model_FFN.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import json\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import tensorflow as tf\n",
    "\n",
    "from src.utils import eval_RtnFcst\n",
    "from src.utils import deco_print\n",
    "from src.utils import load_dataframe\n",
    "from src.utils import sharpe\n",
    "from src.utils import max_1_month_loss\n",
    "from src.utils import max_drawdown\n",
    "from src.utils import construct_decile_portfolios\n",
    "from src.utils import plot_decile_portfolios\n",
    "from src.utils import plotReturnDecile\n",
    "from src.utils import calculateTurnover_with_dl\n",
    "from src.data.data_layer import DataInRamInputLayer\n",
    "from src.model.model_utils import calculateStatistics\n",
    "from src.model.model_utils import getFactorWithLargeCap\n",
    "from src.model.model_RtnFcst import FeedForwardModelWithNA_Return_Ensembled"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Model Configurations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "logdirs = ['model_char_shift/model_FFN/%d' %k for k in range(1,10)]\n",
    "with open(os.path.join('model_char_shift/model_FFN/config.json'), 'r') as file:\n",
    "    config = json.load(file)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Load Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "dl_train = DataInRamInputLayer(config['individual_feature_file'])\n",
    "dl_valid = DataInRamInputLayer(config['individual_feature_file_valid'])\n",
    "dl_test = DataInRamInputLayer(config['individual_feature_file_test'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "tmp = np.load('datasets/char_all_na_rm_all_50/TSplit/CharAll_na_rm_all_50_train.npz')\n",
    "value = tmp['data'][:,:,19]\n",
    "tmp_valid = np.load('datasets/char_all_na_rm_all_50/TSplit/CharAll_na_rm_all_50_valid.npz')\n",
    "value_valid = tmp_valid['data'][:,:,19]\n",
    "tmp_test = np.load('datasets/char_all_na_rm_all_50/TSplit/CharAll_na_rm_all_50_test.npz')\n",
    "value_test = tmp_test['data'][:,:,19]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "### Large Cap Stocks\n",
    "tmp = np.load('datasets/char_all_na_rm_all_50/TSplit/mask_LME.npz')\n",
    "mask_train = tmp['train']\n",
    "mask_valid = tmp['valid']\n",
    "mask_test = tmp['test']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "### Large Cap Stocks 0.001%\n",
    "tmp = np.load('datasets/char_all_na_rm_all_50/TSplit/mask_LME_0001.npz')\n",
    "mask_train_0001 = tmp['train']\n",
    "mask_valid_0001 = tmp['valid']\n",
    "mask_test_0001 = tmp['test']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "### Risk-free rate\n",
    "df_tmp = load_dataframe('datasets/F-F_Research_Data_5_Factors_2x3.CSV', skiprows=2, nrows=659)\n",
    "df_train = df_tmp.loc['196701':'198612']\n",
    "df_valid = df_tmp.loc['198701':'199112']\n",
    "df_test = df_tmp.loc['199201':'201612']\n",
    "rf_train = df_train.loc[:,'RF'].values / 100\n",
    "rf_valid = df_valid.loc[:,'RF'].values / 100\n",
    "rf_test = df_test.loc[:,'RF'].values / 100"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Load Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "tf.reset_default_graph()\n",
    "global_step = tf.train.get_or_create_global_step()\n",
    "model = FeedForwardModelWithNA_Return_Ensembled(logdirs, config, 'test', global_step=global_step)\n",
    "gpu_options = tf.GPUOptions(allow_growth=True)\n",
    "sess_config = tf.ConfigProto(gpu_options=gpu_options)\n",
    "sess = tf.Session(config=sess_config)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Variable Importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA0AAAAIECAYAAAAw3U6MAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAIABJREFUeJzs3XuYHUWZ+PFvSUDAgEEFowyXsEIFRERERFbWSLwhGIVVgYNIVEQUMSIiXtjttILiimj8CcsiQgA5IqBCWBdEAsmKiMEbugItlwkhyCVcAglJgED//qgeOAwzySSZOZec7+d5+jmnq6u6q5M8POelqt4KZVkiSZIkSd3gBa3ugCRJkiQ1iwGQJEmSpK5hACRJkiSpaxgASZIkSeoaBkCSJEmSusaoVneglWKMLwTeANwDPNXi7kiSJDXTOsArgBuKoni81Z1plhjjS4CNW/DoR4uieKgFz1U/XR0AkYKfX7e6E5IkSS20J3BtqzvRDDHGl7x4/acefGTZOq14/MMxxlcZBLVetwdA9wCcf/75jB07ttV9kSRJapp7772Xgw8+GKrfQ11i40eWrcP5H5nH2I2XN+2h9z46ioPP3nIT0siTAVCLdXsA9BTA2LFj6enpaXVfJEmSWqHrlgGM3Xg5PWOaFwCpvZgEQZIkSVLXMACSJEmS1DUMgCRJkiR1DQMgSZIkSV3DAEiSJElS1zAAkiRJktQ1DIAkSZIkdQ0DIEmSJEldwwBIkiRJUtcwAJIkSZLUNQyAJEmSJHUNAyBJkiRJXcMASJIkSVLXMACSJEmS1DUMgCRJkiR1DQMgSZIkSV3DAEiSJElS1zAAkiRJktQ1DIAkSZIkdY1Rre5AWxg3DpYvb3UvJElStyvLVvdAWuuNzAhQCIEQTiaEBYSwhBCuJ4TXE0I5wDF9BfeZ1VBvOSHcQwifqa5NHeBe1xLC0YTwNCFs0XCfiwjhlhF5V0mSJEkdY6SmwL0DOAY4BNgWeASYDmxSHQCfq75/aiX3+llV7+XAVcC3CWHj6tr9DffcBHgXUAeeBg4EIIQXAfsA563xW0mSJEnqaCM1Be6R6vNY4ELgCGAeZfkUACEALKUsFw7hXk8+Uy+EB0jBzVPVtXKAeywmhCuBGvAt4D3A+hgASZIkSV1vZEaAyvJ64HBgC+B04Hbgq6t5t/0JYSEhLAU+AHyCsnysurZZda3v+GJVfi6wMyGMBw4AZlOW81b7fSRJkiStFUZqDdBOwALKcjtgK+As4MuEMHY17vZLYCJwG3Av8IuGaw8COzccp1fllwKPAp8gTYtz9EeSJEnSiK0B2gW4mBDeBzxBCkaWAo+tsNXAHqMs/wDsD4wnrSXq8zSwsOF4GoCyXApcDBwFlMBFq/MSkiRJktYuI7UG6BwgAqcBLyNNgTuIsly02ncsy1sJYQpwJiEcWpVuBjzcUOs+oG+U6Vzgo8Ala/RcSZIkSWuNkQmAyrIEvlQdA10PQ7zPhH7nPwR+2FAydQVtZwNDe05vL/T0DKmqJEmSpM41UlPghi6EXfslMug7Pt/qrkmSJElau4zUFLhV8VdSAoP+Hh6gbGTMGAdjljftcZK0SmruDC9J0nBpfQBUlo8Dcwe9Xg+BtJ/PocCLgL8AR5ISHIynVu5e1ZsMnA1sDPQCl1ArD6uuvRr4P6BGrfzxiLyHJEmSpLbX+ilwK/cO4BjgEGBb0iarZ62g/pPABcD+1MN6VdkBpEx0l4xgPyVJkiS1uU4IgB6pPo8F9gWOIKXZXpFzgU2AvavzDwIXUyuXjkgPJUmSJHWE9g+AauX1wOHAFqSNTm8Hvkra32cgJbVyDlAAB1EPryWl5D63Cb2VJEmS1MZavwZoZephJ2ABtXI76mFL4N+BL5M2RG3s/7qkoOiJ6vy8qt59wJ3A/zary5IkSdKKxBi/DbyetIflhsAdwIKiKD6wGve6tyiKsTHGLwJXF0UxZzX7tDVpvf0fG4qvLoriq/3qXQB8uCiKJ+gnxrgl8NqiKC6LMX4XOKUoinmr05+R0v4BUJrudib18H7getJanqWkv5iDqYe3AzcB7wH+Sq3sGxk6D/gaKWHCSQ3lkiRJUksVRXEMQIxxMjC+KIovDsM9T1rTewA3FUUxYSXPOXAFl/cCxgOXFUXx2WHoz7DrhADoHNIUttOAl5GmwB0EXAHsCFxEyg73V+Bjz7SqlfOoh9nABJz+JkmSpA4QY5wAfJM0q+kM0v/4P5JnZzvtR9ou5gzg1aTfxi+s2k4nJQMbC7ybNLL0T8A3i6KYHmPcDTgVWATcDywrimLyavTpa6QgZwvgTGA9YAlQA74IbBhjvA74HGn9/r3Aj0jZmkcBxxdFcXWM8S/AbGCn6t3eW93rJ6SlOusDRxRF8edV+CNcqfYPgNLIzZeqo79PVMdgbd86pGdM6oWentXpnSRJkjTc1i+K4o0AMcYvA/sURbEkxvhfwDuBx6s6u1dTzt4/wD1eXBTFO2OM2wKXkZaPnA4cUhTF32KMJwKbD9BuhxjjrIbzgwfo09eqspOBbxRFcUWMcRLwWuAk0ojWjBjj56p6xwO/KopiWoxxc+DaGOM2pIDox0VRHBVjPJ+UwGwR8CDwYWAH0kDHsGr/AEiSJEnqLkXD9/uBc2KMi0mjLr8FtgLmABRFMS/GeNcA9+gbNbmLNJIC8MqiKP5Wff81MNBUtudNgauCqGKAurHqD0VRzKjqTh6g3vbA+VW9u2OMjwKbVdf+1K+fF5K2vrmUtL3NCQPcb420fxa4EE4hhHsbzj9ICL2EsIgQfkII66+g7QRCuIkQHiOEXxLCpgPWGzcOQvDw8BiJQ5IkraqnAWKMLwZyUqByGGk6XCCtf39TVeeVDDySM9D697tijDtU33dfnT71czPwhqofB8cYj6rq9Y8xbgb2rOptTtqu5sFB+jkBuKcoineQgp+vr2I/V6p9A6AQXkoI04GjG8peCJxdHW8m7Qt02ArucibwG+B1pIj5uBHqrSRJkjTcHiX9lv0tacRmKfBK0ujIgzHG3wHfBR4Y4v0+BZwVY7wK2I00wrImjgW+VE2ZO5g0yvNX4L0xxsbRpa8De8UY/xe4BDi8KIrlg9zzRuCw6p7fAr6xhn18nlC2U3K0ECaTFmb9GdiDlMntEeADlOVYQngBacjvftICqbuBL1GW0wa535akeYSLSQvEfkpZPhNQVan+emfecQc9ywf7O5C0RtrpvzGSpGfMnz+fiRMnAowrimJui7vTFM/89ptyBz1jmvfbb/7CUUyctg20+M86xngkcGFRFAtijCcAT/RPcd0N2nEEaEPg+8DWlOWHeXZ4DMryacqyl9TvS0nZJs4f9E5lOY8U/JxPmmd4+oj1WpIkSWpv9wFXxhh/DexMGnjoOu2aBOFyynLhgFdCeDEpBfb2wLsoy8GH/EJYj7SQam/gg5TlQIu3JEmSpLVeURQXAxe3uh+t1o4jQADLVnDtfFLEuj9wCytKggCnAJOAycBsQthw2HooSZIkqeO0awA0sBB2BvYhpcibSdoEauogdV9K2ngpAPWq7lnN6KYkSZKk9tReU+DKcjppk6bGsqn0BTll+WdSQDOUez3IUN+v141QJUmSpG7QWSNAAwlhM0JYOMDxvVZ3TZIkSVJ7aa8RoNXzIGlNUH+LhnyHGeOgiakQpbZXM3W1JElaO7V/AFQPE4BrgO2plbf0Kys5n7HUyrnUwy+AdwPnUCsnUw9zqYet+t3t49TKM5vXeUmSJEntpNOnwD0OvJV6GAX8C8/PHvddYJOG49zmdk+SJElSO2n/EaAV+x2wF3BXdf63ftcfpzbIfkKSJEmSuk6nB0CzgRopAPoNMKbf9c9SD0c0nO9ErZzXrM5JkiRJai+dPgVuNvAq4ODqe39nkxIk9B3/aF7XJEmSJLWbThoB2pJ66EvVtnn1+Q/gNmA8KQB6b782y4DGKXCjANO9SZIkSV2qk0aAfgncWh0fbyifDSwBbhigzWeBhxuOqSPbRUmSJEntrP1HgGrlLCCsoMZh1QGwe0O7rYf8jEm90NOz6n2TJEmS1FE6aQRIkiRJktZI+48ANcO4cbDcpUHqImXZ6h5IkiS1RGeNAIWwOSHMIoQlhPBXQthtCG0+QwglIazfhB5KkiRJamOdFQBBTtrr59VAL3DaoDVD2IAQTga+05yuSZIkSWp37R8AhTCZEB4jhN8AE4F9KMteUua3dVbQ8hXAdsDXmtBLSZIkSR2g/QOgZEPg+8AEyvJuQpgCHABMG7RFWd5BWU4C5jalh5IkSZLaXiclQbicslxICMcBJwH/QVlOb3GfJEmSJHWQThkBAlhGCO8nBT/fB75BCBu3uE+SJEmSOkgnBUAAx1efnwYeBv7ewr5IkiRJ6jDtPwUuTXObXp3tvIbtB9bbCz09q3xrSZIkSZ2l00aAni+EiwlhYb/jgVZ3S5IkSVL7WbURoHqYAFwDbE+tvKUqmwW85Xl1a2VY084N0ZHABv3KVm2b+xnjYMzyYeuQtMZqq/ZPWJIkSUMzHFPg9q3ucxqwLfD2Ybjn0JXlfU19niRJkqSOteYBUK1cDEA9PAE8Ra1c+My1eiiBT1IrT6ceLgDGUisnUA9zgbuBHYGPAt8GrgV2AV4OfIFa+UPq4Q2kwGpHoACmAL3AHcD7qZWXUA/fBN4HjCclR/hK9fSTqZUnUw9TgY8Bi4GCWvm+NX5nSZIkSR2plWuARgG7ApdX5zsA7wX+AHyeelgH+Dkp2NkWmAVcAiwEfgFMph5GAR8GTge2J22MOoU0KnUC9bBLde8e4HBSgCRJkiSpSzUjAOpbC7ROv/LfUitvpVYuqc5nUitvBf5EWtOzKbA5cAG1cj5wLjCGFAydCuwDHAq8mJTlbcfqWWcAVwHrAXtU936EWvnr6j6SJEmSutTqToHbknroyxqwlFp59yD1lgCRengZ8DqgMQBZ1q/uU9Vn3+rvBcA9wIHUww2kkZ6FpL1/FpOmwn2HFCA9TD3cUrU7HpgHHESaVve+AZ4lSZIkqQut7gjQL4Fbq+P8FdTLSYHL5azqpqW18ilgP2Cb6jkTgPdRKxdRK0vgP4GNSGuEoFb+BTgaOA6okwKqm1fpmZIkSZLWaqEsuzfdboxxa6B35pQ76DENttqJabAlSSNs/vz5TJw4EWBcURRzW9ydpmjVb7/5C0cxcdo20EV/1u1sONJgd75JvdDT0+peSJIkSRphrcwCJ0mSJElN1d4jQPUwAbimoWQJcDFwGLXyyQHqrw8cCXxvwOuDmTEOnAKnduIUOEmSpBHR3gHQs3YjJULYnZRQ4WfApQPUOwA4mZQme+gBkCRJkrrGtCmfZfTy0U173uJRi2GbGU17nlasUwKgRdTKhdTDQ8+c18NRwFeq85OplSeT9gMCWAoE6uE04EPAuqTA6QNVdjlJkiRJXahTAqA51ENJSnv9a+ABYBppr5/bgWuph6tJ099OBcZSD2OAx4B3kDZJ/QFpE9Vbnn97SZIkSd2gUwKg/YA7gM2BnwMXAAE4g7Rx6nrAHqQNUgEeqa6PBr5J2kAVYP3mdVmSJElSu+mUAOhR4GHSVLYnSaM42wPHA/NII0HXAq+p6m9DGu05grRu6E3AJFJQJEmSJKlLdUoANKf6fAr4PXAs8L/AccAmwCXAzaQscf8AZgFvAP5YfZ8NPEEKjP7UvG5LkiRJaiftHQDVylkMPmrz3epo9HfSNLk+rx/Sc9wIVZIkSeoKboQqSZIkqWu09whQs7gRqlrNjU8lSZKaYngCoHqYAFzTULIUuAE4iFr5jwHqTwbOBjYgJSs4C3g1cCPwYWrlrVXa609SK0+v2lwAjKVWTqAeZgH3UisPXEGfVl5HkiRJUlcZ7ilwu5GSEuxJyrw2ZQhtvkPa12dHYB3gG0Nosy/w0dXsoyRJkqQuNdxT4BZRKxdSD3eTMrYtpx6OAr5SXT+ZWnlyvzYfAp6mVt5PPTxFCoJW5r+Be4EDqYePAl8n7QH0e2ACtXJsVW8r6mEOaZTpG9TK/1ijt5MkSZLU0YY7AJpTTV17IXA18EtSGuqDgNuBa6mHq5/TolbeC0A9nAK8EZjYcPUU6uGk6vuGwHXPaVsPmwL/CZwMnAn8rF9/eoC3A0cD/w4YAEmSJEldbLinwO0PnFDd91JgLCmN9RnAVcB6wB7PaVEPgXo4Ffgs8GlqZeNaohOBnavjlwM8b5vqnhdRK3uBK/td/y218hbS3j8vWqM3kyRJktTxhjsAmk+t/BYpwcGppGlqAMcDhwIXAtf2a3MM8CnSpqZ16mF0w7UHqZVzqZVzgccGeN4dwJPAv1IP2wDv7Hf96TV4F0mSJElrmZHaB+ho4Dbgv4Av0hfcpHVBN/d7/per7/8BPEwaKRqaWrkA+CTwceB/gHsw6JEkSZI0iOFZA1QrZ5GmuvWdLwHGN9T4Zr8W06sD4CWD3DP0Oz+w4fsEAOphPeCtwFTg56R1QHc/p076fjpw+qD9n9QLPT2DXpYkSZK0dhipEaDmqJVPkEaUpgLzSdnejmlllyRJkiS1r+HOAtd8tfJEUrKE1TZt3DRGLx+98orSMMrKrNVdkCRJ6jptFwDlId8W+C4pW9wy4CLSGqIl/aremZXZ1nnIjyKtI9oAOB/4NCnhwtlVvadIG63+e1ZmZ4z8G0iSJElqV201BS4P+YtISRCeAF4HTAY+BvxbVeVzwCbVsVMe8i2AaaQAaH9SNrl9Gm45FtiClHnu+3nIHeaRJEmSuli7jQC9G9gSeGtWZnOBuXnIdwTmAV8ClmZltrCvch7ypUAE5gL/VBWv03C/R4DHgftJwV7jNUmSJEldpt0CoHHV5/y+gqzMegHykAOckof8pOrSUVmZnQfcmof8FcDPgL8DVwAHVXXuBdYHSuDYrMweGfE3kCRJktS22i0A6gt8tgBuB8hDfiDwf1X5iaR1PpDW9ZCHvAe4mrQGaK+szB6vgiWAN5I2UH0gK7NlI957SZIkSW2t3QKgXwD/AL6Zh/zzwM7AuaTpbwAPVlPjAMhDPgq4jLSX0ERgQR7y9Rrud6eBjyRJkqQ+bZUEoZqi9nZgY+AvwH8CpwDfGaTJvqQg6aXAn4GHgY+OfE8lSZIkdaJ2GwEiK7ObgHcMcCkMUPeSgcor04f6zCm9U+jp6RlqdUmSJEkdqq1GgCRJkiRpJK36CFA9TACuaShZCtwAHESt/Mcq3GcucAG18ov9yi8AxlIrJwzS7kZgAbXybdX5JsB9QEat/MaQn99oxjgYs3y1mkqrrVa2ugeSJEldZ01GgHYjbUi6J/AmYMoqtt8JmLoazz0PmEA9jK3O9yPt7/Oj1biXJEmSpC6yJmuAFlErF1IPdwNPAcuph6OAr1TXT6ZWnkw97A18H+gh7dPzYWrln0hJDi4Avkg9ZKQAah7wJCl19WDOB04CDgCmAR8EZlEr76Ie9gFOBTYH7gAOJCVG6CWNWu0KRGrlPWvw3pIkSZI61JqMAM2hHh4B5pKCi1+SApIppOxsJ1APuwCHA3eR9uQ5B9jsOXeph9eRRoK+SNrAdPMVPjUFL1cBB1EPLyWlvz63uroZKTjaHhhDCpL6/A3YxeBHkiRJ6l5rEgDtD5xQ3eNSYCwpI9sZpABlPWAP4N+ARcBsYDIpxXWjbavPn1ArbwbmDOHZ55ECqs8DTwA/rcqXAQcD3wJKYP2GNjOplbcN+e0kSZIkrXXWJACaT638FnA2adrZvVX58cChwIXAtaRpaP8HvBq4nWenyPW5ufo8mHrYnhTYrMzPSUHVscDPqZWLq/LTSAkZppLerTFFthuiSpIkSV1uONJgHw3cBvwXaRrbcUCdtC7oZmAW8F5S8LMD8O/PaV0r/wp8mTSaVAf+tNIn1solpFGfdXh2+hvAmcAnqmt3A9us5jtJkiRJWguFsuzeVLwxxq2B3pkzZ7oRqiRJ6irz589n4sSJAOOKopjb4u40Rd9vv0l3TGL08tFNe+7iUYuZsc0M6KI/63a2JlngRlY9XAy8rV/pcmrly1rRHUmSJEmdr30DIDgS2KBf2cgMV7kRqprBjU8lSVIlxjgbyIuiuLqhbBrw16IozuxXd2vggqIodm9uL9dOzQ2A6mEyKWnCBtTKZVXZu4DLgXGkLHFHUCvHUivvG6D9ZtTDk0BOrTyhKtsH+G9gD2rlbwd57izgXmrlgcP6PpIkSdLq+QHwYeBqgBjjesB7SGvjNYKGIwnCcDoJ2G7Qq7XyftJ+Qwc1lB4A3Dpo8CNJkiS1n4uBvWKMG1bn7wWuBHaNMV4dY5wVY/x9jPE5v41jjHNjjOtX30+KMU6uvn8jxnhtjPG3McYPNPE9Ok6rAqCdqYebqYcHgMMayr8I/B2AengV9XA19bCUeiioh3+t6pwL7EA9vJZ6eCEwibQvENTD3tTDTdTDEurhWurhNU17I0mSJGmIiqJYBlwC7FcVfYSUVfnVwIeKopgA/AxYaTATY9yblGDhzcBbga/EGMeMRL/XBq0KgM4B7gReR9rPZyDnkabojSelt/4x9fBPwAzgEdIo0LtIG6ueRz1sQoqkryCNIt1D+kclSZIktaMfAIfEGDcHNimK4k+krVy+F2OcTgpm1l1B+749L18DvD7GOIv0W3hdYOsR6nPHa1UAtC1wObXyLuDSQeq8jrTJ6Z3AD0l/kTtXa4cuIm2wegBwLbVyLhCBDYHzqJXzgR8D21APLx3RN5EkSZJWQ1EUfwU2Aj4DnFUV/wD4SFEUk4F/8GyQ02cZ8IoYYwB2rspuAa6pRo32Ai4k7cGpAbQqAPoj8B7qYUueHfbr70bgfdTDVsDHgCd4dpPUc4GtgA/y7Eaofyf9gziEeughjRDdRq18cGReQZIkSVpjZwEfJ/3Pe4AfAb+OMf6GFBy9sl/9/wD+pzoersouAxbHGH8N/AEoi6IYbJZV12tVGuzPAmeQ/oJ+OkidQ4DTSRHtXUCNWnlHde1aoBd4BWk0CGrlQ9TD+0n/KI4gBUv7j1D/JUmSpDVWFMUPSbOd+s4/N0jV3avrZ/HsaFGjwdqpn+YGQLVyOjC9Otuh4coR1efU6oBa+XfSEN5A9ymBbQYo/wXwiwHKJ6ywX5N6oadnhVUkSZIkdb52S4MtSZIkSSOmVVPg2sq0adMYPXp0q7uhtVyWZa3ugiRJUtdrqwAoz/MJwDX9is8h7Q90CvA20qjV/wBHZ1n2YJ7nR5F2zN0AOJ+UReO3wIuB12RZ9kSe558GpgF7ZFn2u2a8iyRJkqT201YBUIPdgFur708As4AngT2BMaTU2f8vz/PjSIHNx0j7Cs0ELq/O/wAck+f5D4CvAt8z+JEkSZK6W7uuAVqUZdnCLMsWknbDfQPwb1mWFVUQsxswGbiXtP/Pj0h50gHWybLsr8A3gONJWTUWVt8lSZIkdbF2HQGak+f509X3b1Sf8/suZlk2r6HurXmevwL4GWkvoCuq8hOBfwUmAe/Isuyxke2yJEmSpHbXriNA+5F2tt2ZtJ4HYIu+i3mevyfP892r7z3AbNJGUftmWfY4QJZlTwAXA/dlWfarJvZdkiRJUptq1xGgu7MsmwuQ5/l84EZgap7ndwEvB84FfpTn+e9JO9++BJgILMjzfL0q+JEkSZKk52jXEaBnZFm2HNgXuA+4HriENLLzhap8Z+ClwJ+Bh4GPtqankiRJktpdKMuy1X1omRjj1kDvzJkz6enpaXV3JEmSmmb+/PlMnDgRYFxRFHNb3J2m6PvtN+mOSYxe3rw9IBePWsyMbWZAF/1Zt7O2HwGSJEmSpOEyfGuA6uFGYAG18m3V+SakaWsZtfIbK2pa1d8a6AX2plZeseLKw2zGOBizvKmPVBeqde9oqyRJUrsYzhGg84AJ1MPY6nw/YB3SHj1DMQ/YBLhqGPskSZIkSc8Yzixw5wMnAQcA04APArOolXdRD/sApwKbA3cAB5ISFvQC1wC7Au8ErgP2ph7mkBId7AEsAU6mVn6depgOvKoq2x24gFp5OPWwJXAW8CbShqifolb+ino4CvhK1b+TqZUnD+P7SpIkSeowwzcCVCvvIY3eHEQ9vJSUlvrc6upmpOBoe2AMKUjq8zdgF+CehrItgN8D40lZ3z7TcO21wL+RgqyPUw8vB74NvBh4NfBT4O3Uww5VnSmkbHEnUA+7DNfrSpIkSeo8w70P0HmkKW+fB54gBSMAy0jpqd8JlMD6DW1mUitvq9YA9XkE2AE4A3hRv/q3UCt/Rz28pDrfgBT4/JJaORf4IgD18EEgVPcogfVII0p/XPPXlCRJktSJhjsA+jmwCDiWND1tcVV+GnA2cA7wK1Jg0mfZAPeZQgpqJgInADs2XHuq+mxcUf5/wFuqIOoo0sjRl6prx5PWFx0EXLsa7yRJkiRpLTG8abBr5RLSqM86PDv9DeBM4BPVtbuBbVZyp58D65I2N30JsBH1sOkK6n8eWAjcRJru9l1q5V+Ao4HjgDopcLp5Fd9IkiRJ0lpkuEeAoFZ+BPhIv7JjSaNC/YWGOnN57sjQQDuTTm6of0W/+nsN0JfvAt9dSY8lSZIkdYnhD4A60aRe6Bko3pIkSZK0NhneKXCSJEmS1MaaOwJUDwH4FnAoKbvbX4AjqZV/GOHnjiet/3krtXLW867PGAdjlo9oF9RlauXK60iSJKnpmj0C9A7gGOAQYFtSuuuzmtwHSZIkSV2q2WuAHqk+jwUuBI4A5lEPs0jpsF8BbA78O7XytKp8NLA1cCJpD5/TgVeSNkj9BCkd9vnAu0mZ3s6lVh5JPWxXlW8LXN6Ed5MkSZLU5po7AlQrrwcOB7YgBTK3A1+trr4a+CDwHeB71MPmVfmmwJuA6cAFwJXATsAuwCeBscCdwOurtp+iHjYgTbVbF3gdcNsIv5kkSZKkDtDcAKgedgIWUCu3A7YiTX/7MimI+T21suDZfYT+qWr1J2rlraRgZixwGHAjEIF/BhaTRoROJQVKAC8kjfxcQ63sre4pSZIkqcs1ew3QLsDF1MP7gCeAR4GlwGPAm6iHHYD3Ast5dtRmWfX5IPAA8GNgEvAT4CLgQ9X5EaTACNL+QDcBb6MetgE+MLKvJUmSJKkTNDsAOoc0Ne00YB6wN3AQsAg5+AR5AAAgAElEQVS4ixTUHAt8hlr5j+e0rJVPAQeQRnl+SRohmkOaEnc38CdgfFV7G+ALwBLgD8BGI/hOkiRJkjpEc5Mg1MoS+FJ1PKsejgZup1Ye2K/+hH7nV5PWCvW33SBPfGPD988M2i83QpUkSZK6ghuhSpIkSeoazU6DPbD+Iz1NNm3aNEaPHt3KLqiDZFnW6i5IkiRpNbVHALQa8jyfAFxTnT5N2mNoGikBQgbclGXZq/M8Xwd4CNgY+EiWZdOb31tJkiRJ7WBtmAK3Gyk99o+AqaT1QI8D2+d5/nLgDcB6LeudJEmSpLaxNgRAi7IsW0DKBAcp2FkI/B3Yqzqub1HfJEmSJLWRjp0C12BOnuejSBulnkJKqf1mYDYp+BlXfZ/Qqg5KkiRJag9rwwjQfkAExmRZdkxD+WzgHcAe1XdJkiRJXW5tGAG6O8uyuwYonw1sSdoM9ffN7ZIkSZLa1RSm0EPz9oCcz3xmMKNpz9OKrQ0jQAPKsuxu4HbgOuDJFndHkiRJUhsIZVm2ug8tE2PcGuidOXMmPT3N+78AkiRJrTZ//nwmTpwIMK4oirkt7k5TPPPb746Z9Cxv4gjQqPlM3Ka7/qzb2Vo7AiRJkiRJ/a0Na4CgHj4BnA6cSK08viqbStoQtb+3UitnPadkxjgYs3xk+6i1Q617R0wlSZLWBmvLCNDhwIPAx6iHdauyk4BNgCOr87HV+bXN754kSZKkdtD5AVA97ArsAtSATUlpsaFWLqNWLiRlgQN4hFq5kFrpUI8kSZLUpTo/AEqjP9dRK68ELgc+2eL+SJIkSWpTnb0GqB42Ag4CNqAelgHrAKOoh+2plTe3tnOSJEmS2k2njwDVSEHcLsDOwGuAu4EjWtkpSZIkSe2p0wOgw4ErqZV/oVbeQq28BbgIOJR62LDFfZMkSZLUZjp7ClytfP0AZUcDRzecTwemN6tLkiRJktpXZwdAw2VSL/Q0bzdgSZIkSa3R6VPgJEmSJGnI2noEKM/zCcA1DUVLgRtImd/uBq4D3pxlWZnn+XRgfJZluze0/y3wRiBmWXbrYM+ZNm0ao0ePHv4X0Fohy7JWd0GSJEnDpFNGgHYDNgH2BN4ETKnK9wA+PlCDPM93AnYHHsKscJIkSZLonABoUZZlC0mjPk8By6vye4CT8jzfbIA2hwNzgG8Dk/M8X78pPZUkSZLUtjolAJqT5/kjwFzSlLjvVOWnAIsbzgHI83xD4EPAmcDZwMbAAc3qrCRJkqT21CkB0P7ACaT+Xppl2QNV+WLgSNKGqG9uqH8A8GLg+6SgaR3gk83qrCRJkqT21CkB0Pwsy75FGs05Nc/zZxIdZFl2GfAz4J8a6h8O/BR4LbAzcCjwxjzPX9e8LkuSJElqN50SAPU5GriNFAg1Ogp4FCDP89eQkh+cn2XZLVmW3QL8GHgA+FQT+ypJkiSpzYSyLFvdh5aJMW4N9M6cOZMeN0KVJEldZP78+UycOBFgXFEUc1vcnaZ45rffHTPpWd68337zR81n4jbd9WfdzjptBEiSJEmSVltbb4TaLG6EqsG4CaokSdLapa0DoDzPJ5DSXvdZCtwAHASUpPTXe5P2BroB+HyWZX9taP8J4HTgxCzLjm9StyVJkiS1qU6ZArcbsAmwJ/Am4BjgMuDVwNuAXYEHgVl5no9taHd4Vf6xPM/XbWqPJUmSJLWdTgmAFmVZthC4mzTaswvweuALWZbdkGXZHaRg58XAIQB5nu9a1asBmwL7taLjkiRJktpHpwRAc/I8f4S0qek1wA+r8t6+ClmWLSalut66KjocuC7LsiuBy3EjVEmSJKnrdUoAtD9wAqm/lwK3VuXb9lXI83wM8DLgzjzPNyKtE3pjnufLgHcBE/I8376pvZYkSZLUVto6CUKD+VmWXZXn+auAU4G3AL8H/iPP8/tJIz/fBBYDPyJNextFmgL3RHWPq4AjgClN7rskSZKkNtEpI0B9jgZuA84E3k/K/HYJ8AdgDLBXlmX/IE1/uzLLsr9kWXZLlmW3ABcBh+Z5vmFrui5JkiSp1UJZlq3uQ8s8sxvwzJn09DRvN2BJkqRWmz9/PhMnTgQYVxTF3BZ3pyme+e13x0x6ljfvt9/8UfOZuE13/Vm3s04bAZIkSZKk1dYpa4BG1rhxsHx5q3uhdtPFo6OSJElrq/YeAQrhFEK4t+H8g4TQSwiLCOEnhLD+CtqW/Y6rmtFlSZIkSe2rPQOgEF5KCNNJSQ/6yl4InF0dbwb2BQ5bwV02qY63AI8DXxmh3kqSJEnqEO0TAIUwmRAeI4TfkNJavwD4fkONJ4EdgW8D84AArDPo/cpyIWW5kJQe+/uU5e9GquuSJEmSOkP7BEDJhqSgZ2vK8sPAg89cKcunKctent0MdQlw/grvFsJbgJ2Bk0aov5IkSdJqiTFuHWO8fgXXD48xrjtMz3pXdb8BnxljnB5jfNdwPKvdtVsABHA5ZXnngFdCeDFwJbATsC9l+cBK7nUgcMUQ6kmSJEnt5susaMbTKiiK4oqiKM4Yjnt1unbMArdsBdfOJ43o7APcQgjrU5Yrqj8B+M9h7JskSZI0rGKMs4A/k5Z7bAx8AHgbMBa4IMb4r8B/AVsArwBmFEVxfIxxOmmt+9ZV+eSiKP4YY/wY8GngIeAJ4CfVo8YDpwObxhhnAC8H/rsoiq819GXdqs62pMGS44uimBVjPBF4Kyl++GlRFN8cmT+NkdeOI0ADC6Ev8FkfmAk8DExdSastgNtGtmOSJEnSGptTFMXbgF8BBxVF8UPgXtKMpi2A64uieCewG3BEQ7s7q/L/BxweY3wZcBzwz8A7gBcN8KzRwCHAHsDeMcbXNlw7DHigKIp/Ad4LnFqVHwzUgD2BhcPwvi3TPiNAZTkdmN6vbCp9QU5Z/pmU+GBV7jl6SPV6e6GnebsBS5IkSf38qfq8izTy0+gh4A0xxrcCjwIvHKTdPwOvAm4qimIJQIzxugGedWNRFI9U1+cA2zVcew2wZ4zxjdX5qCqoOpi0rn4scPmqv1776JwRoIGEsBkhLBzg+F6ruyZJkiStgoF2YH+a9Ht9MrCwKIqDSRmRN4wxhkHa3QaMjzFuEGN8AWnEqL/tY4yjY4yjgDcCf2u4dgvw46IoJgB7AxcBi0jT8g4iTYObHGPcatVfsT20zQhQHvJPkOYbnpiV2fEN5V8nDfM9DXwnK7MT85BPBTLIbsrId17Mi17wbT7/Zwgbbcp9x27Ggqv+FvKB/hFtm5XZ86bETRs3jdHLhzZYpO6RlVmruyBJkrrbr4H/AY4E6jHGN5HW/NwKvHKgBkVRPBBj/GbV9iFgA9J2Mo3Z5B4irQvaFPhJURQ3xRj7rv0X8IMY42zSeqTTiqJ4PMb4EHA9sJSUlGzecL5oM7VNAAQcTkp7/bE85HlWZk/mIX8z8CXSIrAtgLPzkF9a1X8cwvY5U5cC46j+Uhfw8gcW8PK+eYnvJ60X6vNoM15EkiRJWpmiKOYCu/crO73h+6ENlxrX6fSZ3FD3CuCKalTnlUVR7FqNEv0vcFdRFP+7onsVRTG54fTDA1z/KvDVFbxOx2iLACgP+a7ALsA7SVHufsCFwO+B8VmZFXnI+/KS96UCXFgde5ECoOtJWd8aPZaVWUcv0pIkSZKGqiiK5THGF8UY/0jKAPc70miQKm0RAJFGf67LyuzKPOSXA58ELszKbBlQ5CEfD5wFzMrK7MY85PtV7WbzbAA0m+cHQBfnIV9efZ+XldlOI/0ikiRJUisVRfFl0h5CGkDLkyDkId+ItKDqjXnIlwHvAibkId++ur4jKbh5EDigX/PZpPR+e1Tf+zuStG/QzsC7R+QFJEmSJHWMlgdApHzio0hT4HYmpd67GzgiD/kmpDR7jwKTgCfykDeOWs0GtiRlv/j9APdezLNT5Rb3aytJkiSpy7RDQHA4cGVWZn/pK8hDfhHwEeA+oG+Dnjuqz7376mVldnce8tuBXlJ2i/4u7ne+N3DFMPVbkiRJUodpeQCUldnrByg7Gji6Ov36AM2uoNogNSuzVzWUh0G+S5IkSVLrA6B2MKV3Cj09PSuvKEmSJKmjtcMaIEmSJElqipaPAOUhnwBc01C0hLR25zBS7vL+3k5aF3Q2KTnCS7IyeyoP+d+AHYA8K7OpecjLgdpmZXZV/8Jp46YxevnoNXsRrXWyMmt1FyRJkjTM2mkEaDdgE+BfSbvP9qWt/lxV3nfMamizHrBrHvKxwPbA4/3uuaK2kiRJkrpMy0eAGizKymxhHvKH+s6rz6VZmS1srJiHvO/r70gbod4JFKQgp9Hz2kqSJEnqXu0UAM2ppq1tBPwauLYqPyUP+Ul9lbIyG9PQZjbPBkCzgff1u+eK2kqSJEnqMu00BW4/0kao/0KaznZyVX5iVd53NJoN/DNpXdDsAe65oraSJEmSukw7BUCPAg8D95M2NX1JVb4EWNh35CFfr6HNDUAJbMnAAdCK2kqSJEnqMu0UAM0hBUA3AfOAvhRcp1TlfcdHG9o8CfwWuC0rs38McM8VtZUkSZLUZVq+Bigrs1lAGOTyYOUA06vPtzXca2zD9xW1fQ43QpUkSZK6QzuNAEmSJEnSiGr5CFA7cCNU9ecmqJIkSe0rxvh54JyiKBasatuOCYDykE8Arhng0tuB7wFbkVJnfwh4EdBbXS+BxcCPgSOyMitHvLOSJEmSRtJi4OcxxnuBHwJXFEUxpN/5nTgFbjfShqd9x+nAb4DXAeOB4xrqvh/YFDgBOJwULEmSJEnqYEVRnF4UxZtJidMOAe6MMU6NMW6ysrYdMwLUYFFWZgv7TvKQ7wUsIkWBAVinoe5jWZk9mId8XnVuGmxJkiSpw8UYxwAHAh8mbXkzhRQH/Ddpn9BBdWIANCcP+dPV92uzMts3D/m6wPnAZqQRoT4X5yEH2BCoA1c0taeSJEmSRsINwI+AA4ui6BvsIMb4upU17MQAaD/g9ur7smpz0wuBvYEPZmVW5CHfurp+JGnd0MKszB5tdkclSZIkjYjLiqLI+xcWRfGVlTXsxADoUdIwV5/vApOAg4HZecg3bLh2X1Zm85AkSZK0Ntk+xjimKIqFK6/6XJ0YAM3pd/4Uae1PvTr/CfDFpvZIkiRJUjPtADwYY1xAyvpcFkXxyqE07JgAKCuzWaRAZ6iGXHdK7xR6enpWuU+SJEmSmq8oiq1Wt23HBECSJEmSBBBjfA1wFtAD3At8tCiKPw2lbccEQHnIA/At4FDSRqd/ISU52Ag4jTXYCHXauGmMXj56pF9BHSIrs1Z3QZIkSSv2PeCwoihujDHuDJzKStJf9+mkjVDfARxD2uhoW+ARUtR3Jm6EKkmSJHWTUBTFjQBFUfwZWD7Uhh0zAkQKeACOJaW9PgKYB2yOG6FKkiRJ3eSpGOO+wK+BfwEeH2rDjhkBysrsetIozhakzU5vB75apblezOAboS4mZYhzI1RJkiRp7fBR0tKY35BmiH18qA07JgDKQ74TsCArs+1I633OAr6ch3xL4CLgvcABWZkVDc2OJKXIG5OV2cFZmQ15aEySJElS2ypJy2P2AT4PPBFjXHcoDTtpCtwuwJl5yN8PXE/aEHUpkOFGqJIkSVI3+W9SBrhbgO2AJcCoGOMXiqL40YoadswIEHAOKQvcaaS1P3sDnyQNffVthPowaWRIkiRJ0tqrF9iuKIo9SAnSbgB2BI5aWcOOGQGq0ld/qToanTNIk1XZNFWSJElS53h5URQPABRF8XCM8eVFUTwUY3x6ZQ07JgAaSVN6p9DT09PqbkiSJEkamj/EGH8M/BZ4E/DnGOMBwH0ra2gAJEmSpK4y7bPTGD16dNOet3jxYpjRtMd1haIojowxTiLtA3peURT/E2OMwGUra9v6NUAhbEYIVxDCo4RwJSFsVpW/gRD+TAiLCeEcQlhnJfcZQwi3E8JJ1fk6hPB9QniIEP5GCG8ctO24cRCCh0c6JEmS1NZijBsBu5ECoPVjjK8qkiUra9v6ACit6Xkp8FrSwqXjSMHOz0gbnk4gbXT6ikHvEMKewHXANg2l+wKTSRsj3UFKniBJkiSp851F+o2/LXAv8MOhNmxNABTCZEJ4jBB+Qwpw3gE8CDwFLCNFcj2k+Xw/BRZRlvNXcMf3AP8PuPOZkrK8lLRf0E3A8uq+kiRJkjrfS4uiOAt4siiK61iFuKaVa4A2BL4PXEdZPkwID5P29TmVZ0dyeoEzgBmEMJuyvGLAO5XlFwAI4bh+5Q8SwiWkDZLePQLvIEmSJKkFYozjq88e0oDHkLR6CtzllOWdhBCAtwB/A84jbXIKcClleRlpf5/XrOYzPgt8B7iQEDZa0w5LkiRJarnPAGcDuwAXA58basNWB0DLCOFbwAXAI8ATwGjgZmABsD8h7Aa8GPi/VbpzCAcQwg3ABqQ1RC+i9e8rSZIkac1tXRTFm4qiGFMUxe6ktUBD0g5psM8CpgO3kIKcwyjLJwnhA8DpwMHAdynLy1fxvr8APkDaFfYB4BDK8pFh67UkSZKkpoox7gv8M3BQjHGPqvgFwHtJCdRWqjUBUFlOJwU9kEZ7np+iuixnA9uv4n23bvi+GHj/kNr19oIboUqSJEnt7kZSBumlQFGVPU2aUTYk7TACNDQh7ApcNcCVEyjLk5vdHUmSJEnNVRTFXcA5McbziqJ4uq88xjj4ljn9dE4ABH8Fdh6g/OE1vfG0cdMYvbx5uwGrfWVl1uouSJIkaeWmxhg/CaxHyi79d+DVQ2nYOQFQWT4OzB1q9Tzkrwe+CbyeNET2K+DYrMzuH5H+SZIkSWqWSaR9Q78DnAKcNtSGa2VWtDzkPcA1pM1VdyUtitoVmJGHfK18Z0mSJKmL3FMUxePARkVR3EYaCRqStTUYOJiU/vqwrMxuz8rsBuALpGQLu7a0Z5IkSZLW1PwY40eBx2KMJwFjhtqwc6bArZqtgAeyMlvUUHZH9TkOmNP8LkmSJElaUzHGw4EjgVcAtwKvBQ4aavu1NQC6E9g0D/nGwDqkgGjr6tqCVnVKkiRJ0uqLMU4FdgR+VBTFnTHGABwNbAJ8bSj3WFunwNVJiQ9+ALwNuBr4IWmj1V+3sF+SJEmSVt/ewAeKolgCUBTFXOAAUlKEIVkrA6CszO4C9gJeBpwBrAvcD7wc2KeFXZMkSZK0+hYXRVE2FhRF8SSwaJD6z7O2ToGjSnwwsbEsD/l7eHbH2GdM6Z1CT09Ps7omSZIkafUsjTFuUxRF3/p+YozbAOUK2jzHWhsADSQrs8ta3QdJkiRJq+044JIY40xSkrMtgXcChw71Bl0VAA1m2rhpjF4+utXdUBvIyqzVXZAkSdIgiqL4W4xxT9I+n68E/gh8tSiKtWsKXB7yCaSNTQGeJq3n+XZWZifnIf86cERV/p2szE7MQz4V6Psl+xQwH/hMVmYzmtpxSZIkScOqKIpHgHNXt31HBEANdiOluD4YOCUP+frAl0iZ3rYAzs5DfmlV934gAhuTssKdDhgASZIkSV2s07LALcrK7P6szL5D2vRoZ2B8VmYzgXurOutUn2VWZguBfwAPA+s1vbeSJEmS2kqnBUCN7gM2ysqsyEM+HjgLmJWV2Y3V9c3ykC8ElgB7Ap9uUT8lSZIktYlOmwLX6GXAb/KQ7wjMJE15O6Dh+oPAG4AngfuzMnuy+V2UJEmS1E46LQDaKA/5S4B3A+OBk4DLgUdJu78+kYe8752eyspsbkt6KUmSJKktdVoANIe0ydE9wDHAJkDfDqZ9myHt3YJ+SZIkSeoAHREAZWU2CwiDXP7uAGVXAFOHev8pvVPo6elZeUVJkiRJHa2TkyBIkiRJ0ipp6ghQnucB+BZwKPAi4C/AkVmW/SHP8w8C/wZsR0pb/TPgmCzLllZtx5BSWi8AxmVZ9nRVPgE4DdgKuBb4UJZlC/L/396dh8lVlYkf/14JCBhWkQnaQhLFg4rKoqiMaCAgLmMQFcUgArKjTlRQRGc8nFEUNyCuuMCgQkRxBkQY10BQ4CcIKiia45IOkCAiSzAREAP398e5TYo2nZAmXUvX9/M89VTdc5d666aT1NvvWVI6Dngn8Hjgd8BxMcYfrCyu2bNnM3HixLH50OoJMcbVHyRJkqSe1+4K0EspY3cOBLYF7gbOTCm9EjiXsqLrU5r9M5tjh7wZuI+y4OkrW9q/DFwB7EiZGOH4lNJ2lETrRGAycC1wbkrJipckSZLUx9o9Buju5vndwDeBo4CbKDO5XRVj/Hizf1FKacsY4/0t5x4OfImy+OnRwHea9j2ApcAyyjihdZrXy4EjgQ0pydDhQ1UjSZIkSf2prRWRGONPgSMoVZzTgT8C/0XpvnYjQEpp75TSEuC2lNJPmrZdge0p1Z4vAnunlKY017yJkvCcA2wJnB5jXATsR0nwTgNuAM6wAiRJkiT1t7YmBCmlZwN/iTE+jZL0nAm8DxikjP0B+DGlynMpsFHTdmTz/CtKovOYobaU0nrAecA+wBtijDml9BRgwxjjDsAk4EOULnS7jOkHlCRJktTV2l0R2Qn4Vkrp1cD9lAVM7wU+D+yYUvpP4AmUas/O8NDkB/sBx1MSox2AzwBvaZKfUyiLoB4MXJZS2pBSYTonpXRM875LKOsH3d6GzyhJkiSpS7U7AfoKZTzO5yhjf14OvDHG+G3g9cC+QKZMhnAdMIsyIcJ6wJkxxvkxxvmUrnBPoCRGR1HG/syhzB53ZoxxHvAu4D3AzcDbgKNijH9oz8eUJEmS1I2quq47HUPHhBAmA4MzZsxwGuw+5zTYkqR+s2jRIqZPnw4wJee8sMPhtEWnvvstW7aMCy+8EFrudQhhGmVSsN9QeiptAJyTc/706q4XQng+ZVjIecAUylCPScBzcs7fGeGcycC5OecXDGt/L3AJ8AzKjMqnAR/IOR8TQngxsCTnfP0afuSu1u5Z4LrSrFmzGBgY6HQYkiRJ6i+X5Jz3BwghPBbIIYSv5ZyXrOa8vYHZrclSCGEPSgKz0gRoJDnnk5vzn9Fs3woMDSN5C2WpGhOgcWfKFFi+vNNRqJP6uBIqSZK6wkbAA8CPQggLgM0pa1+eAUylLPVyCmXysLcA94cQFgGnAs8E3gtsGEK4krL0TKQMd5lIWV/zfuAJIYQLgX8BLso5fzCEcBYlyQFWVIqAtwIvA3YKIWwJvCrnvF9zzBXAfjnnW8bsboyhsR8DVFXTqKqaqtqupW1e0/bwx8jXOHjYsUupqgupKvutSZIkqVftEUKYF0K4hNKl7e2U5V2+nnPek7IO5l9yzrsCe1JmNl4AnAWcknM+v7nOA8DJwJyc84WUhOhNOedpwP9Sxs1DSYYOBHYFXh5CeM5IgeWcrwW+RxlT/zXgWSGEzUIIzwRu79XkBzpXAfq35r0/B2wL7PUIz5sE/B14EaW8dyBlBjlJkiSp1zzUBW5ICOE9lEnBAJ4O/Agg57w0hPAb4CmP4LqLgU+FEJYBTwKuaNqvyznf3bzP1axYhmaVcs51COFs4I2UatQZj+S8btWZhUHrehl1vYRSinuAul7SbK/O3c1xNzfbpd9aVX2EqrqTqrqRqjqAqtqQqlpCVb2j2X90s73hGHwaSZIkaW16sHn+LbAbQAhhI+BZlC5wI50z9N3+S8AhOeeDgVsoMyYDPD2EMDGEMAF4PnDDI4hj6Jr/TakkvRj4vzX5MN2mMwnQ6N1KVf0V+H+Uaa+/RlW9jLJG0KsoA7bOADahTLl9cHPeocBXqOt72h6xJEmSNDpfBB4fQrgcmAeknPNtIxz7K2CfEML+wNnAT5qxOhsBT2yOuRP4BnAl8K2c829W8/5XASeHEJ6ec14MLAXm5px7evB8r02CsCulL+RhwDeo6/uoqmdTstqLm2MeS1lE9XPA26mqNzfbb+pAvJIkSdI/yTnPoyQ1w9untby+HzhoJcec2PJ6cvPyF0BoXp/Lyv3TmJ+mSjTcC5p9XwC+0NL+GHq8+xu0NwHamqoayhbvpa4Xj+IaCygLnG4HnNMkP/MpXeGOoiQ/ewLXU9c3UVWXAJ8BLqWu5z/qTyBJkiT1mRDCBsDllDFLf+h0PI9WOxOg77e8vgyYNqqr1HVNVR1CmY/8DGA6ZUrA2cD6lP6JQ2OEPtvs/9yo3kuSJEnqcznneyk9qsaFsU+A6noeKwZeDd938CO8xlmU6f6Gtm8Ftmw54vjmMfy880d871aDg+BCqJIkSdK4112TIFTV1c1sba2Pn3c6LEmSJEnjQ7dNgrAvsO6wtn+M9ZtOmT2F5RN7ejILPQp1HHkNXkmSJI0vHU2AqlRNAy5tNh/kRG4DPlnH+hNVqoa+ldak6l7KSrQHAq+njPPZgDJDxeeAbSgDs95Ux/ovzbX3bs45p461M8BJkiRJ6poucLsAWwEfAz5epWrPpv1dwBaUaa9fwz9PZf1lysq2O1JmhmsdB3QkcAfwuipVW4xd6JIkSZJ6RbckQEvrWN9Wx/pU4PesWMD03jrWd7Jixdv1hp23B/CeZn8FrANQpWoSZWHUI4F7gbeMafSSJEmSekK3JECt/gw8oXl9SpWqpcD/A+bSOhMcUMf6JmAZcA5lVrjTm11vAf4EnE9ZCffIKlWrnw1OkiRJ0rjWbZMgQOnydkXz+iTga5QK0V0AVVqRx1SpWg/4JvBy4PV1rHOT6BwOPAm4h5LkrQsMjQmSJEmS1Ke6JQHaqErV5sArKGN53gscCtzRVHlGcgowAzgAuKxK1YbAbsBkYHfg1ua4rwBHYwIkSZIk9bVuSYCuBmpKt7Vj61h/u7XSM4LHA0dRxv7Madq+Qan2XF/Het7QgVWq5gCfrFK19WoSKkmSJEnjWEcToCZJWWmmU8d6pPazWDEW6BHFX8d6NjB7pP2DswYZGBh4JJeSJEmS1MO6cRIESZIkSRoT3dIFbq1oWVj15DrWJzRtC4FzKWsEHV3H+vTh502ZPYXlE5e3MfAHYUQAACAASURBVFJ1kzrWqz9IkiRJ48J4rQAdW6XqmZ0OQpIkSVJ3Ga8J0O3A6a79I0mSJKnVeE2AjgVeCBzW6UAkSZIkdY9xNQaoxS8os759lDK9tiRJkiSN2woQwAeApcDmnQ5EkiRJUncYtwlQHeu/AW/tdBySJEmSuse46gI3fGHVOtYXtWy/txMxSZIkSeoe4yoBGq3BWYMMDAx0OgxJkiRJY2zcdoGTJEmSpOF6PwGqeB4Vl1NxLxWLqTiDii2oOJGKWx/RNaZQOsr56K+HJEmS+k5vJ0AV2wA/AhYD2wMzgOcCF9Drn02SJEnSWtfrY4AOAjYEDqNmKQAVxwPfBe7uYFySJEmSulCvV0meCNzxUPJTLGyeJ7c9GkmSJEldrdcToEFgCyo2bWnbtnm+sQPxSJIkSepivZ4AzQHuAc6k4qlUPA84GbgKuAZYp2kfemzeyWAlSZIkdVZvJ0A1NwN7AJtQEp4Lm+d9gAeBLYDftzz+vTOBSpIkSeoGvT4JAtRcA0xfyZ4Tm8fqDQKugypJkiSNe71dAZIkSZKkNdD7FaC1YQqwvNNBqC3qTgcgSZKkTureClDFNCrq5vEAFXdSEZt9e1PxMyqWUfEHKo4Zds6DVGzZtF3ctJ3VqY8iSZIkqTt0bwK0wi7AJOBs4EQqdgEuBi4AAvAF4LNU7NVyzt+B3amYALwYuK+9IUuSJEnqRr2QAC2l5i/A4mb7I8Biak6iZjE1HweeSs0PW865ijI73C7N9g3tC1eSJElSt+qFMUBXN5WcdYFTgKezIhkqav447JzLgJnAzcAV8LCFUiVJktTHTuM0lrdxAPgEJjCVqW17P61aL1SA9qV0dduUmmOBRbROWl2xDhVHUzGp5ZzLgKcCBzSvJUmSJKknEqDF1NxMzd+a7TOAJ1FxAhVPoqz181ngyS3n3AL8AdgOEyBJkiRJjV5IgB6u5irg1cDrgd83zwdS87NhR14G3AP/1C5JkiSpT3XvGKCaeUA1wr7vAN9ZzTmHNQ+AF6zyvQZp7VQnSZIkaZzqvQqQJEmSJI1S91aA2mj27NlMnDix02FojMUYOx2CJEmSOqyrE6CU0jTg0pameyljet4I1MCpwMuBB5r242KMv1rJeUPWjTG2b85DSZIkSV2lV7rA7QJsBuwGvBA4ljIG6JnAnsBzgTuAeSmlSSs5bzNgM5MfSZIkqb91dQWoxdIY45KU0mJKtWcnYGfgFTHGnwGklI4AlgAHsmLmt6UxxiWdCFiSJElS9+mVCtDVKaW7gYWUrm1nNO2DQwfEGJcBtwOTh523pHlc1KZYJUmSJHWpXqkAvQbYETgJ+DZl/R+AbYH5ACmlTYEtgBtbztsX+GPz+r62RCpJkiSpa/VKBWhRjPHjwH8Dn6UkbtcAH0spPT+l9BTgy8Ay4OyW8/5K6Ra3BLgvpdQrn1eSJEnSGOi1hOCdwB8oyc7rKGN9LgCuBTYF9ogx3tJy/NXAXS2Pp7U1WkmSJEldparrutMxdEwIYTIwOHfuXAYGBjodjiRJUtssWrSI6dOnA0zJOS/scDhtMfTdb8GMBSyf2L7JgScsm8DUC6dCH93rbtZrFSBJkiRJGrWemAQhpXQkcDpwUozxP5q2E4G4ksN3B65s9h0CbEKZKOGDMcYLVnb92bNnM3HixDGIXN0gxpX9mEiSJKkf9UoF6AjKQqeHppTWbdpOpixw+tZme1KzfTnwMeBtwDFAAOYA56eUXtrOoCVJkiR1l65PgFJKz6UsfDoTeAJlamtijPc1i5ze0xx6d7O9DnAkcHqM8YIY46IY4yeBq4BZbf8AkiRJkrpG1ydAlOrPlTHGHwDfBY5ezfGbAevTskhqYyEPXyRVkiRJUp/p6gQopbQR8Ebg+Sml+4CXAdNSSk9fxWm3UapC2w5r35aHL5IqSZIkqc90+yQIMykx7gTc37T9CDiKEbqzxRgfTCl9ETgipXQlpevbgc01/m3MI5YkSZLUtbo9AToC+EGM8fqhhpTSecAhKaUTYoz3jHDee4G7KZMhTAJ+D+wfY7x4rAOWJEmS1L1cCBUGZ8yY4TTY45jTYEuS9M9cCNWFUPtVt1eA2mLWrFkMDAx0OgxJkiRJY8wECBdCHe+sAEmSJGlIVyRAKaVpwKXN5oOU8TuzgQoY+vZaN+0XAIfFGB9IKT0POBXYGbgT+B5wfIzx9pTSiS3nDjkkxnjW2H0SSZIkSd2s26bB3oUyacHZwInA0yjTWm8GbA68DzgYeGlKaRvKjHCLge2BGcBzgQtSSkOfa+jcocecNn0OSZIkSV2oKypALZbGGP+SUlrcbK8H1DHGJQAppVua9uXAQcCGlGrQ0mb/8ZTFUndujnvoXEmSJEnqtgTo6pTSBGBd4BRgKfCalNISYB1K97hPUyo/rwXuGEp+Ggub58nN85bNuQCLY4zPHNvwJUmSJHWzbusCty8QgE1jjMc2bXdQusZdDNwHzIkx1sAgsEVKadOW87dtnm9sOXeH5rH3GMcuSZIkqct1WwVocYzx5mFtD8QYf5dSOhi4CvhWSml7ynie9wNnppTeQxnjc3JzzDXAK5pzF7YreEmSJEndrdsqQCOKMd4HzAQeD5zaJEp7AJtQEp4Lm+d9YowPdixQSZIkSV2rquu60zF0zNBqwHPnznUhVEmS1FcWLVrE9OnTAabknBd2OJy2GPrut2DGApZPXN62952wbAJTL5wKfXSvu1nPVIAkSZIk6dHqtjFAnTGFMrG2xpf+LW5KkiRpBL1bAap4HhWXU3EvFYupOIOKLVr2f52Kmoo9OxilJEmSpC7SmwlQxTaUtYAWA9sDM4DnAhdQ8ZgmEXoNZRrsozsWpyRJkjSCEMLkEMJPx+jam4cQZo7RtY8IIaw7Ftduh95MgOAgYEPgMGr+SM21wPHAvwI7AwcDtwPvAGZQ8cROBSpJkiR1wLMpRYKx8D5gnTG69pjr1TFATwTuoGZpS9vC5nkycDhwFnAecFqzndoXniRJkvTIhBDmAb+k9GzaGNgP2AfYLOecQgiPBa6jJDVHUpaGqYFzc86fCiG8hlIM+AdwC7A/Zb3M54QQjgB2bfZtAzwWOBd4FbA1sE/O+Y8hhI8Au1ESm1NyzueNENeewKTmGq8ew9syZnq1AjQIbEHFpi1t2zbPU4GnAe8G7qYskHo4Ve9mqZIkSRr3rs457wn8EHgj8DXg9SGEilLJuQh4KvAG4EWUZOXVIYTQHP/xnPOLmuM2Bk4CLsk5f7G5/sKc80uB31Km434F8D/Aq0IIL2/aXgTsDrw/hDD0PfthceWczwBupSRZPalXE6A5wD3AmVQ8lYrnAScDVwE7AD+lZMg7AHsBT2LsSoCSJEnSo/WL5vlmYP2c811N24sowzu+TKnEbAPMbR6PpxQB3gXsEUK4jFLteXAl1/9587wE+E3z+i5gfeBZwM5Nxed7wLqUXlX/FNej+4jdoTcToJqbgT2ATYBrgAub532AfYFvUjO/eVxCKd0d06lwJUmSpNVY2QIeX6KMad8g5zwfyMANwO4552mUIR/XA0cAJ+acXwJUlO/DD/Lw7/qrWiBkPnBpc809gG8Cf1zFecOv3VN6NnBqrqFmOjWbUrMVNQdR82dq1qfm1GHH7kjNXiNea5DyR+tjfD0kSZJ6WM75MkrV56xm+zpK5efyEMI1lOrPYuBq4KIQwlzK+JyLKAnMs0II73gEb/UdYFkI4SfAtUCdc166iuN/Avxf0z2v51R13b/fFEMIk4HBuXPnMjAw0OlwJEmS2mbRokVMnz4dytiPhR0Opy2GvvstmLGA5ROXt+19JyybwNQLp0If3etu1quzwAGQUnoecCpl6us7KX0WjwfeBsTmsAeARcC/xxgvXNl1Zs+ezcSJE8c+YLVNjHH1B0mSJKnv9GwXuJTSNoy0GGr5XLdRZoCbSkmATu9MpJIkSZK6Rc8mQLQshhpj/GOMcfhiqHWMcQllLvS7gPU6FqkkSZKkrtDLXeCeCNwRYxxpMdQtU0pLKEnSPcBRbY1OkiRJUtfp5QrQILBFSmlli6HeCNxBWQdoCvCEGOO5bY5PkiRJUpfp5QrQHOD9wJkppfdQxvsMLYZ6DbBTjHFh58KTJEmS1G16tgIUY1zVYqgrW/1WkiRJUp/r5QoQMcZrgOkr2XVi85AkSZKkh/R0ArS2zJo1y4VQJUmSpD7Qs13gJEmSJGlNdX0FKKW0LXAasCtwH3AecC/wnpUcvjtwJRCBQyjjg+YDH4wxXjDSe8yePZuJEyeu5cjVSTHGTocgSZKkLtTVFaCU0uOAHwH3AzsCBwOHAutSZn17a3PopGb7cuBjwNuAY4BAmS3u/JTSS9sZuyRJkqTu09UJEPAKYGvg2Bjjwhjj94HtgXfHGJdQFjgFuLvZXgc4Ejg9xnhBjHFRjPGTlKmxZ3UgfkmSJEldpNu7wE1pnhcNNcQYB1dx/GbA+pRFUlstBJ61ViOTJEmS1HO6vQI0lPg8eaghpbR/Smn7EY6/jVIV2nZY+7bAjWs/PEmSJEm9pNsrQBcDtwAfTSkdB+wAfBU4Afj18INjjA+mlL4IHJFSupLS9e1AYCfg39oWtSRJkqSu1NUVoBjj3cBewMbA9cDngVOAU1dx2nuBT1ImQ8jAG4D9Y4wXj220kiRJkrpdVdd1p2PomBDCZGBw7ty5LoQqSZL6yqJFi5g+fTrAlJzzwg6H0xZD3/0WzFjA8onL2/a+E5ZNYOqFU6GP7nU36+oKkCRJkiStTd0+Bqgtpsye0tbfAmhs1bF/q5qSJElata5NgKpUTQMuBU6uY31C07YQOBf4MPBRYD9gPcr4oPfVsf5xy/l7A98Dzqlj/aa2Bi9JkiSpK/VCF7hjq1Q9c1jbVymzuu1HWRj1KuCHVapap8c+ErgDeF2Vqi3aEqkkSZKkrtYLCdDtwOlVqqpm+1+AfYAP17G+tI71TcC7KcnO0QBVqiYBr6IkQfcCb2l71JIkSZK6Ti8kQMcCLwQOa7a3ap4Hhw6oY/0gcBMwuWl6C/An4HzgbODIlgRKkiRJUp/qhQToF8BsypifjSiJDcC2QwdUqZpASX5ubBKdw4EnAfdQqkBTgb3bF7IkSZKkbtS1kyAM8wHgdcDWwJ+BC4ATqlT9FvgdpQvc44EvAC+lJEO7A7c253+F0j3ue22NWpIkSVJX6YUKEHWs/wa8taXpIOBbwJnADcBzgZfVsb4OOAK4vo71vDrW8+tYzwfmAK+sUrV1m0OXJEmS1EW6tgJUx3oeULVsX9S6Dfx78xh+3mtX0jab0o1upQZnDTIwMPBowpUkSZLUA3qiAiRJkiRJa0PXVoDaacrsKSyfuLzTYWgtqWPd6RAkSZLUpVZbAapSdXCVqrpK1fpjFUSVqqdUqTqwV68vSZIkqTd0Sxe4LwOv7OHrS5IkSeoBa9IF7qwqVQOUtXVeAJwLvAO4BTixjvVpVaqOBj4CPJGScHwCmAicBRwH7Ny83ha4mTKz21bANIAqVfcBC4GZQAZ2A75EmdZ6L+B/gUOBAcrU1rsA1wIH17EerFK1ELgc2An4F+A9wAOt169jffAafGZJkiRJ48iaVoCeA/wnZUa1wykLk34FOLjZf2izvX7z/GlgV+AA4NXN82OAFwGnAJtSpqi+gpLcHNNc56nAqZTk5zjg4uZ9DwG2AT7evMd2lITptJYYnwHsQ0mMjhvh+pIkSZL60JomQPPrWF8FXNlsbwB8Dnh2lao3Uyo8n6ckMBsAEbgK2AL4V+CTwPXAd4FjgSfUsb4fWA78o471Pc11b69jfQlljR+A84BfNa/XB57dvNevgf2aaw+ZW8f698AvgA1GuL4kSZKkPrSmCdADzfND02zVsc7AJcBngEubhUcHgfuBTwGvpSxa+l3gNcCdlC5q3wM+1HLdTapUPb7ZfnCE9x0yn5LgvJqSgH1mVTGu5PqSJEmS+tDamgb7s8B0SjJCHeu/VKk6CDgJeBcwj5Kw3A4cCPwO+AulixrA/wH/BZwMLH4E7/dO4ExK17g/UcYSrUrr9Q9/hJ9JkiRJ49DgaTDQxhVQFk2A6VPb935ataqu+3fNlBDCZGBw7ty5DAwMdDocSZKktlm0aBHTp08HmJJzXtjhcNrioe9+CxYwsLx9GdCiCROYPnUq9NG97mbdMg22JEmSJI25tdUF7p/NqaYBl1LG4kxiZn0bc6qLgVdQZog7CDiamfXpLeesT+nO9ilm1v9gTtVanvoH8DPgDcysF63iffcFFjCzvu4Rx3rhFNi0jXVQja2Z/VvVlCRJ0qq1owL0d2B35lQTgBcD963i2DdQ1g5ap6XtXcBmlOmtn0KZPW7l5lTbUKa73urRhSxJkiRpPBq7CtAKVwF7UBY+hRVTWxdzqncAH6YsnHpW03ovUD30ema9hDnVYyjTWS9tznsJcDpl0dULgCMplSWA7zKn2h2YSpmI4fGUKbNfzcz6prX66SRJkiT1jHZUgC6jJEB7UBYkbe1rth9lbaDXMLO+lBWzuU1qOeYU5lRLgD8D6wFfZk5VAecCP6CsCbQTcDQrFjp9HXA5sDmlYvQ8YHtK9ztJkiRJfapdCdBTgQOa1612o4zt2bjZHlqo9O6WY04CdgB2pEyffRHwBEqSdBhwHRAoi6EOnf83ZtbLKcnWMcD7m/dZf219KEmSJEm9px0J0C3AH4Dt+OcE6L+AU4BPMKd6HCsWMW2dKf0eYAlwB6X72+bN69uBrwMzgG8A57WcP4k51SbAqcD/AJ+mdPerkCRJktS32jUN