157/1000\n",
      "20/20 - 0s - loss: 0.0864 - MI: -8.5888e-02\n",
      "Epoch 158/1000\n",
      "20/20 - 0s - loss: 0.0764 - MI: -7.5828e-02\n",
      "Epoch 159/1000\n",
      "20/20 - 0s - loss: 0.0769 - MI: -7.6235e-02\n",
      "Epoch 160/1000\n",
      "20/20 - 0s - loss: 0.0686 - MI: -6.8934e-02\n",
      "Epoch 161/1000\n",
      "20/20 - 0s - loss: 0.0631 - MI: -6.2487e-02\n",
      "Epoch 162/1000\n",
      "20/20 - 0s - loss: 0.0726 - MI: -7.1641e-02\n",
      "Epoch 163/1000\n",
      "20/20 - 0s - loss: 0.0655 - MI: -6.4932e-02\n",
      "Epoch 164/1000\n",
      "20/20 - 0s - loss: 0.0632 - MI: -6.2165e-02\n",
      "Epoch 165/1000\n",
      "20/20 - 0s - loss: 0.0613 - MI: -6.0923e-02\n",
      "Epoch 166/1000\n",
      "20/20 - 0s - loss: 0.0565 - MI: -5.6913e-02\n",
      "Epoch 167/1000\n",
      "20/20 - 0s - loss: 0.0594 - MI: -5.9354e-02\n",
      "Epoch 168/1000\n",
      "20/20 - 0s - loss: 0.0573 - MI: -5.6138e-02\n",
      "Epoch 169/1000\n",
      "20/20 - 0s - loss: 0.0493 - MI: -4.9306e-02\n",
      "Epoch 170/1000\n",
      "20/20 - 0s - loss: 0.0513 - MI: -5.0799e-02\n",
      "Epoch 171/1000\n",
      "20/20 - 0s - loss: 0.0464 - MI: -5.1772e-02\n",
      "Epoch 172/1000\n",
      "20/20 - 0s - loss: 0.0379 - MI: -3.7276e-02\n",
      "Epoch 173/1000\n",
      "20/20 - 0s - loss: 0.0414 - MI: -4.0543e-02\n",
      "Epoch 174/1000\n",
      "20/20 - 0s - loss: 0.0407 - MI: -4.0211e-02\n",
      "Epoch 175/1000\n",
      "20/20 - 0s - loss: 0.0412 - MI: -4.0616e-02\n",
      "Epoch 176/1000\n",
      "20/20 - 0s - loss: 0.0361 - MI: -3.9515e-02\n",
      "Epoch 177/1000\n",
      "20/20 - 0s - loss: 0.0288 - MI: -2.9917e-02\n",
      "Epoch 178/1000\n",
      "20/20 - 0s - loss: 0.0237 - MI: -2.3875e-02\n",
      "Epoch 179/1000\n",
      "20/20 - 0s - loss: 0.0198 - MI: -1.9818e-02\n",
      "Epoch 180/1000\n",
      "20/20 - 0s - loss: 0.0269 - MI: -2.7082e-02\n",
      "Epoch 181/1000\n",
      "20/20 - 0s - loss: 0.0208 - MI: -2.0262e-02\n",
      "Epoch 182/1000\n",
      "20/20 - 0s - loss: 0.0171 - MI: -1.7968e-02\n",
      "Epoch 183/1000\n",
      "20/20 - 0s - loss: 0.0184 - MI: -1.8579e-02\n",
      "Epoch 184/1000\n",
      "20/20 - 0s - loss: 0.0213 - MI: -2.0881e-02\n",
      "Epoch 185/1000\n",
      "20/20 - 0s - loss: 0.0209 - MI: -2.0916e-02\n",
      "Epoch 186/1000\n",
      "20/20 - 0s - loss: 0.0155 - MI: -1.5315e-02\n",
      "Epoch 187/1000\n",
      "20/20 - 0s - loss: 0.0133 - MI: -1.2660e-02\n",
      "Epoch 188/1000\n",
      "20/20 - 0s - loss: 0.0147 - MI: -1.5347e-02\n",
      "Epoch 189/1000\n",
      "20/20 - 0s - loss: 0.0065 - MI: -9.2480e-03\n",
      "Epoch 190/1000\n",
      "20/20 - 0s - loss: 0.0118 - MI: -1.2180e-02\n",
      "Epoch 191/1000\n",
      "20/20 - 0s - loss: 0.0069 - MI: -5.6784e-03\n",
      "Epoch 192/1000\n",
      "20/20 - 0s - loss: 0.0026 - MI: -2.8429e-03\n",
      "Epoch 193/1000\n",
      "20/20 - 0s - loss: -2.2432e-03 - MI: 0.0026\n",
      "Epoch 194/1000\n",
      "20/20 - 0s - loss: -2.5241e-03 - MI: 0.0019\n",
      "Epoch 195/1000\n",
      "20/20 - 0s - loss: -2.0771e-03 - MI: 0.0022\n",
      "Epoch 196/1000\n",
      "20/20 - 0s - loss: 4.1817e-04 - MI: -9.5958e-04\n",
      "Epoch 197/1000\n",
      "20/20 - 0s - loss: -6.3833e-03 - MI: 0.0051\n",
      "Epoch 198/1000\n",
      "20/20 - 0s - loss: -5.0722e-03 - MI: 0.0047\n",
      "Epoch 199/1000\n",
      "20/20 - 0s - loss: -3.4052e-03 - MI: 0.0032\n",
      "Epoch 200/1000\n",
      "20/20 - 0s - loss: -6.5127e-03 - MI: 0.0072\n",
      "Epoch 201/1000\n",
      "20/20 - 0s - loss: -9.0580e-03 - MI: 0.0100\n",
      "Epoch 202/1000\n",
      "20/20 - 0s - loss: -1.5842e-02 - MI: 0.0150\n",
      "Epoch 203/1000\n",
      "20/20 - 0s - loss: -1.6754e-02 - MI: 0.0165\n",
      "Epoch 204/1000\n",
      "20/20 - 0s - loss: -1.4530e-02 - MI: 0.0137\n",
      "Epoch 205/1000\n",
      "20/20 - 0s - loss: -8.4428e-03 - MI: 0.0090\n",
      "Epoch 206/1000\n",
      "20/20 - 0s - loss: -1.6721e-02 - MI: 0.0173\n",
      "Epoch 207/1000\n",
      "20/20 - 0s - loss: -1.5461e-02 - MI: 0.0151\n",
      "Epoch 208/1000\n",
      "20/20 - 0s - loss: -1.2003e-02 - MI: 0.0118\n",
      "Epoch 209/1000\n",
      "20/20 - 0s - loss: -1.9423e-02 - MI: 0.0194\n",
      "Epoch 210/1000\n",
      "20/20 - 0s - loss: -1.6478e-02 - MI: 0.0168\n",
      "Epoch 211/1000\n",
      "20/20 - 0s - loss: -2.5634e-02 - MI: 0.0270\n",
      "Epoch 212/1000\n",
      "20/20 - 0s - loss: -1.6407e-02 - MI: 0.0182\n",
      "Epoch 213/1000\n",
      "20/20 - 0s - loss: -2.2367e-02 - MI: 0.0230\n",
      "Epoch 214/1000\n",
      "20/20 - 0s - loss: -3.0069e-02 - MI: 0.0290\n",
      "Epoch 215/1000\n",
      "20/20 - 0s - loss: -2.4555e-02 - MI: 0.0249\n",
      "Epoch 216/1000\n",
      "20/20 - 0s - loss: -2.8862e-02 - MI: 0.0288\n",
      "Epoch 217/1000\n",
      "20/20 - 0s - loss: -2.2838e-02 - MI: 0.0228\n",
      "Epoch 218/1000\n",
      "20/20 - 0s - loss: -2.1387e-02 - MI: 0.0218\n",
      "Epoch 219/1000\n",
      "20/20 - 0s - loss: -2.6352e-02 - MI: 0.0272\n",
      "Epoch 220/1000\n",
      "20/20 - 0s - loss: -2.5563e-02 - MI: 0.0272\n",
      "Epoch 221/1000\n",
      "20/20 - 0s - loss: -3.0787e-02 - MI: 0.0304\n",
      "Epoch 222/1000\n",
      "20/20 - 0s - loss: -3.0975e-02 - MI: 0.0315\n",
      "Epoch 223/1000\n",
      "20/20 - 0s - loss: -3.1675e-02 - MI: 0.0305\n",
      "Epoch 224/1000\n",
      "20/20 - 0s - loss: -3.1995e-02 - MI: 0.0316\n",
      "Epoch 225/1000\n",
      "20/20 - 0s - loss: -3.0389e-02 - MI: 0.0305\n",
      "Epoch 226/1000\n",
      "20/20 - 0s - loss: -3.3057e-02 - MI: 0.0317\n",
      "Epoch 227/1000\n",
      "20/20 - 0s - loss: -2.7079e-02 - MI: 0.0281\n",
      "Epoch 228/1000\n",
      "20/20 - 0s - loss: -3.1003e-02 - MI: 0.0318\n",
      "Epoch 229/1000\n",
      "20/20 - 0s - loss: -3.5375e-02 - MI: 0.0358\n",
      "Epoch 230/1000\n",
      "20/20 - 0s - loss: -3.3752e-02 - MI: 0.0338\n",
      "Epoch 231/1000\n",
      "20/20 - 0s - loss: -3.0071e-02 - MI: 0.0299\n",
      "Epoch 232/1000\n",
      "20/20 - 0s - loss: -3.5118e-02 - MI: 0.0357\n",
      "Epoch 233/1000\n",
      "20/20 - 0s - loss: -3.3431e-02 - MI: 0.0341\n",
      "Epoch 234/1000\n",
      "20/20 - 0s - loss: -3.3996e-02 - MI: 0.0336\n",
      "Epoch 235/1000\n",
      "20/20 - 0s - loss: -3.4698e-02 - MI: 0.0352\n",
      "Epoch 236/1000\n",
      "20/20 - 0s - loss: -3.7887e-02 - MI: 0.0370\n",
      "Epoch 237/1000\n",
      "20/20 - 0s - loss: -3.5281e-02 - MI: 0.0347\n",
      "Epoch 238/1000\n",
      "20/20 - 0s - loss: -4.2813e-02 - MI: 0.0428\n",
      "Epoch 239/1000\n",
      "20/20 - 0s - loss: -3.6849e-02 - MI: 0.0366\n",
      "Epoch 240/1000\n",
      "20/20 - 0s - loss: -3.7656e-02 - MI: 0.0369\n",
      "Epoch 241/1000\n",
      "20/20 - 0s - loss: -4.2946e-02 - MI: 0.0432\n",
      "Epoch 242/1000\n",
      "20/20 - 0s - loss: -4.0589e-02 - MI: 0.0417\n",
      "Epoch 243/1000\n",
      "20/20 - 0s - loss: -4.6905e-02 - MI: 0.0472\n",
      "Epoch 244/1000\n",
      "20/20 - 0s - loss: -3.8086e-02 - MI: 0.0372\n",
      "Epoch 245/1000\n",
      "20/20 - 0s - loss: -3.0646e-02 - MI: 0.0332\n",
      "Epoch 246/1000\n",
      "20/20 - 0s - loss: -3.8599e-02 - MI: 0.0384\n",
      "Epoch 247/1000\n",
      "20/20 - 0s - loss: -4.7467e-02 - MI: 0.0487\n",
      "Epoch 248/1000\n",
      "20/20 - 0s - loss: -4.3603e-02 - MI: 0.0433\n",
      "Epoch 249/1000\n",
      "20/20 - 0s - loss: -4.4686e-02 - MI: 0.0455\n",
      "Epoch 250/1000\n",
      "20/20 - 0s - loss: -3.4876e-02 - MI: 0.0344\n",
      "Epoch 251/1000\n",
      "20/20 - 0s - loss: -4.2314e-02 - MI: 0.0424\n",
      "Epoch 252/1000\n",
      "20/20 - 0s - loss: -4.0288e-02 - MI: 0.0379\n",
      "Epoch 253/1000\n",
      "20/20 - 0s - loss: -3.8286e-02 - MI: 0.0381\n",
      "Epoch 254/1000\n",
      "20/20 - 0s - loss: -4.1770e-02 - MI: 0.0405\n",
      "Epoch 255/1000\n",
      "20/20 - 0s - loss: -4.5469e-02 - MI: 0.0457\n",
      "Epoch 256/1000\n",
      "20/20 - 0s - loss: -4.7610e-02 - MI: 0.0491\n",
      "Epoch 257/1000\n",
      "20/20 - 0s - loss: -4.2882e-02 - MI: 0.0429\n",
      "Epoch 258/1000\n",
      "20/20 - 0s - loss: -4.5942e-02 - MI: 0.0466\n",
      "Epoch 259/1000\n",
      "20/20 - 0s - loss: -4.8091e-02 - MI: 0.0485\n",
      "Epoch 260/1000\n",
      "20/20 - 0s - loss: -4.8739e-02 - MI: 0.0483\n",
      "Epoch 261/1000\n",
      "20/20 - 0s - loss: -4.6437e-02 - MI: 0.0464\n",
      "Epoch 262/1000\n",
      "20/20 - 0s - loss: -4.9883e-02 - MI: 0.0509\n",
      "Epoch 263/1000\n",
      "20/20 - 0s - loss: -4.4520e-02 - MI: 0.0446\n",
      "Epoch 264/1000\n",
      "20/20 - 0s - loss: -4.8534e-02 - MI: 0.0487\n",
      "Epoch 265/1000\n",
      "20/20 - 0s - loss: -4.4967e-02 - MI: 0.0452\n",
      "Epoch 266/1000\n",
      "20/20 - 0s - loss: -4.8342e-02 - MI: 0.0480\n",
      "Epoch 267/1000\n",
      "20/20 - 0s - loss: -4.6408e-02 - MI: 0.0472\n",
      "Epoch 268/1000\n",
      "20/20 - 0s - loss: -5.4384e-02 - MI: 0.0547\n",
      "Epoch 269/1000\n",
      "20/20 - 0s - loss: -5.1246e-02 - MI: 0.0508\n",
      "Epoch 270/1000\n",
      "20/20 - 0s - loss: -4.9512e-02 - MI: 0.0502\n",
      "Epoch 271/1000\n",
      "20/20 - 0s - loss: -4.7310e-02 - MI: 0.0470\n",
      "Epoch 272/1000\n",
      "20/20 - 0s - loss: -5.1403e-02 - MI: 0.0525\n",
      "Epoch 273/1000\n",
      "20/20 - 0s - loss: -5.1539e-02 - MI: 0.0523\n",
      "Epoch 274/1000\n",
      "20/20 - 0s - loss: -4.9923e-02 - MI: 0.0507\n",
      "Epoch 275/1000\n",
      "20/20 - 0s - loss: -4.8972e-02 - MI: 0.0492\n",
      "Epoch 276/1000\n",
      "20/20 - 0s - loss: -4.6714e-02 - MI: 0.0471\n",
      "Epoch 277/1000\n",
      "20/20 - 0s - loss: -4.9371e-02 - MI: 0.0495\n",
      "Epoch 278/1000\n",
      "20/20 - 0s - loss: -5.8677e-02 - MI: 0.0589\n",
      "Epoch 279/1000\n",
      "20/20 - 0s - loss: -4.8688e-02 - MI: 0.0486\n",
      "Epoch 280/1000\n",
      "20/20 - 0s - loss: -5.2792e-02 - MI: 0.0539\n",
      "Epoch 281/1000\n",
      "20/20 - 0s - loss: -4.6928e-02 - MI: 0.0478\n",
      "Epoch 282/1000\n",
      "20/20 - 0s - loss: -4.6594e-02 - MI: 0.0465\n",
      "Epoch 283/1000\n",
      "20/20 - 0s - loss: -4.6322e-02 - MI: 0.0475\n",
      "Epoch 284/1000\n",
      "20/20 - 0s - loss: -5.3603e-02 - MI: 0.0543\n",
      "Epoch 285/1000\n",
      "20/20 - 0s - loss: -5.3347e-02 - MI: 0.0538\n",
      "Epoch 286/1000\n",
      "20/20 - 0s - loss: -4.7688e-02 - MI: 0.0482\n",
      "Epoch 287/1000\n",
      "20/20 - 0s - loss: -5.4692e-02 - MI: 0.0559\n",
      "Epoch 288/1000\n",
      "20/20 - 0s - loss: -5.4359e-02 - MI: 0.0551\n",
      "Epoch 289/1000\n",
      "20/20 - 0s - loss: -5.8307e-02 - MI: 0.0571\n",
      "Epoch 290/1000\n",
      "20/20 - 0s - loss: -5.3786e-02 - MI: 0.0533\n",
      "Epoch 291/1000\n",
      "20/20 - 0s - loss: -4.8911e-02 - MI: 0.0494\n",
      "Epoch 292/1000\n",
      "20/20 - 0s - loss: -6.3299e-02 - MI: 0.0632\n",
      "Epoch 293/1000\n",
      "20/20 - 0s - loss: -5.1245e-02 - MI: 0.0517\n",
      "Epoch 294/1000\n",
      "20/20 - 0s - loss: -4.9778e-02 - MI: 0.0502\n",
      "Epoch 295/1000\n",
      "20/20 - 0s - loss: -5.5719e-02 - MI: 0.0554\n",
      "Epoch 296/1000\n",
      "20/20 - 0s - loss: -5.0711e-02 - MI: 0.0506\n",
      "Epoch 297/1000\n",
      "20/20 - 0s - loss: -5.2460e-02 - MI: 0.0517\n",
      "Epoch 298/1000\n",
      "20/20 - 0s - loss: -4.9145e-02 - MI: 0.0495\n",
      "Epoch 299/1000\n",
      "20/20 - 0s - loss: -5.4662e-02 - MI: 0.0548\n",
      "Epoch 300/1000\n",
      "20/20 - 0s - loss: -5.7464e-02 - MI: 0.0581\n",
      "Epoch 301/1000\n",
      "20/20 - 0s - loss: -4.8342e-02 - MI: 0.0488\n",
      "Epoch 302/1000\n",
      "20/20 - 0s - loss: -5.5368e-02 - MI: 0.0551\n",
      "Epoch 303/1000\n",
      "20/20 - 0s - loss: -5.2700e-02 - MI: 0.0525\n",
      "Epoch 304/1000\n",
      "20/20 - 0s - loss: -5.8807e-02 - MI: 0.0592\n",
      "Epoch 305/1000\n",
      "20/20 - 0s - loss: -5.2811e-02 - MI: 0.0532\n",
      "Epoch 306/1000\n",
      "20/20 - 0s - loss: -6.2144e-02 - MI: 0.0622\n",
      "Epoch 307/1000\n",
      "20/20 - 0s - loss: -5.5171e-02 - MI: 0.0542\n",
      "Epoch 308/1000\n",
      "20/20 - 0s - loss: -5.9897e-02 - MI: 0.0601\n",
      "Epoch 309/1000\n",
      "20/20 - 0s - loss: -6.3751e-02 - MI: 0.0636\n",
      "Epoch 310/1000\n",
      "20/20 - 0s - loss: -5.1321e-02 - MI: 0.0512\n",
      "Epoch 311/1000\n",
      "20/20 - 0s - loss: -5.8920e-02 - MI: 0.0586\n",
      "Epoch 312/1000\n",
      "20/20 - 0s - loss: -5.8220e-02 - MI: 0.0575\n",
      "Epoch 313/1000\n",
      "20/20 - 0s - loss: -6.1505e-02 - MI: 0.0609\n",
      "Epoch 314/1000\n",
      "20/20 - 0s - loss: -5.8570e-02 - MI: 0.0577\n",
      "Epoch 315/1000\n",
      "20/20 - 0s - loss: -5.9694e-02 - MI: 0.0596\n",
      "Epoch 316/1000\n",
      "20/20 - 0s - loss: -5.7232e-02 - MI: 0.0589\n",
      "Epoch 317/1000\n",
      "20/20 - 0s - loss: -5.6349e-02 - MI: 0.0565\n",
      "Epoch 318/1000\n",
      "20/20 - 0s - loss: -6.5779e-02 - MI: 0.0655\n",
      "Epoch 319/1000\n",
      "20/20 - 0s - loss: -5.9072e-02 - MI: 0.0591\n",
      "Epoch 320/1000\n",
      "20/20 - 0s - loss: -6.2697e-02 - MI: 0.0634\n",
      "Epoch 321/1000\n",
      "20/20 - 0s - loss: -5.9859e-02 - MI: 0.0590\n",
      "Epoch 322/1000\n",
      "20/20 - 0s - loss: -6.0962e-02 - MI: 0.0605\n",
      "Epoch 323/1000\n",
      "20/20 - 0s - loss: -5.8201e-02 - MI: 0.0575\n",
      "Epoch 324/1000\n",
      "20/20 - 0s - loss: -6.2237e-02 - MI: 0.0626\n",
      "Epoch 325/1000\n",
      "20/20 - 0s - loss: -6.0680e-02 - MI: 0.0590\n",
      "Epoch 326/1000\n",
      "20/20 - 0s - loss: -6.2919e-02 - MI: 0.0632\n",
      "Epoch 327/1000\n",
      "20/20 - 0s - loss: -6.6453e-02 - MI: 0.0664\n",
      "Epoch 328/1000\n",
      "20/20 - 0s - loss: -6.1737e-02 - MI: 0.0623\n",
      "Epoch 329/1000\n",
      "20/20 - 0s - loss: -6.4467e-02 - MI: 0.0643\n",
      "Epoch 330/1000\n",
      "20/20 - 0s - loss: -6.2756e-02 - MI: 0.0640\n",
      "Epoch 331/1000\n",
      "20/20 - 0s - loss: -6.3969e-02 - MI: 0.0641\n",
      "Epoch 332/1000\n",
      "20/20 - 0s - loss: -6.2954e-02 - MI: 0.0639\n",
      "Epoch 333/1000\n",
      "20/20 - 0s - loss: -5.3263e-02 - MI: 0.0530\n",
      "Epoch 334/1000\n",
      "20/20 - 0s - loss: -5.9019e-02 - MI: 0.0593\n",
      "Epoch 335/1000\n",
      "20/20 - 0s - loss: -6.4440e-02 - MI: 0.0656\n",
      "Epoch 336/1000\n",
      "20/20 - 0s - loss: -5.8530e-02 - MI: 0.0600\n",
      "Epoch 337/1000\n",
      "20/20 - 0s - loss: -6.0835e-02 - MI: 0.0606\n",
      "Epoch 338/1000\n",
      "20/20 - 0s - loss: -5.9188e-02 - MI: 0.0584\n",
      "Epoch 339/1000\n",
      "20/20 - 0s - loss: -5.9819e-02 - MI: 0.0597\n",
      "Epoch 340/1000\n",
      "20/20 - 0s - loss: -5.7725e-02 - MI: 0.0577\n",
      "Epoch 341/1000\n",
      "20/20 - 0s - loss: -6.2983e-02 - MI: 0.0625\n",
      "Epoch 342/1000\n",
      "20/20 - 0s - loss: -6.0090e-02 - MI: 0.0596\n",
      "Epoch 343/1000\n",
      "20/20 - 0s - loss: -7.0005e-02 - MI: 0.0693\n",
      "Epoch 344/1000\n",
      "20/20 - 0s - loss: -6.8586e-02 - MI: 0.0695\n",
      "Epoch 345/1000\n",
      "20/20 - 0s - loss: -6.7224e-02 - MI: 0.0673\n",
      "Epoch 346/1000\n",
      "20/20 - 0s - loss: -6.8213e-02 - MI: 0.0681\n",
      "Epoch 347/1000\n",
      "20/20 - 0s - loss: -6.1543e-02 - MI: 0.0622\n",
      "Epoch 348/1000\n",
      "20/20 - 0s - loss: -6.2310e-02 - MI: 0.0624\n",
      "Epoch 349/1000\n",
      "20/20 - 0s - loss: -6.1043e-02 - MI: 0.0616\n",
      "Epoch 350/1000\n",
      "20/20 - 0s - loss: -6.3326e-02 - MI: 0.0635\n",
      "Epoch 351/1000\n",
      "20/20 - 0s - loss: -6.4565e-02 - MI: 0.0652\n",
      "Epoch 352/1000\n",
      "20/20 - 0s - loss: -6.2197e-02 - MI: 0.0627\n",
      "Epoch 353/1000\n",
      "20/20 - 0s - loss: -6.4340e-02 - MI: 0.0644\n",
      "Epoch 354/1000\n",
      "20/20 - 0s - loss: -6.5184e-02 - MI: 0.0645\n",
      "Epoch 355/1000\n",
      "20/20 - 0s - loss: -6.7643e-02 - MI: 0.0673\n",
      "Epoch 356/1000\n",
      "20/20 - 0s - loss: -6.8384e-02 - MI: 0.0685\n",
      "Epoch 357/1000\n",
      "20/20 - 0s - loss: -6.4504e-02 - MI: 0.0633\n",
      "Epoch 358/1000\n",
      "20/20 - 0s - loss: -6.1750e-02 - MI: 0.0625\n",
      "Epoch 359/1000\n",
      "20/20 - 0s - loss: -6.9102e-02 - MI: 0.0698\n",
      "Epoch 360/1000\n",
      "20/20 - 0s - loss: -6.7968e-02 - MI: 0.0679\n",
      "Epoch 361/1000\n",
      "20/20 - 0s - loss: -5.9181e-02 - MI: 0.0595\n",
      "Epoch 362/1000\n",
      "20/20 - 0s - loss: -7.6044e-02 - MI: 0.0759\n",
      "Epoch 363/1000\n",
      "20/20 - 0s - loss: -6.2549e-02 - MI: 0.0620\n",
      "Epoch 364/1000\n",
      "20/20 - 0s - loss: -6.3636e-02 - MI: 0.0627\n",
      "Epoch 365/1000\n",
      "20/20 - 0s - loss: -7.4109e-02 - MI: 0.0754\n",
      "Epoch 366/1000\n",
      "20/20 - 0s - loss: -6.9648e-02 - MI: 0.0681\n",
      "Epoch 367/1000\n",
      "20/20 - 0s - loss: -6.7350e-02 - MI: 0.0686\n",
      "Epoch 368/1000\n",
      "20/20 - 0s - loss: -7.3165e-02 - MI: 0.0722\n",
      "Epoch 369/1000\n",
      "20/20 - 0s - loss: -7.3351e-02 - MI: 0.0729\n",
      "Epoch 370/1000\n",
      "20/20 - 0s - loss: -6.8987e-02 - MI: 0.0681\n",
      "Epoch 371/1000\n",
      "20/20 - 0s - loss: -6.9318e-02 - MI: 0.0676\n",
      "Epoch 372/1000\n",
      "20/20 - 0s - loss: -7.2776e-02 - MI: 0.0736\n",
      "Epoch 373/1000\n",
      "20/20 - 0s - loss: -7.3103e-02 - MI: 0.0728\n",
      "Epoch 374/1000\n",
      "20/20 - 0s - loss: -5.9030e-02 - MI: 0.0586\n",
      "Epoch 375/1000\n",
      "20/20 - 0s - loss: -6.7138e-02 - MI: 0.0670\n",
      "Epoch 376/1000\n",
      "20/20 - 0s - loss: -7.0426e-02 - MI: 0.0702\n",
      "Epoch 377/1000\n",
      "20/20 - 0s - loss: -6.9935e-02 - MI: 0.0708\n",
      "Epoch 378/1000\n",
      "20/20 - 0s - loss: -7.0837e-02 - MI: 0.0709\n",
      "Epoch 379/1000\n",
      "20/20 - 0s - loss: -7.5735e-02 - MI: 0.0757\n",
      "Epoch 380/1000\n",
      "20/20 - 0s - loss: -7.2277e-02 - MI: 0.0721\n",
      "Epoch 381/1000\n",
      "20/20 - 0s - loss: -7.1713e-02 - MI: 0.0716\n",
      "Epoch 382/1000\n",
      "20/20 - 0s - loss: -6.7061e-02 - MI: 0.0657\n",
      "Epoch 383/1000\n",
      "20/20 - 0s - loss: -7.3721e-02 - MI: 0.0734\n",
      "Epoch 384/1000\n",
      "20/20 - 0s - loss: -6.7608e-02 - MI: 0.0685\n",
      "Epoch 385/1000\n",
      "20/20 - 0s - loss: -7.2424e-02 - MI: 0.0712\n",
      "Epoch 386/1000\n",
      "20/20 - 0s - loss: -7.5986e-02 - MI: 0.0757\n",
      "Epoch 387/1000\n",
      "20/20 - 0s - loss: -7.4918e-02 - MI: 0.0757\n",
      "Epoch 388/1000\n",
      "20/20 - 0s - loss: -8.1129e-02 - MI: 0.0816\n",
      "Epoch 389/1000\n",
      "20/20 - 0s - loss: -7.6228e-02 - MI: 0.0757\n",
      "Epoch 390/1000\n",
      "20/20 - 0s - loss: -7.0345e-02 - MI: 0.0707\n",
      "Epoch 391/1000\n",
      "20/20 - 0s - loss: -7.9029e-02 - MI: 0.0778\n",
      "Epoch 392/1000\n",
      "20/20 - 0s - loss: -7.7388e-02 - MI: 0.0787\n",
      "Epoch 393/1000\n",
      "20/20 - 0s - loss: -7.7733e-02 - MI: 0.0797\n",
      "Epoch 394/1000\n",
      "20/20 - 0s - loss: -7.7791e-02 - MI: 0.0786\n",
      "Epoch 395/1000\n",
      "20/20 - 0s - loss: -8.1799e-02 - MI: 0.0817\n",
      "Epoch 396/1000\n",
      "20/20 - 0s - loss: -7.7661e-02 - MI: 0.0782\n",
      "Epoch 397/1000\n",
      "20/20 - 0s - loss: -7.1963e-02 - MI: 0.0713\n",
      "Epoch 398/1000\n",
      "20/20 - 0s - loss: -7.8033e-02 - MI: 0.0777\n",
      "Epoch 399/1000\n",
      "20/20 - 0s - loss: -7.7839e-02 - MI: 0.0775\n",
      "Epoch 400/1000\n",
      "20/20 - 0s - loss: -8.0379e-02 - MI: 0.0801\n",
      "Epoch 401/1000\n",
      "20/20 - 0s - loss: -7.9777e-02 - MI: 0.0803\n",
      "Epoch 402/1000\n",
      "20/20 - 0s - loss: -8.0995e-02 - MI: 0.0815\n",
      "Epoch 403/1000\n",
      "20/20 - 0s - loss: -7.4290e-02 - MI: 0.0749\n",
      "Epoch 404/1000\n",
      "20/20 - 0s - loss: -7.6474e-02 - MI: 0.0762\n",
      "Epoch 405/1000\n",
      "20/20 - 0s - loss: -7.3652e-02 - MI: 0.0746\n",
      "Epoch 406/1000\n",
      "20/20 - 0s - loss: -7.4216e-02 - MI: 0.0735\n",
      "Epoch 407/1000\n",
      "20/20 - 0s - loss: -8.1707e-02 - MI: 0.0810\n",
      "Epoch 408/1000\n",
      "20/20 - 0s - loss: -7.6597e-02 - MI: 0.0770\n",
      "Epoch 409/1000\n",
      "20/20 - 0s - loss: -7.5059e-02 - MI: 0.0744\n",
      "Epoch 410/1000\n",
      "20/20 - 0s - loss: -7.7642e-02 - MI: 0.0770\n",
      "Epoch 411/1000\n",
      "20/20 - 0s - loss: -7.5512e-02 - MI: 0.0757\n",
      "Epoch 412/1000\n",
      "20/20 - 0s - loss: -7.9160e-02 - MI: 0.0794\n",
      "Epoch 413/1000\n",
      "20/20 - 0s - loss: -7.7864e-02 - MI: 0.0773\n",
      "Epoch 414/1000\n",
      "20/20 - 0s - loss: -8.1370e-02 - MI: 0.0817\n",
      "Epoch 415/1000\n",
      "20/20 - 0s - loss: -7.1748e-02 - MI: 0.0711\n",
      "Epoch 416/1000\n",
      "20/20 - 0s - loss: -7.8235e-02 - MI: 0.0779\n",
      "Epoch 417/1000\n",
      "20/20 - 0s - loss: -7.9359e-02 - MI: 0.0788\n",
      "Epoch 418/1000\n",
      "20/20 - 0s - loss: -7.1784e-02 - MI: 0.0707\n",
      "Epoch 419/1000\n",
      "20/20 - 0s - loss: -7.9304e-02 - MI: 0.0792\n",
      "Epoch 420/1000\n",
      "20/20 - 0s - loss: -8.4270e-02 - MI: 0.0831\n",
      "Epoch 421/1000\n",
      "20/20 - 0s - loss: -8.0883e-02 - MI: 0.0813\n",
      "Epoch 422/1000\n",
      "20/20 - 0s - loss: -7.6703e-02 - MI: 0.0772\n",
      "Epoch 423/1000\n",
      "20/20 - 0s - loss: -8.0539e-02 - MI: 0.0812\n",
      "Epoch 424/1000\n",
      "20/20 - 0s - loss: -8.1508e-02 - MI: 0.0800\n",
      "Epoch 425/1000\n",
      "20/20 - 0s - loss: -8.4164e-02 - MI: 0.0824\n",
      "Epoch 426/1000\n",
      "20/20 - 0s - loss: -7.7236e-02 - MI: 0.0762\n",
      "Epoch 427/1000\n",
      "20/20 - 0s - loss: -8.0839e-02 - MI: 0.0816\n",
      "Epoch 428/1000\n",
      "20/20 - 0s - loss: -8.4381e-02 - MI: 0.0841\n",
      "Epoch 429/1000\n",
      "20/20 - 0s - loss: -8.5990e-02 - MI: 0.0865\n",
      "Epoch 430/1000\n",
      "20/20 - 0s - loss: -8.0988e-02 - MI: 0.0807\n",
      "Epoch 431/1000\n",
      "20/20 - 0s - loss: -8.3688e-02 - MI: 0.0844\n",
      "Epoch 432/1000\n",
      "20/20 - 0s - loss: -9.0813e-02 - MI: 0.0914\n",
      "Epoch 433/1000\n",
      "20/20 - 0s - loss: -8.1092e-02 - MI: 0.0814\n",
      "Epoch 434/1000\n",
      "20/20 - 0s - loss: -8.1016e-02 - MI: 0.0820\n",
      "Epoch 435/1000\n",
      "20/20 - 0s - loss: -8.8524e-02 - MI: 0.0887\n",
      "Epoch 436/1000\n",
      "20/20 - 0s - loss: -8.5500e-02 - MI: 0.0846\n",
      "Epoch 437/1000\n",
      "20/20 - 0s - loss: -8.7583e-02 - MI: 0.0886\n",
      "Epoch 438/1000\n",
      "20/20 - 0s - loss: -9.0168e-02 - MI: 0.0915\n",
      "Epoch 439/1000\n",
      "20/20 - 0s - loss: -8.5815e-02 - MI: 0.0845\n",
      "Epoch 440/1000\n",
      "20/20 - 0s - loss: -8.6668e-02 - MI: 0.0862\n",
      "Epoch 441/1000\n",
      "20/20 - 0s - loss: -8.8985e-02 - MI: 0.0892\n",
      "Epoch 442/1000\n",
      "20/20 - 0s - loss: -8.2325e-02 - MI: 0.0825\n",
      "Epoch 443/1000\n",
      "20/20 - 0s - loss: -8.2367e-02 - MI: 0.0815\n",
      "Epoch 444/1000\n",
      "20/20 - 0s - loss: -8.6778e-02 - MI: 0.0866\n",
      "Epoch 445/1000\n",
      "20/20 - 0s - loss: -8.2470e-02 - MI: 0.0830\n",
      "Epoch 446/1000\n",
      "20/20 - 0s - loss: -8.4768e-02 - MI: 0.0846\n",
      "Epoch 447/1000\n",
      "20/20 - 0s - loss: -8.3510e-02 - MI: 0.0839\n",
      "Epoch 448/1000\n",
      "20/20 - 0s - loss: -8.7747e-02 - MI: 0.0881\n",
      "Epoch 449/1000\n",
      "20/20 - 0s - loss: -8.3557e-02 - MI: 0.0833\n",
      "Epoch 450/1000\n",
      "20/20 - 0s - loss: -7.9839e-02 - MI: 0.0804\n",
      "Epoch 451/1000\n",
      "20/20 - 0s - loss: -7.9613e-02 - MI: 0.0813\n",
      "Epoch 452/1000\n",
      "20/20 - 0s - loss: -8.4011e-02 - MI: 0.0841\n",
      "Epoch 453/1000\n",
      "20/20 - 0s - loss: -8.6910e-02 - MI: 0.0868\n",
      "Epoch 454/1000\n",
      "20/20 - 0s - loss: -9.0511e-02 - MI: 0.0900\n",
      "Epoch 455/1000\n",
      "20/20 - 0s - loss: -8.8809e-02 - MI: 0.0892\n",
      "Epoch 456/1000\n",
      "20/20 - 0s - loss: -8.9248e-02 - MI: 0.0893\n",
      "Epoch 457/1000\n",
      "20/20 - 0s - loss: -8.8665e-02 - MI: 0.0888\n",
      "Epoch 458/1000\n",
      "20/20 - 0s - loss: -9.1660e-02 - MI: 0.0915\n",
      "Epoch 459/1000\n",
      "20/20 - 0s - loss: -8.7926e-02 - MI: 0.0882\n",
      "Epoch 460/1000\n",
      "20/20 - 0s - loss: -8.9522e-02 - MI: 0.0896\n",
      "Epoch 461/1000\n",
      "20/20 - 0s - loss: -8.7018e-02 - MI: 0.0863\n",
      "Epoch 462/1000\n",
      "20/20 - 0s - loss: -9.2411e-02 - MI: 0.0929\n",
      "Epoch 463/1000\n",
      "20/20 - 0s - loss: -9.0473e-02 - MI: 0.0906\n",
      "Epoch 464/1000\n",
      "20/20 - 0s - loss: -9.5531e-02 - MI: 0.0947\n",
      "Epoch 465/1000\n",
      "20/20 - 0s - loss: -8.8666e-02 - MI: 0.0863\n",
      "Epoch 466/1000\n",
      "20/20 - 0s - loss: -8.5307e-02 - MI: 0.0854\n",
      "Epoch 467/1000\n",
      "20/20 - 0s - loss: -9.3183e-02 - MI: 0.0940\n",
      "Epoch 468/1000\n",
      "20/20 - 0s - loss: -9.3638e-02 - MI: 0.0946\n",
      "Epoch 469/1000\n",
      "20/20 - 0s - loss: -8.8813e-02 - MI: 0.0896\n",
      "Epoch 470/1000\n",
      "20/20 - 0s - loss: -9.2366e-02 - MI: 0.0922\n",
      "Epoch 471/1000\n",
      "20/20 - 0s - loss: -9.2641e-02 - MI: 0.0939\n",
      "Epoch 472/1000\n",
      "20/20 - 0s - loss: -9.8451e-02 - MI: 0.0981\n",
      "Epoch 473/1000\n",
      "20/20 - 0s - loss: -8.9143e-02 - MI: 0.0890\n",
      "Epoch 474/1000\n",
      "20/20 - 0s - loss: -9.1960e-02 - MI: 0.0914\n",
      "Epoch 475/1000\n",
      "20/20 - 0s - loss: -9.5486e-02 - MI: 0.0950\n",
      "Epoch 476/1000\n",
      "20/20 - 0s - loss: -9.8077e-02 - MI: 0.0998\n",
      "Epoch 477/1000\n",
      "20/20 - 0s - loss: -9.1753e-02 - MI: 0.0920\n",
      "Epoch 478/1000\n",
      "20/20 - 0s - loss: -8.3363e-02 - MI: 0.0835\n",
      "Epoch 479/1000\n",
      "20/20 - 0s - loss: -8.9233e-02 - MI: 0.0881\n",
      "Epoch 480/1000\n",
      "20/20 - 0s - loss: -9.9500e-02 - MI: 0.1000\n",
      "Epoch 481/1000\n",
      "20/20 - 0s - loss: -9.2581e-02 - MI: 0.0932\n",
      "Epoch 482/1000\n",
      "20/20 - 0s - loss: -9.6787e-02 - MI: 0.0965\n",
      "Epoch 483/1000\n",
      "20/20 - 0s - loss: -9.6548e-02 - MI: 0.0957\n",
      "Epoch 484/1000\n",
      "20/20 - 0s - loss: -9.1764e-02 - MI: 0.0912\n",
      "Epoch 485/1000\n",
      "20/20 - 0s - loss: -9.5289e-02 - MI: 0.0966\n",
      "Epoch 486/1000\n",
      "20/20 - 0s - loss: -9.3890e-02 - MI: 0.0937\n",
      "Epoch 487/1000\n",
      "20/20 - 0s - loss: -9.5223e-02 - MI: 0.0944\n",
      "Epoch 488/1000\n",
      "20/20 - 0s - loss: -9.5209e-02 - MI: 0.0958\n",
      "Epoch 489/1000\n",
      "20/20 - 0s - loss: -9.6013e-02 - MI: 0.0963\n",
      "Epoch 490/1000\n",
      "20/20 - 0s - loss: -9.5197e-02 - MI: 0.0950\n",
      "Epoch 491/1000\n",
      "20/20 - 0s - loss: -9.6688e-02 - MI: 0.0959\n",
      "Epoch 492/1000\n",
      "20/20 - 0s - loss: -9.6072e-02 - MI: 0.0966\n",
      "Epoch 493/1000\n",
      "20/20 - 0s - loss: -1.0095e-01 - MI: 0.1001\n",
      "Epoch 494/1000\n",
      "20/20 - 0s - loss: -9.4545e-02 - MI: 0.0941\n",
      "Epoch 495/1000\n",
      "20/20 - 0s - loss: -9.8179e-02 - MI: 0.0978\n",
      "Epoch 496/1000\n",
      "20/20 - 0s - loss: -1.0134e-01 - MI: 0.1013\n",
      "Epoch 497/1000\n",
      "20/20 - 0s - loss: -1.0025e-01 - MI: 0.1005\n",
      "Epoch 498/1000\n",
      "20/20 - 0s - loss: -1.0121e-01 - MI: 0.1010\n",
      "Epoch 499/1000\n",
      "20/20 - 0s - loss: -9.5408e-02 - MI: 0.0949\n",
      "Epoch 500/1000\n",
      "20/20 - 0s - loss: -1.0322e-01 - MI: 0.1044\n",
      "Epoch 501/1000\n",
      "20/20 - 0s - loss: -9.7504e-02 - MI: 0.0984\n",
      "Epoch 502/1000\n",
      "20/20 - 0s - loss: -9.0284e-02 - MI: 0.0901\n",
      "Epoch 503/1000\n",
      "20/20 - 0s - loss: -1.0558e-01 - MI: 0.1052\n",
      "Epoch 504/1000\n",
      "20/20 - 0s - loss: -9.0584e-02 - MI: 0.0898\n",
      "Epoch 505/1000\n",
      "20/20 - 0s - loss: -9.4859e-02 - MI: 0.0942\n",
      "Epoch 506/1000\n",
      "20/20 - 0s - loss: -9.9839e-02 - MI: 0.1003\n",
      "Epoch 507/1000\n",
      "20/20 - 0s - loss: -1.0076e-01 - MI: 0.1005\n",
      "Epoch 508/1000\n",
      "20/20 - 0s - loss: -1.0039e-01 - MI: 0.0987\n",
      "Epoch 509/1000\n",
      "20/20 - 0s - loss: -9.4144e-02 - MI: 0.0936\n",
      "Epoch 510/1000\n",
      "20/20 - 0s - loss: -1.0004e-01 - MI: 0.1001\n",
      "Epoch 511/1000\n",
      "20/20 - 0s - loss: -9.6933e-02 - MI: 0.0984\n",
      "Epoch 512/1000\n",
      "20/20 - 0s - loss: -1.0127e-01 - MI: 0.1008\n",
      "Epoch 513/1000\n",
      "20/20 - 0s - loss: -9.7441e-02 - MI: 0.0972\n",
      "Epoch 514/1000\n",
      "20/20 - 0s - loss: -1.0439e-01 - MI: 0.1044\n",
      "Epoch 515/1000\n",
      "20/20 - 0s - loss: -1.0272e-01 - MI: 0.1024\n",
      "Epoch 516/1000\n",
      "20/20 - 0s - loss: -1.0127e-01 - MI: 0.1005\n",
      "Epoch 517/1000\n",
      "20/20 - 0s - loss: -1.0153e-01 - MI: 0.1014\n",
      "Epoch 518/1000\n",
      "20/20 - 0s - loss: -9.4283e-02 - MI: 0.0952\n",
      "Epoch 519/1000\n",
      "20/20 - 0s - loss: -1.0307e-01 - MI: 0.1039\n",
      "Epoch 520/1000\n",
      "20/20 - 0s - loss: -1.0007e-01 - MI: 0.0989\n",
      "Epoch 521/1000\n",
      "20/20 - 0s - loss: -1.0214e-01 - MI: 0.1023\n",
      "Epoch 522/1000\n",
      "20/20 - 0s - loss: -9.6845e-02 - MI: 0.0967\n",
      "Epoch 523/1000\n",
      "20/20 - 0s - loss: -9.8238e-02 - MI: 0.0972\n",
      "Epoch 524/1000\n",
      "20/20 - 0s - loss: -1.0209e-01 - MI: 0.1027\n",
      "Epoch 525/1000\n",
      "20/20 - 0s - loss: -1.0037e-01 - MI: 0.1001\n",
      "Epoch 526/1000\n",
      "20/20 - 0s - loss: -1.0268e-01 - MI: 0.1029\n",
      "Epoch 527/1000\n",
      "20/20 - 0s - loss: -1.0033e-01 - MI: 0.1010\n",
      "Epoch 528/1000\n",
      "20/20 - 0s - loss: -9.8633e-02 - MI: 0.0985\n",
      "Epoch 529/1000\n",
      "20/20 - 0s - loss: -9.8245e-02 - MI: 0.0982\n",
      "Epoch 530/1000\n",
      "20/20 - 0s - loss: -1.0382e-01 - MI: 0.1039\n",
      "Epoch 531/1000\n",
      "20/20 - 0s - loss: -9.8170e-02 - MI: 0.0981\n",
      "Epoch 532/1000\n",
      "20/20 - 0s - loss: -1.0488e-01 - MI: 0.1039\n",
      "Epoch 533/1000\n",
      "20/20 - 0s - loss: -9.1807e-02 - MI: 0.0911\n",
      "Epoch 534/1000\n",
      "20/20 - 0s - loss: -1.0614e-01 - MI: 0.1058\n",
      "Epoch 535/1000\n",
      "20/20 - 0s - loss: -1.0279e-01 - MI: 0.1027\n",
      "Epoch 536/1000\n",
      "20/20 - 0s - loss: -1.0733e-01 - MI: 0.1068\n",
      "Epoch 537/1000\n",
      "20/20 - 0s - loss: -1.0805e-01 - MI: 0.1084\n",
      "Epoch 538/1000\n",
      "20/20 - 0s - loss: -1.0421e-01 - MI: 0.1033\n",
      "Epoch 539/1000\n",
      "20/20 - 0s - loss: -9.7897e-02 - MI: 0.0986\n",
      "Epoch 540/1000\n",
      "20/20 - 0s - loss: -1.0440e-01 - MI: 0.1045\n",
      "Epoch 541/1000\n",
      "20/20 - 0s - loss: -1.0323e-01 - MI: 0.1036\n",
      "Epoch 542/1000\n",
      "20/20 - 0s - loss: -1.0819e-01 - MI: 0.1084\n",
      "Epoch 543/1000\n",
      "20/20 - 0s - loss: -1.0311e-01 - MI: 0.1034\n",
      "Epoch 544/1000\n",
      "20/20 - 0s - loss: -1.0316e-01 - MI: 0.1040\n",
      "Epoch 545/1000\n",
      "20/20 - 0s - loss: -1.0530e-01 - MI: 0.1052\n",
      "Epoch 546/1000\n",
      "20/20 - 0s - loss: -9.6463e-02 - MI: 0.0970\n",
      "Epoch 547/1000\n",
      "20/20 - 0s - loss: -1.0507e-01 - MI: 0.1060\n",
      "Epoch 548/1000\n",
      "20/20 - 0s - loss: -1.0585e-01 - MI: 0.1062\n",
      "Epoch 549/1000\n",
      "20/20 - 0s - loss: -1.1027e-01 - MI: 0.1103\n",
      "Epoch 550/1000\n",
      "20/20 - 0s - loss: -1.0880e-01 - MI: 0.1097\n",
      "Epoch 551/1000\n",
      "20/20 - 0s - loss: -1.0424e-01 - MI: 0.1043\n",
      "Epoch 552/1000\n",
      "20/20 - 0s - loss: -1.0301e-01 - MI: 0.1035\n",
      "Epoch 553/1000\n",
      "20/20 - 0s - loss: -1.0593e-01 - MI: 0.1062\n",
      "Epoch 554/1000\n",
      "20/20 - 0s - loss: -1.0898e-01 - MI: 0.1092\n",
      "Epoch 555/1000\n",
      "20/20 - 0s - loss: -1.0339e-01 - MI: 0.1026\n",
      "Epoch 556/1000\n",
      "20/20 - 0s - loss: -1.0626e-01 - MI: 0.1068\n",
      "Epoch 557/1000\n",
      "20/20 - 0s - loss: -1.0397e-01 - MI: 0.1035\n",
      "Epoch 558/1000\n",
      "20/20 - 0s - loss: -1.0788e-01 - MI: 0.1079\n",
      "Epoch 559/1000\n",
      "20/20 - 0s - loss: -1.0700e-01 - MI: 0.1073\n",
      "Epoch 560/1000\n",
      "20/20 - 0s - loss: -1.0460e-01 - MI: 0.1042\n",
      "Epoch 561/1000\n",
      "20/20 - 0s - loss: -1.0584e-01 - MI: 0.1055\n",
      "Epoch 562/1000\n",
      "20/20 - 0s - loss: -1.0611e-01 - MI: 0.1063\n",
      "Epoch 563/1000\n",
      "20/20 - 0s - loss: -1.0960e-01 - MI: 0.1099\n",
      "Epoch 564/1000\n",
      "20/20 - 0s - loss: -1.0795e-01 - MI: 0.1093\n",
      "Epoch 565/1000\n",
      "20/20 - 0s - loss: -1.0103e-01 - MI: 0.1012\n",
      "Epoch 566/1000\n",
      "20/20 - 0s - loss: -1.1006e-01 - MI: 0.1087\n",
      "Epoch 567/1000\n",
      "20/20 - 0s - loss: -1.0824e-01 - MI: 0.1090\n",
      "Epoch 568/1000\n",
      "20/20 - 0s - loss: -1.0110e-01 - MI: 0.1003\n",
      "Epoch 569/1000\n",
      "20/20 - 0s - loss: -1.1027e-01 - MI: 0.1100\n",
      "Epoch 570/1000\n",
      "20/20 - 0s - loss: -1.0583e-01 - MI: 0.1061\n",
      "Epoch 571/1000\n",
      "20/20 - 0s - loss: -1.0762e-01 - MI: 0.1085\n",
      "Epoch 572/1000\n",
      "20/20 - 0s - loss: -1.0045e-01 - MI: 0.1001\n",
      "Epoch 573/1000\n",
      "20/20 - 0s - loss: -1.0955e-01 - MI: 0.1096\n",
      "Epoch 574/1000\n",
      "20/20 - 0s - loss: -1.1013e-01 - MI: 0.1095\n",
      "Epoch 575/1000\n",
      "20/20 - 0s - loss: -1.0349e-01 - MI: 0.1035\n",
      "Epoch 576/1000\n",
      "20/20 - 0s - loss: -1.0304e-01 - MI: 0.1031\n",
      "Epoch 577/1000\n",
      "20/20 - 0s - loss: -1.0567e-01 - MI: 0.1063\n",
      "Epoch 578/1000\n",
      "20/20 - 0s - loss: -1.0993e-01 - MI: 0.1100\n",
      "Epoch 579/1000\n",
      "20/20 - 0s - loss: -1.1241e-01 - MI: 0.1125\n",
      "Epoch 580/1000\n",
      "20/20 - 0s - loss: -1.1056e-01 - MI: 0.1107\n",
      "Epoch 581/1000\n",
      "20/20 - 0s - loss: -1.0254e-01 - MI: 0.1033\n",
      "Epoch 582/1000\n",
      "20/20 - 0s - loss: -1.1218e-01 - MI: 0.1117\n",
      "Epoch 583/1000\n",
      "20/20 - 0s - loss: -1.0520e-01 - MI: 0.1049\n",
      "Epoch 584/1000\n",
      "20/20 - 0s - loss: -1.1147e-01 - MI: 0.1116\n",
      "Epoch 585/1000\n",
      "20/20 - 0s - loss: -1.0857e-01 - MI: 0.1085\n",
      "Epoch 586/1000\n",
      "20/20 - 0s - loss: -1.1299e-01 - MI: 0.1127\n",
      "Epoch 587/1000\n",
      "20/20 - 0s - loss: -1.0997e-01 - MI: 0.1097\n",
      "Epoch 588/1000\n",
      "20/20 - 0s - loss: -1.0743e-01 - MI: 0.1077\n",
      "Epoch 589/1000\n",
      "20/20 - 0s - loss: -1.1519e-01 - MI: 0.1156\n",
      "Epoch 590/1000\n",
      "20/20 - 0s - loss: -1.0891e-01 - MI: 0.1087\n",
      "Epoch 591/1000\n",
      "20/20 - 0s - loss: -1.1257e-01 - MI: 0.1126\n",
      "Epoch 592/1000\n",
      "20/20 - 0s - loss: -1.0978e-01 - MI: 0.1099\n",
      "Epoch 593/1000\n",
      "20/20 - 0s - loss: -1.1408e-01 - MI: 0.1144\n",
      "Epoch 594/1000\n",
      "20/20 - 0s - loss: -1.0969e-01 - MI: 0.1097\n",
      "Epoch 595/1000\n",
      "20/20 - 0s - loss: -1.0807e-01 - MI: 0.1091\n",
      "Epoch 596/1000\n",
      "20/20 - 0s - loss: -1.0226e-01 - MI: 0.1030\n",
      "Epoch 597/1000\n",
      "20/20 - 0s - loss: -1.0796e-01 - MI: 0.1088\n",
      "Epoch 598/1000\n",
      "20/20 - 0s - loss: -1.1360e-01 - MI: 0.1136\n",
      "Epoch 599/1000\n",
      "20/20 - 0s - loss: -1.0461e-01 - MI: 0.1038\n",
      "Epoch 600/1000\n",
      "20/20 - 0s - loss: -1.1227e-01 - MI: 0.1135\n",
      "Epoch 601/1000\n",
      "20/20 - 0s - loss: -1.1241e-01 - MI: 0.1118\n",
      "Epoch 602/1000\n",
      "20/20 - 0s - loss: -1.1325e-01 - MI: 0.1134\n",
      "Epoch 603/1000\n",
      "20/20 - 0s - loss: -1.0719e-01 - MI: 0.1084\n",
      "Epoch 604/1000\n",
      "20/20 - 0s - loss: -1.1063e-01 - MI: 0.1111\n",
      "Epoch 605/1000\n",
      "20/20 - 0s - loss: -1.1138e-01 - MI: 0.1119\n",
      "Epoch 606/1000\n",
      "20/20 - 0s - loss: -1.0876e-01 - MI: 0.1092\n",
      "Epoch 607/1000\n",
      "20/20 - 0s - loss: -1.1104e-01 - MI: 0.1107\n",
      "Epoch 608/1000\n",
      "20/20 - 0s - loss: -1.1102e-01 - MI: 0.1119\n",
      "Epoch 609/1000\n",
      "20/20 - 0s - loss: -1.1584e-01 - MI: 0.1155\n",
      "Epoch 610/1000\n",
      "20/20 - 0s - loss: -1.1463e-01 - MI: 0.1145\n",
      "Epoch 611/1000\n",
      "20/20 - 0s - loss: -1.0919e-01 - MI: 0.1081\n",
      "Epoch 612/1000\n",
      "20/20 - 0s - loss: -1.1528e-01 - MI: 0.1144\n",
      "Epoch 613/1000\n",
      "20/20 - 0s - loss: -1.0531e-01 - MI: 0.1058\n",
      "Epoch 614/1000\n",
      "20/20 - 0s - loss: -1.1107e-01 - MI: 0.1132\n",
      "Epoch 615/1000\n",
      "20/20 - 0s - loss: -1.1074e-01 - MI: 0.1104\n",
      "Epoch 616/1000\n",
      "20/20 - 0s - loss: -1.1180e-01 - MI: 0.1120\n",
      "Epoch 617/1000\n",
      "20/20 - 0s - loss: -1.1011e-01 - MI: 0.1096\n",
      "Epoch 618/1000\n",
      "20/20 - 0s - loss: -1.1494e-01 - MI: 0.1151\n",
      "Epoch 619/1000\n",
      "20/20 - 0s - loss: -1.0955e-01 - MI: 0.1088\n",
      "Epoch 620/1000\n",
      "20/20 - 0s - loss: -1.0583e-01 - MI: 0.1056\n",
      "Epoch 621/1000\n",
      "20/20 - 0s - loss: -1.0891e-01 - MI: 0.1098\n",
      "Epoch 622/1000\n",
      "20/20 - 0s - loss: -1.1122e-01 - MI: 0.1097\n",
      "Epoch 623/1000\n",
      "20/20 - 0s - loss: -1.0773e-01 - MI: 0.1086\n",
      "Epoch 624/1000\n",
      "20/20 - 0s - loss: -1.1426e-01 - MI: 0.1141\n",
      "Epoch 625/1000\n",
      "20/20 - 0s - loss: -1.1790e-01 - MI: 0.1176\n",
      "Epoch 626/1000\n",
      "20/20 - 0s - loss: -1.0757e-01 - MI: 0.1078\n",
      "Epoch 627/1000\n",
      "20/20 - 0s - loss: -1.1335e-01 - MI: 0.1142\n",
      "Epoch 628/1000\n",
      "20/20 - 0s - loss: -1.1075e-01 - MI: 0.1104\n",
      "Epoch 629/1000\n",
      "20/20 - 0s - loss: -1.1163e-01 - MI: 0.1120\n",
      "Epoch 630/1000\n",
      "20/20 - 0s - loss: -1.0926e-01 - MI: 0.1098\n",
      "Epoch 631/1000\n",
      "20/20 - 0s - loss: -1.1054e-01 - MI: 0.1096\n",
      "Epoch 632/1000\n",
      "20/20 - 0s - loss: -1.1057e-01 - MI: 0.1101\n",
      "Epoch 633/1000\n",
      "20/20 - 0s - loss: -1.1206e-01 - MI: 0.1113\n",
      "Epoch 634/1000\n",
      "20/20 - 0s - loss: -1.0995e-01 - MI: 0.1091\n",
      "Epoch 635/1000\n",
      "20/20 - 0s - loss: -1.1169e-01 - MI: 0.1114\n",
      "Epoch 636/1000\n",
      "20/20 - 0s - loss: -1.1503e-01 - MI: 0.1155\n",
      "Epoch 637/1000\n",
      "20/20 - 0s - loss: -1.1428e-01 - MI: 0.1142\n",
      "Epoch 638/1000\n",
      "20/20 - 0s - loss: -1.0940e-01 - MI: 0.1083\n",
      "Epoch 639/1000\n",
      "20/20 - 0s - loss: -1.1418e-01 - MI: 0.1135\n",
      "Epoch 640/1000\n",
      "20/20 - 0s - loss: -1.0566e-01 - MI: 0.1051\n",
      "Epoch 641/1000\n",
      "20/20 - 0s - loss: -1.1441e-01 - MI: 0.1143\n",
      "Epoch 642/1000\n",
      "20/20 - 0s - loss: -1.0730e-01 - MI: 0.1070\n",
      "Epoch 643/1000\n",
      "20/20 - 0s - loss: -1.1094e-01 - MI: 0.1101\n",
      "Epoch 644/1000\n",
      "20/20 - 0s - loss: -1.1374e-01 - MI: 0.1130\n",
      "Epoch 645/1000\n",
      "20/20 - 0s - loss: -1.0985e-01 - MI: 0.1093\n",
      "Epoch 646/1000\n",
      "20/20 - 0s - loss: -1.1231e-01 - MI: 0.1124\n",
      "Epoch 647/1000\n",
      "20/20 - 0s - loss: -1.0854e-01 - MI: 0.1091\n",
      "Epoch 648/1000\n",
      "20/20 - 0s - loss: -1.0948e-01 - MI: 0.1088\n",
      "Epoch 649/1000\n",
      "20/20 - 0s - loss: -1.1180e-01 - MI: 0.1121\n",
      "Epoch 650/1000\n",
      "20/20 - 0s - loss: -1.0861e-01 - MI: 0.1078\n",
      "Epoch 651/1000\n",
      "20/20 - 0s - loss: -1.1283e-01 - MI: 0.1123\n",
      "Epoch 652/1000\n",
      "20/20 - 0s - loss: -1.0452e-01 - MI: 0.1050\n",
      "Epoch 653/1000\n",
      "20/20 - 0s - loss: -1.1472e-01 - MI: 0.1152\n",
      "Epoch 654/1000\n",
      "20/20 - 0s - loss: -1.0885e-01 - MI: 0.1092\n",
      "Epoch 655/1000\n",
      "20/20 - 0s - loss: -1.0740e-01 - MI: 0.1085\n",
      "Epoch 656/1000\n",
      "20/20 - 0s - loss: -1.1008e-01 - MI: 0.1106\n",
      "Epoch 657/1000\n",
      "20/20 - 0s - loss: -1.1206e-01 - MI: 0.1123\n",
      "Epoch 658/1000\n",
      "20/20 - 0s - loss: -1.1809e-01 - MI: 0.1197\n",
      "Epoch 659/1000\n",
      "20/20 - 0s - loss: -1.1490e-01 - MI: 0.1156\n",
      "Epoch 660/1000\n",
      "20/20 - 0s - loss: -1.1042e-01 - MI: 0.1086\n",
      "Epoch 661/1000\n",
      "20/20 - 0s - loss: -1.1669e-01 - MI: 0.1163\n",
      "Epoch 662/1000\n",
      "20/20 - 0s - loss: -1.0890e-01 - MI: 0.1091\n",
      "Epoch 663/1000\n",
      "20/20 - 0s - loss: -1.0892e-01 - MI: 0.1074\n",
      "Epoch 664/1000\n",
      "20/20 - 0s - loss: -1.0782e-01 - MI: 0.1071\n",
      "Epoch 665/1000\n",
      "20/20 - 0s - loss: -1.0393e-01 - MI: 0.1031\n",
      "Epoch 666/1000\n",
      "20/20 - 0s - loss: -1.0816e-01 - MI: 0.1089\n",
      "Epoch 667/1000\n",
      "20/20 - 0s - loss: -1.1876e-01 - MI: 0.1182\n",
      "Epoch 668/1000\n",
      "20/20 - 0s - loss: -1.0716e-01 - MI: 0.1071\n",
      "Epoch 669/1000\n",
      "20/20 - 0s - loss: -1.1074e-01 - MI: 0.1100\n",
      "Epoch 670/1000\n",
      "20/20 - 0s - loss: -1.0997e-01 - MI: 0.1093\n",
      "Epoch 671/1000\n",
      "20/20 - 0s - loss: -1.1525e-01 - MI: 0.1161\n",
      "Epoch 672/1000\n",
      "20/20 - 0s - loss: -1.1153e-01 - MI: 0.1110\n",
      "Epoch 673/1000\n",
      "20/20 - 0s - loss: -1.0376e-01 - MI: 0.1027\n",
      "Epoch 674/1000\n",
      "20/20 - 0s - loss: -1.0582e-01 - MI: 0.1075\n",
      "Epoch 675/1000\n",
      "20/20 - 0s - loss: -1.1273e-01 - MI: 0.1134\n",
      "Epoch 676/1000\n",
      "20/20 - 0s - loss: -1.0989e-01 - MI: 0.1097\n",
      "Epoch 677/1000\n",
      "20/20 - 0s - loss: -1.0679e-01 - MI: 0.1074\n",
      "Epoch 678/1000\n",
      "20/20 - 0s - loss: -1.1158e-01 - MI: 0.1111\n",
      "Epoch 679/1000\n",
      "20/20 - 0s - loss: -1.0745e-01 - MI: 0.1081\n",
      "Epoch 680/1000\n",
      "20/20 - 0s - loss: -1.1377e-01 - MI: 0.1132\n",
      "Epoch 681/1000\n",
      "20/20 - 0s - loss: -1.1013e-01 - MI: 0.1101\n",
      "Epoch 682/1000\n",
      "20/20 - 0s - loss: -1.0927e-01 - MI: 0.1088\n",
      "Epoch 683/1000\n",
      "20/20 - 0s - loss: -1.1203e-01 - MI: 0.1118\n",
      "Epoch 684/1000\n",
      "20/20 - 0s - loss: -1.1280e-01 - MI: 0.1135\n",
      "Epoch 685/1000\n",
      "20/20 - 0s - loss: -1.0654e-01 - MI: 0.1063\n",
      "Epoch 686/1000\n",
      "20/20 - 0s - loss: -1.0679e-01 - MI: 0.1059\n",
      "Epoch 687/1000\n",
      "20/20 - 0s - loss: -1.1289e-01 - MI: 0.1137\n",
      "Epoch 688/1000\n",
      "20/20 - 0s - loss: -1.1323e-01 - MI: 0.1140\n",
      "Epoch 689/1000\n",
      "20/20 - 0s - loss: -1.1111e-01 - MI: 0.1114\n",
      "Epoch 690/1000\n",
      "20/20 - 0s - loss: -1.0985e-01 - MI: 0.1103\n",
      "Epoch 691/1000\n",
      "20/20 - 0s - loss: -1.0918e-01 - MI: 0.1099\n",
      "Epoch 692/1000\n",
      "20/20 - 0s - loss: -1.0848e-01 - MI: 0.1076\n",
      "Epoch 693/1000\n",
      "20/20 - 0s - loss: -1.0872e-01 - MI: 0.1092\n",
      "Epoch 694/1000\n",
      "20/20 - 0s - loss: -1.0491e-01 - MI: 0.1053\n",
      "Epoch 695/1000\n",
      "20/20 - 0s - loss: -1.0858e-01 - MI: 0.1088\n",
      "Epoch 696/1000\n",
      "20/20 - 0s - loss: -1.1685e-01 - MI: 0.1167\n",
      "Epoch 697/1000\n",
      "20/20 - 0s - loss: -1.1489e-01 - MI: 0.1155\n",
      "Epoch 698/1000\n",
      "20/20 - 0s - loss: -1.1685e-01 - MI: 0.1170\n",
      "Epoch 699/1000\n",
      "20/20 - 0s - loss: -1.1681e-01 - MI: 0.1167\n",
      "Epoch 700/1000\n",
      "20/20 - 0s - loss: -1.1193e-01 - MI: 0.1123\n",
      "Epoch 701/1000\n",
      "20/20 - 0s - loss: -1.1262e-01 - MI: 0.1129\n",
      "Epoch 702/1000\n",
      "20/20 - 0s - loss: -1.1159e-01 - MI: 0.1121\n",
      "Epoch 703/1000\n",
      "20/20 - 0s - loss: -1.1476e-01 - MI: 0.1146\n",
      "Epoch 704/1000\n",
      "20/20 - 0s - loss: -1.0200e-01 - MI: 0.1016\n",
      "Epoch 705/1000\n",
      "20/20 - 0s - loss: -1.1064e-01 - MI: 0.1107\n",
      "Epoch 706/1000\n",
      "20/20 - 0s - loss: -1.1521e-01 - MI: 0.1143\n",
      "Epoch 707/1000\n",
      "20/20 - 0s - loss: -1.0949e-01 - MI: 0.1095\n",
      "Epoch 708/1000\n",
      "20/20 - 0s - loss: -1.1436e-01 - MI: 0.1149\n",
      "Epoch 709/1000\n",
      "20/20 - 0s - loss: -1.1398e-01 - MI: 0.1139\n",
      "Epoch 710/1000\n",
      "20/20 - 0s - loss: -1.0725e-01 - MI: 0.1079\n",
      "Epoch 711/1000\n",
      "20/20 - 0s - loss: -1.1536e-01 - MI: 0.1150\n",
      "Epoch 712/1000\n",
      "20/20 - 0s - loss: -1.0489e-01 - MI: 0.1051\n",
      "Epoch 713/1000\n",
      "20/20 - 0s - loss: -1.1327e-01 - MI: 0.1131\n",
      "Epoch 714/1000\n",
      "20/20 - 0s - loss: -1.1446e-01 - MI: 0.1142\n",
      "Epoch 715/1000\n",
      "20/20 - 0s - loss: -1.1558e-01 - MI: 0.1157\n",
      "Epoch 716/1000\n",
      "20/20 - 0s - loss: -1.1411e-01 - MI: 0.1133\n",
      "Epoch 717/1000\n",
      "20/20 - 0s - loss: -1.1514e-01 - MI: 0.1163\n",
      "Epoch 718/1000\n",
      "20/20 - 0s - loss: -1.1593e-01 - MI: 0.1171\n",
      "Epoch 719/1000\n",
      "20/20 - 0s - loss: -1.1356e-01 - MI: 0.1128\n",
      "Epoch 720/1000\n",
      "20/20 - 0s - loss: -1.0709e-01 - MI: 0.1074\n",
      "Epoch 721/1000\n",
      "20/20 - 0s - loss: -1.1221e-01 - MI: 0.1122\n",
      "Epoch 722/1000\n",
      "20/20 - 0s - loss: -1.0966e-01 - MI: 0.1097\n",
      "Epoch 723/1000\n",
      "20/20 - 0s - loss: -1.1471e-01 - MI: 0.1151\n",
      "Epoch 724/1000\n",
      "20/20 - 0s - loss: -1.1365e-01 - MI: 0.1137\n",
      "Epoch 725/1000\n",
      "20/20 - 0s - loss: -1.1759e-01 - MI: 0.1174\n",
      "Epoch 726/1000\n",
      "20/20 - 0s - loss: -1.1778e-01 - MI: 0.1178\n",
      "Epoch 727/1000\n",
      "20/20 - 0s - loss: -1.0687e-01 - MI: 0.1067\n",
      "Epoch 728/1000\n",
      "20/20 - 0s - loss: -1.0888e-01 - MI: 0.1076\n",
      "Epoch 729/1000\n",
      "20/20 - 0s - loss: -1.1395e-01 - MI: 0.1137\n",
      "Epoch 730/1000\n",
      "20/20 - 0s - loss: -1.1041e-01 - MI: 0.1107\n",
      "Epoch 731/1000\n",
      "20/20 - 0s - loss: -1.1356e-01 - MI: 0.1134\n",
      "Epoch 732/1000\n",
      "20/20 - 0s - loss: -1.1092e-01 - MI: 0.1109\n",
      "Epoch 733/1000\n",
      "20/20 - 0s - loss: -1.1355e-01 - MI: 0.1132\n",
      "Epoch 734/1000\n",
      "20/20 - 0s - loss: -1.0764e-01 - MI: 0.1073\n",
      "Epoch 735/1000\n",
      "20/20 - 0s - loss: -1.1110e-01 - MI: 0.1105\n",
      "Epoch 736/1000\n",
      "20/20 - 0s - loss: -1.1072e-01 - MI: 0.1121\n",
      "Epoch 737/1000\n",
      "20/20 - 0s - loss: -1.1791e-01 - MI: 0.1174\n",
      "Epoch 738/1000\n",
      "20/20 - 0s - loss: -1.0852e-01 - MI: 0.1071\n",
      "Epoch 739/1000\n",
      "20/20 - 0s - loss: -1.0397e-01 - MI: 0.1055\n",
      "Epoch 740/1000\n",
      "20/20 - 0s - loss: -1.0434e-01 - MI: 0.1043\n",
      "Epoch 741/1000\n",
      "20/20 - 0s - loss: -1.1476e-01 - MI: 0.1156\n",
      "Epoch 742/1000\n",
      "20/20 - 0s - loss: -1.1338e-01 - MI: 0.1126\n",
      "Epoch 743/1000\n",
      "20/20 - 0s - loss: -1.1188e-01 - MI: 0.1121\n",
      "Epoch 744/1000\n",
      "20/20 - 0s - loss: -1.0686e-01 - MI: 0.1065\n",
      "Epoch 745/1000\n",
      "20/20 - 0s - loss: -1.1550e-01 - MI: 0.1153\n",
      "Epoch 746/1000\n",
      "20/20 - 0s - loss: -1.1119e-01 - MI: 0.1112\n",
      "Epoch 747/1000\n",
      "20/20 - 0s - loss: -1.0677e-01 - MI: 0.1071\n",
      "Epoch 748/1000\n",
      "20/20 - 0s - loss: -1.1521e-01 - MI: 0.1157\n",
      "Epoch 749/1000\n",
      "20/20 - 0s - loss: -1.1052e-01 - MI: 0.1102\n",
      "Epoch 750/1000\n",
      "20/20 - 0s - loss: -1.0818e-01 - MI: 0.1068\n",
      "Epoch 751/1000\n",
      "20/20 - 0s - loss: -1.1679e-01 - MI: 0.1160\n",
      "Epoch 752/1000\n",
      "20/20 - 0s - loss: -1.1009e-01 - MI: 0.1103\n",
      "Epoch 753/1000\n",
      "20/20 - 0s - loss: -1.1379e-01 - MI: 0.1135\n",
      "Epoch 754/1000\n",
      "20/20 - 0s - loss: -1.1241e-01 - MI: 0.1122\n",
      "Epoch 755/1000\n",
      "20/20 - 0s - loss: -1.0973e-01 - MI: 0.1095\n",
      "Epoch 756/1000\n",
      "20/20 - 0s - loss: -1.0691e-01 - MI: 0.1070\n",
      "Epoch 757/1000\n",
      "20/20 - 0s - loss: -1.0735e-01 - MI: 0.1092\n",
      "Epoch 758/1000\n",
      "20/20 - 0s - loss: -1.1845e-01 - MI: 0.1192\n",
      "Epoch 759/1000\n",
      "20/20 - 0s - loss: -1.0895e-01 - MI: 0.1095\n",
      "Epoch 760/1000\n",
      "20/20 - 0s - loss: -1.1406e-01 - MI: 0.1139\n",
      "Epoch 761/1000\n",
      "20/20 - 0s - loss: -1.0993e-01 - MI: 0.1101\n",
      "Epoch 762/1000\n",
      "20/20 - 0s - loss: -1.0496e-01 - MI: 0.1054\n",
      "Epoch 763/1000\n",
      "20/20 - 0s - loss: -1.0287e-01 - MI: 0.1029\n",
      "Epoch 764/1000\n",
      "20/20 - 0s - loss: -1.1434e-01 - MI: 0.1146\n",
      "Epoch 765/1000\n",
      "20/20 - 0s - loss: -1.1440e-01 - MI: 0.1145\n",
      "Epoch 766/1000\n",
      "20/20 - 0s - loss: -1.0281e-01 - MI: 0.1036\n",
      "Epoch 767/1000\n",
      "20/20 - 0s - loss: -1.1868e-01 - MI: 0.1180\n",
      "Epoch 768/1000\n",
      "20/20 - 0s - loss: -1.0697e-01 - MI: 0.1068\n",
      "Epoch 769/1000\n",
      "20/20 - 0s - loss: -1.0901e-01 - MI: 0.1101\n",
      "Epoch 770/1000\n",
      "20/20 - 0s - loss: -1.1160e-01 - MI: 0.1119\n",
      "Epoch 771/1000\n",
      "20/20 - 0s - loss: -1.0758e-01 - MI: 0.1077\n",
      "Epoch 772/1000\n",
      "20/20 - 0s - loss: -1.1054e-01 - MI: 0.1111\n",
      "Epoch 773/1000\n",
      "20/20 - 0s - loss: -1.1230e-01 - MI: 0.1115\n",
      "Epoch 774/1000\n",
      "20/20 - 0s - loss: -1.0873e-01 - MI: 0.1092\n",
      "Epoch 775/1000\n",
      "20/20 - 0s - loss: -1.0577e-01 - MI: 0.1049\n",
      "Epoch 776/1000\n",
      "20/20 - 0s - loss: -1.0893e-01 - MI: 0.1075\n",
      "Epoch 777/1000\n",
      "20/20 - 0s - loss: -1.1257e-01 - MI: 0.1131\n",
      "Epoch 778/1000\n",
      "20/20 - 0s - loss: -1.0714e-01 - MI: 0.1073\n",
      "Epoch 779/1000\n",
      "20/20 - 0s - loss: -1.1535e-01 - MI: 0.1154\n",
      "Epoch 780/1000\n",
      "20/20 - 0s - loss: -1.1561e-01 - MI: 0.1155\n",
      "Epoch 781/1000\n",
      "20/20 - 0s - loss: -1.1121e-01 - MI: 0.1115\n",
      "Epoch 782/1000\n",
      "20/20 - 0s - loss: -1.1337e-01 - MI: 0.1138\n",
      "Epoch 783/1000\n",
      "20/20 - 0s - loss: -1.1966e-01 - MI: 0.1190\n",
      "Epoch 784/1000\n",
      "20/20 - 0s - loss: -1.0692e-01 - MI: 0.1071\n",
      "Epoch 785/1000\n",
      "20/20 - 0s - loss: -1.1451e-01 - MI: 0.1155\n",
      "Epoch 786/1000\n",
      "20/20 - 0s - loss: -1.1220e-01 - MI: 0.1118\n",
      "Epoch 787/1000\n",
      "20/20 - 0s - loss: -1.1364e-01 - MI: 0.1132\n",
      "Epoch 788/1000\n",
      "20/20 - 0s - loss: -1.1515e-01 - MI: 0.1143\n",
      "Epoch 789/1000\n",
      "20/20 - 0s - loss: -1.1599e-01 - MI: 0.1158\n",
      "Epoch 790/1000\n",
      "20/20 - 0s - loss: -1.0826e-01 - MI: 0.1078\n",
      "Epoch 791/1000\n",
      "20/20 - 0s - loss: -1.1102e-01 - MI: 0.1104\n",
      "Epoch 792/1000\n",
      "20/20 - 0s - loss: -1.1011e-01 - MI: 0.1090\n",
      "Epoch 793/1000\n",
      "20/20 - 0s - loss: -1.1420e-01 - MI: 0.1140\n",
      "Epoch 794/1000\n",
      "20/20 - 0s - loss: -1.0909e-01 - MI: 0.1086\n",
      "Epoch 795/1000\n",
      "20/20 - 0s - loss: -1.1146e-01 - MI: 0.1115\n",
      "Epoch 796/1000\n",
      "20/20 - 0s - loss: -1.1047e-01 - MI: 0.1106\n",
      "Epoch 797/1000\n",
      "20/20 - 0s - loss: -1.1061e-01 - MI: 0.1111\n",
      "Epoch 798/1000\n",
      "20/20 - 0s - loss: -1.0284e-01 - MI: 0.1032\n",
      "Epoch 799/1000\n",
      "20/20 - 0s - loss: -1.0957e-01 - MI: 0.1094\n",
      "Epoch 800/1000\n",
      "20/20 - 0s - loss: -1.1204e-01 - MI: 0.1108\n",
      "Epoch 801/1000\n",
      "20/20 - 0s - loss: -1.0922e-01 - MI: 0.1097\n",
      "Epoch 802/1000\n",
      "20/20 - 0s - loss: -1.0883e-01 - MI: 0.1080\n",
      "Epoch 803/1000\n",
      "20/20 - 0s - loss: -1.1043e-01 - MI: 0.1106\n",
      "Epoch 804/1000\n",
      "20/20 - 0s - loss: -1.1347e-01 - MI: 0.1135\n",
      "Epoch 805/1000\n",
      "20/20 - 0s - loss: -1.1643e-01 - MI: 0.1167\n",
      "Epoch 806/1000\n",
      "20/20 - 0s - loss: -1.1156e-01 - MI: 0.1116\n",
      "Epoch 807/1000\n",
      "20/20 - 0s - loss: -1.0936e-01 - MI: 0.1093\n",
      "Epoch 808/1000\n",
      "20/20 - 0s - loss: -1.1324e-01 - MI: 0.1142\n",
      "Epoch 809/1000\n",
      "20/20 - 0s - loss: -1.1171e-01 - MI: 0.1116\n",
      "Epoch 810/1000\n",
      "20/20 - 0s - loss: -1.1375e-01 - MI: 0.1133\n",
      "Epoch 811/1000\n",
      "20/20 - 0s - loss: -1.1479e-01 - MI: 0.1144\n",
      "Epoch 812/1000\n",
      "20/20 - 0s - loss: -1.0920e-01 - MI: 0.1112\n",
      "Epoch 813/1000\n",
      "20/20 - 0s - loss: -1.1229e-01 - MI: 0.1126\n",
      "Epoch 814/1000\n",
      "20/20 - 0s - loss: -1.1067e-01 - MI: 0.1108\n",
      "Epoch 815/1000\n",
      "20/20 - 0s - loss: -1.1567e-01 - MI: 0.1154\n",
      "Epoch 816/1000\n",
      "20/20 - 0s - loss: -1.1234e-01 - MI: 0.1123\n",
      "Epoch 817/1000\n",
      "20/20 - 0s - loss: -1.1280e-01 - MI: 0.1130\n",
      "Epoch 818/1000\n",
      "20/20 - 0s - loss: -1.0719e-01 - MI: 0.1075\n",
      "Epoch 819/1000\n",
      "20/20 - 0s - loss: -1.1675e-01 - MI: 0.1169\n",
      "Epoch 820/1000\n",
      "20/20 - 0s - loss: -1.1158e-01 - MI: 0.1099\n",
      "Epoch 821/1000\n",
      "20/20 - 0s - loss: -1.1508e-01 - MI: 0.1135\n",
      "Epoch 822/1000\n",
      "20/20 - 0s - loss: -1.1010e-01 - MI: 0.1089\n",
      "Epoch 823/1000\n",
      "20/20 - 0s - loss: -1.0785e-01 - MI: 0.1075\n",
      "Epoch 824/1000\n",
      "20/20 - 0s - loss: -1.1268e-01 - MI: 0.1114\n",
      "Epoch 825/1000\n",
      "20/20 - 0s - loss: -1.0774e-01 - MI: 0.1075\n",
      "Epoch 826/1000\n",
      "20/20 - 0s - loss: -1.0351e-01 - MI: 0.1031\n",
      "Epoch 827/1000\n",
      "20/20 - 0s - loss: -1.0741e-01 - MI: 0.1065\n",
      "Epoch 828/1000\n",
      "20/20 - 0s - loss: -1.1186e-01 - MI: 0.1114\n",
      "Epoch 829/1000\n",
      "20/20 - 0s - loss: -1.0982e-01 - MI: 0.1094\n",
      "Epoch 830/1000\n",
      "20/20 - 0s - loss: -1.1010e-01 - MI: 0.1099\n",
      "Epoch 831/1000\n",
      "20/20 - 0s - loss: -1.0873e-01 - MI: 0.1092\n",
      "Epoch 832/1000\n",
      "20/20 - 0s - loss: -1.1104e-01 - MI: 0.1108\n",
      "Epoch 833/1000\n",
      "20/20 - 0s - loss: -1.1144e-01 - MI: 0.1112\n",
      "Epoch 834/1000\n",
      "20/20 - 0s - loss: -1.1514e-01 - MI: 0.1142\n",
      "Epoch 835/1000\n",
      "20/20 - 0s - loss: -1.1600e-01 - MI: 0.1148\n",
      "Epoch 836/1000\n",
      "20/20 - 0s - loss: -1.1136e-01 - MI: 0.1104\n",
      "Epoch 837/1000\n",
      "20/20 - 0s - loss: -1.0959e-01 - MI: 0.1109\n",
      "Epoch 838/1000\n",
      "20/20 - 0s - loss: -1.1258e-01 - MI: 0.1109\n",
      "Epoch 839/1000\n",
      "20/20 - 0s - loss: -1.0840e-01 - MI: 0.1078\n",
      "Epoch 840/1000\n",
      "20/20 - 0s - loss: -1.0956e-01 - MI: 0.1088\n",
      "Epoch 841/1000\n",
      "20/20 - 0s - loss: -1.0436e-01 - MI: 0.1043\n",
      "Epoch 842/1000\n",
      "20/20 - 0s - loss: -1.1202e-01 - MI: 0.1118\n",
      "Epoch 843/1000\n",
      "20/20 - 0s - loss: -1.1345e-01 - MI: 0.1130\n",
      "Epoch 844/1000\n",
      "20/20 - 0s - loss: -1.1777e-01 - MI: 0.1181\n",
      "Epoch 845/1000\n",
      "20/20 - 0s - loss: -1.0744e-01 - MI: 0.1064\n",
      "Epoch 846/1000\n",
      "20/20 - 0s - loss: -1.1327e-01 - MI: 0.1137\n",
      "Epoch 847/1000\n",
      "20/20 - 0s - loss: -1.1124e-01 - MI: 0.1119\n",
      "Epoch 848/1000\n",
      "20/20 - 0s - loss: -1.1458e-01 - MI: 0.1131\n",
      "Epoch 849/1000\n",
      "20/20 - 0s - loss: -1.1062e-01 - MI: 0.1121\n",
      "Epoch 850/1000\n",
      "20/20 - 0s - loss: -1.1066e-01 - MI: 0.1110\n",
      "Epoch 851/1000\n",
      "20/20 - 0s - loss: -1.1617e-01 - MI: 0.1163\n",
      "Epoch 852/1000\n",
      "20/20 - 0s - loss: -1.1000e-01 - MI: 0.1097\n",
      "Epoch 853/1000\n",
      "20/20 - 0s - loss: -1.0676e-01 - MI: 0.1064\n",
      "Epoch 854/1000\n",
      "20/20 - 0s - loss: -1.1400e-01 - MI: 0.1140\n",
      "Epoch 855/1000\n",
      "20/20 - 0s - loss: -1.1177e-01 - MI: 0.1116\n",
      "Epoch 856/1000\n",
      "20/20 - 0s - loss: -1.1351e-01 - MI: 0.1140\n",
      "Epoch 857/1000\n",
      "20/20 - 0s - loss: -1.1220e-01 - MI: 0.1120\n",
      "Epoch 858/1000\n",
      "20/20 - 0s - loss: -1.1783e-01 - MI: 0.1181\n",
      "Epoch 859/1000\n",
      "20/20 - 0s - loss: -1.0801e-01 - MI: 0.1082\n",
      "Epoch 860/1000\n",
      "20/20 - 0s - loss: -1.0892e-01 - MI: 0.1086\n",
      "Epoch 861/1000\n",
      "20/20 - 0s - loss: -1.1222e-01 - MI: 0.1132\n",
      "Epoch 862/1000\n",
      "20/20 - 0s - loss: -1.1117e-01 - MI: 0.1110\n",
      "Epoch 863/1000\n",
      "20/20 - 0s - loss: -1.1336e-01 - MI: 0.1148\n",
      "Epoch 864/1000\n",
      "20/20 - 0s - loss: -1.0912e-01 - MI: 0.1086\n",
      "Epoch 865/1000\n",
      "20/20 - 0s - loss: -1.0697e-01 - MI: 0.1064\n",
      "Epoch 866/1000\n",
      "20/20 - 0s - loss: -1.0491e-01 - MI: 0.1049\n",
      "Epoch 867/1000\n",
      "20/20 - 0s - loss: -1.1073e-01 - MI: 0.1116\n",
      "Epoch 868/1000\n",
      "20/20 - 0s - loss: -1.0991e-01 - MI: 0.1097\n",
      "Epoch 869/1000\n",
      "20/20 - 0s - loss: -1.0882e-01 - MI: 0.1097\n",
      "Epoch 870/1000\n",
      "20/20 - 0s - loss: -1.0849e-01 - MI: 0.1081\n",
      "Epoch 871/1000\n",
      "20/20 - 0s - loss: -1.0764e-01 - MI: 0.1082\n",
      "Epoch 872/1000\n",
      "20/20 - 0s - loss: -1.1028e-01 - MI: 0.1095\n",
      "Epoch 873/1000\n",
      "20/20 - 0s - loss: -1.0964e-01 - MI: 0.1097\n",
      "Epoch 874/1000\n",
      "20/20 - 0s - loss: -1.1796e-01 - MI: 0.1173\n",
      "Epoch 875/1000\n",
      "20/20 - 0s - loss: -1.1696e-01 - MI: 0.1169\n",
      "Epoch 876/1000\n",
      "20/20 - 0s - loss: -1.1204e-01 - MI: 0.1124\n",
      "Epoch 877/1000\n",
      "20/20 - 0s - loss: -1.2038e-01 - MI: 0.1203\n",
      "Epoch 878/1000\n",
      "20/20 - 0s - loss: -1.1504e-01 - MI: 0.1151\n",
      "Epoch 879/1000\n",
      "20/20 - 0s - loss: -1.1692e-01 - MI: 0.1184\n",
      "Epoch 880/1000\n",
      "20/20 - 0s - loss: -1.1277e-01 - MI: 0.1127\n",
      "Epoch 881/1000\n",
      "20/20 - 0s - loss: -1.1083e-01 - MI: 0.1114\n",
      "Epoch 882/1000\n",
      "20/20 - 0s - loss: -1.1289e-01 - MI: 0.1131\n",
      "Epoch 883/1000\n",
      "20/20 - 0s - loss: -1.1467e-01 - MI: 0.1128\n",
      "Epoch 884/1000\n",
      "20/20 - 0s - loss: -1.0972e-01 - MI: 0.1104\n",
      "Epoch 885/1000\n",
      "20/20 - 0s - loss: -1.1607e-01 - MI: 0.1169\n",
      "Epoch 886/1000\n",
      "20/20 - 0s - loss: -1.1238e-01 - MI: 0.1130\n",
      "Epoch 887/1000\n",
      "20/20 - 0s - loss: -1.1614e-01 - MI: 0.1158\n",
      "Epoch 888/1000\n",
      "20/20 - 0s - loss: -1.1030e-01 - MI: 0.1116\n",
      "Epoch 889/1000\n",
      "20/20 - 0s - loss: -1.1525e-01 - MI: 0.1157\n",
      "Epoch 890/1000\n",
      "20/20 - 0s - loss: -1.1159e-01 - MI: 0.1109\n",
      "Epoch 891/1000\n",
      "20/20 - 0s - loss: -1.0765e-01 - MI: 0.1085\n",
      "Epoch 892/1000\n",
      "20/20 - 0s - loss: -1.0280e-01 - MI: 0.1025\n",
      "Epoch 893/1000\n",
      "20/20 - 0s - loss: -1.1306e-01 - MI: 0.1130\n",
      "Epoch 894/1000\n",
      "20/20 - 0s - loss: -1.0497e-01 - MI: 0.1049\n",
      "Epoch 895/1000\n",
      "20/20 - 0s - loss: -1.1134e-01 - MI: 0.1121\n",
      "Epoch 896/1000\n",
      "20/20 - 0s - loss: -1.0661e-01 - MI: 0.1067\n",
      "Epoch 897/1000\n",
      "20/20 - 0s - loss: -1.1176e-01 - MI: 0.1119\n",
      "Epoch 898/1000\n",
      "20/20 - 0s - loss: -1.1382e-01 - MI: 0.1142\n",
      "Epoch 899/1000\n",
      "20/20 - 0s - loss: -1.1016e-01 - MI: 0.1104\n",
      "Epoch 900/1000\n",
      "20/20 - 0s - loss: -1.0991e-01 - MI: 0.1099\n",
      "Epoch 901/1000\n",
      "20/20 - 0s - loss: -1.0678e-01 - MI: 0.1086\n",
      "Epoch 902/1000\n",
      "20/20 - 0s - loss: -1.0534e-01 - MI: 0.1042\n",
      "Epoch 903/1000\n",
      "20/20 - 0s - loss: -1.1255e-01 - MI: 0.1125\n",
      "Epoch 904/1000\n",
      "20/20 - 0s - loss: -1.1394e-01 - MI: 0.1131\n",
      "Epoch 905/1000\n",
      "20/20 - 0s - loss: -1.1091e-01 - MI: 0.1109\n",
      "Epoch 906/1000\n",
      "20/20 - 0s - loss: -1.0665e-01 - MI: 0.1070\n",
      "Epoch 907/1000\n",
      "20/20 - 0s - loss: -1.1155e-01 - MI: 0.1107\n",
      "Epoch 908/1000\n",
      "20/20 - 0s - loss: -1.0388e-01 - MI: 0.1036\n",
      "Epoch 909/1000\n",
      "20/20 - 0s - loss: -1.1180e-01 - MI: 0.1116\n",
      "Epoch 910/1000\n",
      "20/20 - 0s - loss: -1.1274e-01 - MI: 0.1124\n",
      "Epoch 911/1000\n",
      "20/20 - 0s - loss: -1.0793e-01 - MI: 0.1085\n",
      "Epoch 912/1000\n",
      "20/20 - 0s - loss: -1.1332e-01 - MI: 0.1135\n",
      "Epoch 913/1000\n",
      "20/20 - 0s - loss: -1.0883e-01 - MI: 0.1091\n",
      "Epoch 914/1000\n",
      "20/20 - 0s - loss: -1.1223e-01 - MI: 0.1128\n",
      "Epoch 915/1000\n",
      "20/20 - 0s - loss: -1.0996e-01 - MI: 0.1098\n",
      "Epoch 916/1000\n",
      "20/20 - 0s - loss: -1.1213e-01 - MI: 0.1116\n",
      "Epoch 917/1000\n",
      "20/20 - 0s - loss: -1.1131e-01 - MI: 0.1114\n",
      "Epoch 918/1000\n",
      "20/20 - 0s - loss: -1.0731e-01 - MI: 0.1081\n",
      "Epoch 919/1000\n",
      "20/20 - 0s - loss: -1.1027e-01 - MI: 0.1101\n",
      "Epoch 920/1000\n",
      "20/20 - 0s - loss: -1.1199e-01 - MI: 0.1124\n",
      "Epoch 921/1000\n",
      "20/20 - 0s - loss: -1.0874e-01 - MI: 0.1093\n",
      "Epoch 922/1000\n",
      "20/20 - 0s - loss: -1.1336e-01 - MI: 0.1134\n",
      "Epoch 923/1000\n",
      "20/20 - 0s - loss: -1.0852e-01 - MI: 0.1097\n",
      "Epoch 924/1000\n",
      "20/20 - 0s - loss: -1.0684e-01 - MI: 0.1065\n",
      "Epoch 925/1000\n",
      "20/20 - 0s - loss: -1.0825e-01 - MI: 0.1082\n",
      "Epoch 926/1000\n",
      "20/20 - 0s - loss: -1.0908e-01 - MI: 0.1094\n",
      "Epoch 927/1000\n",
      "20/20 - 0s - loss: -1.1479e-01 - MI: 0.1139\n",
      "Epoch 928/1000\n",
      "20/20 - 0s - loss: -1.1358e-01 - MI: 0.1128\n",
      "Epoch 929/1000\n",
      "20/20 - 0s - loss: -1.1034e-01 - MI: 0.1110\n",
      "Epoch 930/1000\n",
      "20/20 - 0s - loss: -1.1611e-01 - MI: 0.1163\n",
      "Epoch 931/1000\n",
      "20/20 - 0s - loss: -1.1676e-01 - MI: 0.1179\n",
      "Epoch 932/1000\n",
      "20/20 - 0s - loss: -1.1388e-01 - MI: 0.1138\n",
      "Epoch 933/1000\n",
      "20/20 - 0s - loss: -1.0896e-01 - MI: 0.1098\n",
      "Epoch 934/1000\n",
      "20/20 - 0s - loss: -1.1170e-01 - MI: 0.1113\n",
      "Epoch 935/1000\n",
      "20/20 - 0s - loss: -1.0920e-01 - MI: 0.1093\n",
      "Epoch 936/1000\n",
      "20/20 - 0s - loss: -1.1399e-01 - MI: 0.1131\n",
      "Epoch 937/1000\n",
      "20/20 - 0s - loss: -1.0675e-01 - MI: 0.1058\n",
      "Epoch 938/1000\n",
      "20/20 - 0s - loss: -1.0813e-01 - MI: 0.1074\n",
      "Epoch 939/1000\n",
      "20/20 - 0s - loss: -1.0898e-01 - MI: 0.1083\n",
      "Epoch 940/1000\n",
      "20/20 - 0s - loss: -1.1930e-01 - MI: 0.1188\n",
      "Epoch 941/1000\n",
      "20/20 - 0s - loss: -1.1278e-01 - MI: 0.1140\n",
      "Epoch 942/1000\n",
      "20/20 - 0s - loss: -1.1417e-01 - MI: 0.1138\n",
      "Epoch 943/1000\n",
      "20/20 - 0s - loss: -1.1541e-01 - MI: 0.1166\n",
      "Epoch 944/1000\n",
      "20/20 - 0s - loss: -1.1170e-01 - MI: 0.1109\n",
      "Epoch 945/1000\n",
      "20/20 - 0s - loss: -1.1741e-01 - MI: 0.1175\n",
      "Epoch 946/1000\n",
      "20/20 - 0s - loss: -1.1307e-01 - MI: 0.1121\n",
      "Epoch 947/1000\n",
      "20/20 - 0s - loss: -1.0728e-01 - MI: 0.1078\n",
      "Epoch 948/1000\n",
      "20/20 - 0s - loss: -1.1029e-01 - MI: 0.1107\n",
      "Epoch 949/1000\n",
      "20/20 - 0s - loss: -1.0874e-01 - MI: 0.1072\n",
      "Epoch 950/1000\n",
      "20/20 - 0s - loss: -1.1024e-01 - MI: 0.1096\n",
      "Epoch 951/1000\n",
      "20/20 - 0s - loss: -1.0853e-01 - MI: 0.1095\n",
      "Epoch 952/1000\n",
      "20/20 - 0s - loss: -1.0674e-01 - MI: 0.1060\n",
      "Epoch 953/1000\n",
      "20/20 - 0s - loss: -1.0710e-01 - MI: 0.1070\n",
      "Epoch 954/1000\n",
      "20/20 - 0s - loss: -1.1236e-01 - MI: 0.1128\n",
      "Epoch 955/1000\n",
      "20/20 - 0s - loss: -1.0952e-01 - MI: 0.1090\n",
      "Epoch 956/1000\n",
      "20/20 - 0s - loss: -1.0840e-01 - MI: 0.1080\n",
      "Epoch 957/1000\n",
      "20/20 - 0s - loss: -1.0816e-01 - MI: 0.1073\n",
      "Epoch 958/1000\n",
      "20/20 - 0s - loss: -1.1163e-01 - MI: 0.1118\n",
      "Epoch 959/1000\n",
      "20/20 - 0s - loss: -1.0829e-01 - MI: 0.1091\n",
      "Epoch 960/1000\n",
      "20/20 - 0s - loss: -1.0622e-01 - MI: 0.1055\n",
      "Epoch 961/1000\n",
      "20/20 - 0s - loss: -1.0629e-01 - MI: 0.1062\n",
      "Epoch 962/1000\n",
      "20/20 - 0s - loss: -1.1204e-01 - MI: 0.1122\n",
      "Epoch 963/1000\n",
      "20/20 - 0s - loss: -1.1020e-01 - MI: 0.1102\n",
      "Epoch 964/1000\n",
      "20/20 - 0s - loss: -1.0997e-01 - MI: 0.1095\n",
      "Epoch 965/1000\n",
      "20/20 - 0s - loss: -1.1926e-01 - MI: 0.1185\n",
      "Epoch 966/1000\n",
      "20/20 - 0s - loss: -1.0920e-01 - MI: 0.1093\n",
      "Epoch 967/1000\n",
      "20/20 - 0s - loss: -1.1402e-01 - MI: 0.1150\n",
      "Epoch 968/1000\n",
      "20/20 - 0s - loss: -1.1305e-01 - MI: 0.1140\n",
      "Epoch 969/1000\n",
      "20/20 - 0s - loss: -1.1115e-01 - MI: 0.1105\n",
      "Epoch 970/1000\n",
      "20/20 - 0s - loss: -1.1420e-01 - MI: 0.1147\n",
      "Epoch 971/1000\n",
      "20/20 - 0s - loss: -1.1467e-01 - MI: 0.1154\n",
      "Epoch 972/1000\n",
      "20/20 - 0s - loss: -1.1229e-01 - MI: 0.1124\n",
      "Epoch 973/1000\n",
      "20/20 - 0s - loss: -1.1173e-01 - MI: 0.1111\n",
      "Epoch 974/1000\n",
      "20/20 - 0s - loss: -1.1247e-01 - MI: 0.1126\n",
      "Epoch 975/1000\n",
      "20/20 - 0s - loss: -1.1188e-01 - MI: 0.1119\n",
      "Epoch 976/1000\n",
      "20/20 - 0s - loss: -1.1219e-01 - MI: 0.1124\n",
      "Epoch 977/1000\n",
      "20/20 - 0s - loss: -1.1556e-01 - MI: 0.1163\n",
      "Epoch 978/1000\n",
      "20/20 - 0s - loss: -1.1252e-01 - MI: 0.1116\n",
      "Epoch 979/1000\n",
      "20/20 - 0s - loss: -1.1349e-01 - MI: 0.1135\n",
      "Epoch 980/1000\n",
      "20/20 - 0s - loss: -1.1587e-01 - MI: 0.1163\n",
      "Epoch 981/1000\n",
      "20/20 - 0s - loss: -1.1011e-01 - MI: 0.1101\n",
      "Epoch 982/1000\n",
      "20/20 - 0s - loss: -1.1315e-01 - MI: 0.1131\n",
      "Epoch 983/1000\n",
      "20/20 - 0s - loss: -1.0426e-01 - MI: 0.1038\n",
      "Epoch 984/1000\n",
      "20/20 - 0s - loss: -1.1530e-01 - MI: 0.1150\n",
      "Epoch 985/1000\n",
      "20/20 - 0s - loss: -1.1299e-01 - MI: 0.1135\n",
      "Epoch 986/1000\n",
      "20/20 - 0s - loss: -1.0652e-01 - MI: 0.1071\n",
      "Epoch 987/1000\n",
      "20/20 - 0s - loss: -1.0971e-01 - MI: 0.1100\n",
      "Epoch 988/1000\n",
      "20/20 - 0s - loss: -1.0431e-01 - MI: 0.1050\n",
      "Epoch 989/1000\n",
      "20/20 - 0s - loss: -1.0606e-01 - MI: 0.1062\n",
      "Epoch 990/1000\n",
      "20/20 - 0s - loss: -1.1375e-01 - MI: 0.1150\n",
      "Epoch 991/1000\n",
      "20/20 - 0s - loss: -1.1451e-01 - MI: 0.1149\n",
      "Epoch 992/1000\n",
      "20/20 - 0s - loss: -1.1193e-01 - MI: 0.1114\n",
      "Epoch 993/1000\n",
      "20/20 - 0s - loss: -1.1562e-01 - MI: 0.1160\n",
      "Epoch 994/1000\n",
      "20/20 - 0s - loss: -1.0627e-01 - MI: 0.1073\n",
      "Epoch 995/1000\n",
      "20/20 - 0s - loss: -1.1106e-01 - MI: 0.1110\n",
      "Epoch 996/1000\n",
      "20/20 - 0s - loss: -1.1289e-01 - MI: 0.1126\n",
      "Epoch 997/1000\n",
      "20/20 - 0s - loss: -1.1071e-01 - MI: 0.1098\n",
      "Epoch 998/1000\n",
      "20/20 - 0s - loss: -1.1370e-01 - MI: 0.1123\n",
      "Epoch 999/1000\n",
      "20/20 - 0s - loss: -1.0675e-01 - MI: 0.1067\n",
      "Epoch 1000/1000\n",
      "20/20 - 0s - loss: -1.0926e-01 - MI: 0.1075\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f60f027fe20>]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "# GA numerical bias\n",
    "from ifn.Architectures.utils import build_gIFN_DNN\n",
    "from ifn.Architectures.ifn import IFN, mine_loss, MI\n",
    "from ifn.Architectures.dnn import DNN\n",
    "from ifn.Architectures.ifn import IFN, gIFN\n",
    "\n",
    "\n",
    "x_dim = 1\n",
    "y_dim = 1\n",
    "A_layers = [16,16,16]\n",
    "B_layers = [1,]\n",
    "C_layers = [1,]\n",
    "D_layers = [1,]\n",
    "l2_reg = 1e-9\n",
    "d_multiplier = 0.0\n",
    "d_l1_reg = 1e-3\n",
    "\n",
    "\n",
    "X, Y = np.expand_dims(gauss_reco_prior,1), np.expand_dims(gauss_truth_prior,1)\n",
    "\n",
    "# # Build model\n",
    "# ifn = build_gIFN_DNN(1, 1, [16,16,16], l2_reg = 1e-9, acts = 'relu', d_l1_reg= 1e-3, d_multiplier=0)\n",
    "\n",
    "model_A = DNN(x_dim, A_layers, 1, l2_regs = l2_reg, acts = \"relu\")\n",
    "model_B = DNN(x_dim, B_layers, y_dim, l2_regs = l2_reg, acts = \"linear\")\n",
    "model_C = DNN([x_dim, y_dim], C_layers, [y_dim, y_dim], symmetrize=True, l2_regs = l2_reg, acts = \"relu\")\n",
    "model_D = DNN(x_dim, D_layers, y_dim, l2_regs = l2_reg, acts = \"relu\")\n",
    "\n",
    "ifn = gIFN(model_A, model_B, model_C, model_D, d_multiplier= d_multiplier, d_l1_reg = d_l1_reg, y_dim = y_dim)\n",
    "\n",
    "\n",
    "ifn.pre_train([X, Y], epochs = 5, batch_size = 64, verbose = True)\n",
    "opt = tf.keras.optimizers.Adam(lr = 1e-3)\n",
    "ifn.compile(loss=mine_loss, optimizer=opt, metrics = [MI,])\n",
    "\n",
    "\n",
    "# Fit\n",
    "history = ifn.fit([X, Y],\n",
    "            batch_size= 1024,\n",
    "            epochs = 1000,\n",
    "            shuffle=True, verbose= 2)\n",
    "\n",
    "plt.plot(history.history[\"MI\"][5:])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "GA_meanvals =  np.squeeze(ifn.maximum_likelihood(np.expand_dims(np.linspace(-5,5,20), 1)))\n",
    "\n",
    "\n",
    "fig = plt.figure(figsize=(6, 6))\n",
    "gs = gridspec.GridSpec(1, 1) \n",
    "plt.subplots_adjust(hspace = .001)\n",
    "\n",
    "ax0 = plt.subplot(gs[0])\n",
    "ax0.yaxis.set_ticks_position('both')\n",
    "ax0.xaxis.set_ticks_position('both')\n",
    "ax0.tick_params(direction=\"in\",which=\"both\")\n",
    "plt.xticks(fontsize=20)\n",
    "plt.yticks(fontsize=20)\n",
    "ax0.minorticks_on()\n",
    "\n",
    "plt.plot(np.linspace(-5,5,20),meanvals,marker='o',ls=\"\",label=\"Polyfit Numerical Bias\")\n",
    "plt.plot(np.linspace(-5,5,20),GA_meanvals,marker='^',ls=\"\",label=\"GA Numerical Bias\", color = \"black\")\n",
    "plt.plot(np.linspace(-5,5,20),meanvals_analytic,ls=\":\",label=\"Analytic Bias\")\n",
    "plt.plot(np.linspace(-5,5,20),np.linspace(-5,5,20),ls=\"--\",label=\"Unbiased Calibration\")\n",
    "\n",
    "plt.xlim([-5,5])\n",
    "plt.ylim([-5,5])\n",
    "plt.xlabel(\"True $X$\",fontsize=20)\n",
    "plt.ylabel(\"Calibrated $X$\",fontsize=20)\n",
    "plt.legend(frameon=False,fontsize=15,ncol=1)\n",
    "_=plt.title(\"Gaussian Example\",loc=\"right\",fontsize=20, fontstyle=\"italic\")\n",
    "_=plt.text(-0.5,-4,r\"$(\\mu,\\sigma,\\epsilon)=(0,1,2)$\",fontsize=20)\n",
    "plt.savefig(\"Gaussian_prior2.pdf\",bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(6, 6))\n",
    "gs = gridspec.GridSpec(1, 1) \n",
    "plt.subplots_adjust(hspace = .001)\n",
    "\n",
    "ax0 = plt.subplot(gs[0])\n",
    "ax0.yaxis.set_ticks_position('both')\n",
    "ax0.xaxis.set_ticks_position('both')\n",
    "ax0.tick_params(direction=\"in\",which=\"both\")\n",
    "plt.xticks(fontsize=20)\n",
    "plt.yticks(fontsize=20)\n",
    "ax0.minorticks_on()\n",
    "\n",
    "plt.plot(np.linspace(-5,5,20),np.linspace(-5,5,20),ls=\"--\",label=\"Unbiased Calibration\",color=\"tab:green\")\n",
    "markers = [\"^\",\"o\",\"s\",\"d\",\"v\"]\n",
    "myc=-1\n",
    "for epsilon in np.linspace(0.01,2,5):\n",
    "    myc+=1\n",
    "    gauss_truth_prior_B = np.random.normal(0,1,10000)\n",
    "    gauss_reco_prior_B = np.random.normal(gauss_truth_prior_B,epsilon,10000)\n",
    "    mycalibration_B = np.polyfit(gauss_reco_prior_B, gauss_truth_prior_B, 1)\n",
    "    mypoly_B = np.poly1d(mycalibration_B)\n",
    "\n",
    "    meanvals_B = np.array([np.mean(mypoly_B(np.random.normal(xx,2,10000))) for xx in np.linspace(-5,5,20)])\n",
    "    meanvals_analytic_B = b(np.linspace(-5,5,20),1,epsilon,0)\n",
    "\n",
    "    plt.plot(np.linspace(-5,5,20),meanvals_B,marker=markers[myc],ls=\"\",color=\"tab:blue\",label=\"$\\epsilon=%0.1f$\" % epsilon)\n",
    "    plt.plot(np.linspace(-5,5,20),meanvals_analytic_B,ls=\":\",color=\"tab:red\")\n",
    "\n",
    "plt.xlim([-5,5])\n",
    "plt.ylim([-5,5])\n",
    "plt.xlabel(\"True $X$\",fontsize=20)\n",
    "plt.ylabel(\"Calibrated $X$\",fontsize=20)\n",
    "plt.legend(frameon=False,fontsize=15,ncol=1)\n",
    "_=plt.title(\"Gaussian Example\",loc=\"right\",fontsize=20, fontstyle=\"italic\")\n",
    "_=plt.text(-0.5,-4,r\"$(\\mu,\\sigma)=(0,1)$\",fontsize=20)\n",
    "plt.savefig(\"Gaussian_prior3.pdf\",bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(6, 6))\n",
    "gs = gridspec.GridSpec(1, 1) \n",
    "plt.subplots_adjust(hspace = .001)\n",
    "\n",
    "ax0 = plt.subplot(gs[0])\n",
    "ax0.yaxis.set_ticks_position('both')\n",
    "ax0.xaxis.set_ticks_position('both')\n",
    "ax0.tick_params(direction=\"in\",which=\"both\")\n",
    "plt.xticks(fontsize=20)\n",
    "plt.yticks(fontsize=20)\n",
    "ax0.minorticks_on()\n",
    "\n",
    "plt.plot(np.linspace(-5,5,20),np.linspace(-5,5,20),ls=\"--\",label=\"Unbiased Calibration\",color=\"tab:green\")\n",
    "markers = [\"v\",\"o\",\"s\",\"d\",\"^\"]\n",
    "myc=-1\n",
    "for epsilon in np.linspace(-2,2,5):\n",
    "    myc+=1\n",
    "    gauss_truth_prior_B = np.random.normal(epsilon,1,10000)\n",
    "    gauss_reco_prior_B = np.random.normal(gauss_truth_prior_B,2,10000)\n",
    "    mycalibration_B = np.polyfit(gauss_reco_prior_B, gauss_truth_prior_B, 1)\n",
    "    mypoly_B = np.poly1d(mycalibration_B)\n",
    "\n",
    "    meanvals_B = np.array([np.mean(mypoly_B(np.random.normal(xx,2,10000))) for xx in np.linspace(-5,5,20)])\n",
    "    meanvals_analytic_B = b(np.linspace(-5,5,20),1,2,epsilon)\n",
    "\n",
    "    plt.plot(np.linspace(-5,5,20),meanvals_B,marker=markers[myc],ls=\"\",color=\"tab:blue\",label=\"$\\mu=%0.1f$\" % epsilon)\n",
    "    plt.plot(np.linspace(-5,5,20),meanvals_analytic_B,ls=\":\",color=\"tab:red\")\n",
    "\n",
    "plt.xlim([-5,5])\n",
    "plt.ylim([-5,5])\n",
    "plt.xlabel(\"True $X$\",fontsize=20)\n",
    "plt.ylabel(\"Calibrated $X$\",fontsize=20)\n",
    "plt.legend(frameon=True,fontsize=15,ncol=2,loc=\"lower right\",handlelength=1)\n",
    "_=plt.title(\"Gaussian Example\",loc=\"right\",fontsize=20, fontstyle=\"italic\")\n",
    "_=plt.text(-4,4,r\"$(\\sigma,\\epsilon)=(1,2)$\",fontsize=20)\n",
    "plt.savefig(\"Gaussian_prior4.pdf\",bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Gaussian, OT example"
   ]
  }
 ],
 "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": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}