9mWUROZnw9pvp4z/WQf4D8p4oVuAecypNmqOOQW4i7JA6tMoM8c9QJkw4YXA9ymTI1zdnHs18ClgF0o3uQ9TFmq9iYcnVpIkSZL6zNhNgjCznseKisthzQPgBc3zwS1HP2mE1yNXbGbWlwDPXMme57e8/uHqA5UkSZLUL9oxC1z3mzEIAwOdjkKSJEnSGDMBAmZPmc3E5RM7HYYehVjHTocgSZKkHtBzCVCq0jTgUuDpsY7zOxyOJEmSpB7SrkkQJEmSJKnjeq4CtDKpSusCXwBeC9wKvA3IwALgdbGOF6QqfRR4NbBdrGPdsWAlSZIkdcx4qQAdCrwG2JWy9s/ZwG3AxcDBqUoTgDcDp5v8SJIkSf1rXFSAgGcDGwFXUJK6jYCnUtb/uRg4CNgEOKtD8UmSJEnqAr2cAG2dqrS8ef1n4G5KlecplIToRuAGYJBSFTo31vGuTgQqSZIkqTv0cgL0/ZbXVwHnA18FHgA+Eeu4FCBV6fPAKcDn2h6hJEmSpK7ScwlQrOM8oBph96ErOf5USgVoRLMGZzHgQqiSJEnSuDdeJkGQJEmSpNXquQrQmJgCLF/tUepGzuknSZKkNdD7FaCKI6moqfjQsPaNqfg8FYupWELFt6mY3JkgJUmSJHWD3k+A4AjgDuBQKtZtaT8T2B14FfCvwLbAnPaHJ0mSJKlb9HYCVPFcYCdgJvAEYN+mfUvgtcDHqPk5NTcAezYPSZIkSX2qtxOgUv25kpofAN8Fjm7aJzfPix46suYWau5pa3SSJEmSukrvToJQsRHwRmADKu4D1gEmUPF0ViQ+T245fhqwLjU/bHOkkiRJkrpE7yZApdvbBEoXuPubth8BR1Ezi4qLgeOouLbZ99/Ar8EESJIkSepXvdwF7gjgB9RcT818auYD5wEHUbEhcCBwJfB94DLgWuCQjkUrSZIkqeN6twJUs/NK2t4JvLPZugc49BFdaxAYWFuBSZIkSepWvVwBkiRJkqQ10hMVoJTSNODSZvNB4DbgkzHGT6SUPgwc1bSfGmM8qeW8rwP7A3vFGH800vVnz57NxIkTxyp8PUoxxk6HIEmSpHGi1ypAuwBbAR8DPp5S+g/gBGA/4DjgQyml7QFSSlsAr6Esknr0yi8nSZIkqZ/0WgK0NMZ4W4zxVOD3wA7AdjHGucCtzTHrNM8HA7cD7wBmpJSe2O5gJUmSJHWXXkuAWv0Z2CjGmFNK2wFnAvNijNc1+w8HzqLMDHd3sy1JkiSpj/VyArQFcHPT5e0ySle3NwCklHYHnga8m5L8bAYcnlJaZ4RrSZIkSeoDPTEJQouNUkqbA68AtgNOBr4L/BWYAdyfUppAWSPop6xY9+eJwNzmmPPbHbQkSZLUKoQwDfgm8BugBjYAzsk5f3oNr7MvcFXO+ZZVXHtjYAFwQM75/hGu8yxgs5zzj9f4w/SYXkuArqb8If4JOJZS2RlawWdB83wgsC9wQoxxftM2P6X0S+AYTIAkSZLUHS7JOe8PEEJ4LJBDCF/LOS9Zg2vMosyIfMuw9oeu3Vx/DqUY8K0RrvNayph6E6BuEGOcB1Qj7D5tJW1nr+QaO67NmCRJkqS1aCPgAWB5COElQKQMV5kIzARuolR1NgE2BN4PrEuZFOyrIYQXraK6sx5lJuW7mu2PALtRJg87BbiSMoHY/SGEnzfvs13O+b4QwsnAfGAh8FHgfuCLlKEmlwHPphQo9gHWA77RxL0+cFTO+Zdr5e6sRT2RAI21WbNmMTAwsPoDJUmSpLVnjxDCPMp6lv8A3p5zXhZCeCbwppzzLSGE91GWfLmAMgb+ZcCWwNNyzheHEH5JSTSGJz9D196yuf4Xc85zQwgvB6bknF8UQlifMmxkGmXysFtzzleHEEaKd/2c8/MBQggfBL6ec357COEc4OXAUsq4/DcDzwAe9yjvz5gwAZIkSZI642Hd1FosBj4VQlgGPAm4Iud8QwjhC8DXKZWfTz2Sa4cQHg/8EBhs2p8F7NwkRzTXmryK67T2wsrD9v2ieb6ZUvH5JrAt8G1KQveh1cTYET2fAKUqbUvpBrcrcB9l2uvjgXuGHXpjrOPklV1j9pTZTFw+cSzD1CMU69jpECRJkjrtS8BTcs5LQwhfAapmkoKNcs6vDCFsRem2dhGlujPizM455ztCCG8CLg0h7EDpznZpzvmIEMJjgP8E/jjsOvcBW4UQFlK62P22aX9w2OXrYdvTgD/lnF8aQngh8GFg9zX+9GOsl6fBJlXpccCPKH0Rd6T0XTyU8gcJ8C7KRAmbUfonSpIkSd3ubOAnIYQrKGODngj8HpgWQvgx5Rf+H2iOvZIyBmjzkS6Wc/4NpWL0KeA7wLIQwk+Aa4E657y0ef22EMLuwMeA/2sed61B3NcBhzXVpY8DH1mDc9umquvhiVvvSFXaj1Jqe0qs44KmbQplkNhy4OhYx9NHOj+EMBkYnLFghhWgLmEFSJKk9li0aBHTp0+HMh5kYYfDaYuh735zFyxgYPnytr3vogkTmD51KvTRve5mvd4FbkrzvGioIdZxECBVCeCUVKWTm11vj3X8WnvDkyRJktRNej0BGkp8nkzpu0iq0v7Ar5v2k4Bzmte3tzc0SZIkSd2m1xOgiymLPn00Vek4mnnQgROa/XfEOi7sUGySJEmSukxPT4IQ63g3sBewMXA98HnKYk6ndjIuSZIkSd2p1ytAxDr+BnjpSnZVK2lbqVmDLoQqSZIk9YOergBJkiRJ0pro+QrQ2uBCqGPP6a0lSZLUDcZtApSqtC1wGrArZTXb84DjYx3v7WhgkiRJkjpmXCZAqUqPA34E/BzYEQjA/wLLgPd1MDRJkiRJHTQuEyDgFcDWwO7NNNgLU5W2B27qaFSSJEmSOmq8JkBTmuehhVKJdRzsUCySJEmSusR4nQVuKPF58lBDqtL+TRVIkiRJUp8arxWgi4FbgI+mKh0H7AB8FTgB+HUnA5MkSZLUOeOyAhTreDewF7AxcD3weeAU4NROxiVJkiSps8ZrBYhYx98AL30kx84anMXAwMAYRyRJkiSp08ZlBUiSJEmSVsYESJIkSVLfMAGSJEmS1DdMgCRJkiT1DRMgSZIkSX3DBEiSJElS3zABkiRJktQ3TIAkSZIk9Q0TIEmSJEl9wwRIkiRJUt8wAZIkSZLUN0yAJEmSJPUNEyBJkiRJfcMESJIkSVLfMAGSJEmS1DdMgCRJkiT1DRMgSZIkSX3DBEiSJElS3zABkiRJktQ3TIAkSZIk9Y0JnQ6gw9YBuPXWWzsdhyRJUlu1fP9Zp5NxdMKtE9r7Fbjd76dV6/c/ja0ADjjggE7HIUmS1ClbAX/sdBBt8lfgrgO23nqzDrz3Xc37q8P6PQH6GbAb8CfggQ7HIkmS1E7rUJKfn3U6kHbJOd8ZQngqsHEH3v6vOec7O/C+Gqaq67rTMUiSJElSWzgJgiRJkqS+MW67wIUQHgN8DngO8HfgsJzzH1r2Hw4cCSwHPpRzviiEsAUwB9gAuAU4JOd8T9uDHyOjuSct+94BTMo5v7e9UY+tUf6cbA2cSfn7UwFH5Jxz24MfI6O8J1sBZwPrAXcCb8o5L2178GPkUf7deQlwds75ye2NeuyM8mdkc+B3wK+bw87POc9ub+RjZ5T35HHA54EplL87b885X9324MfIKO/JacAOzSGTgCU55xe0N/Kx8yj+z/ka5f+bO4GZ4+m7idQJ47kC9Gpg/ZzzC4H3Ap8c2hFCmAT8O/CvwN7AR0IIjwU+AMzJOe8G/ILyj9B4ssb3JISwQQjhHOCtnQi4DUbzc/JB4DM552nAh4GPtDvoMTaae3I88JWWvzuHtT3qsTWae0II4cnAu4B12x7x2BrN/dgJ+HrOeVrzGDfJT2M09+TdwK+bvzeHA6HtUY+tNb4nOed3NP+27gXcTbkv48lofk7eCXwj5/xi4Abg0LZHLY0z4zkBehHwPYCc80+B57bs2wW4Iuf895zz3cAfgGe3ngN8F9izfeG2xWjuyfrAV4CT2hxru4zmnhwLXNwcMwG4r33htsVo7sk7gbOb324+GVjS3pDH3BrfkxDC+sDpwDHtDrYNRvMzsjOwcwjhshDCeU3VcDwZzT3ZG7g/hPB94D+B77c35DE3mnsy5O3AD3LOv2pXsG0ymnvyS2BoxrKNgX+0L1xpJX+DbwAAAs5JREFUfBrPCdDGlN8eDXkghDBhhH1LgU2GtQ+1jSdrfE9yznflnH/QrgA7YDT35Pac8z9CCAH4BJDaE2rbjOae1JTZhH4N7A5c0o5A22g0/558BvhEznlxe0Jsq9Hcj/nAB3LOLwEuAD7djkDbaDT3ZAtgs5zz3sB3KP+ejCejuSeEENaj9MAYb/cDRndPFgFvCyHcALwcOK8dgUrj2XhOgP4KbNSy/Zic8/IR9m1E+Y11a/tQ23gymnsy3o3qnoQQdqd8iTtwPI3/aYzqnuSc/5FzfgZwBPDVdgTaRmt6T+6nTLEfQwjzgM1DCOe2I9A2Gc3PyCXApU3b+cCOYx1km43mntwBXNi0fYeHVwPGg9H+n7Mn8OOmCjLejOaefBw4OOf8TGAW4+/fV6ntxnMCdAXwCoAQwguA1jL61cBuIYT1QwibAE+n/Ob6oXMov2X5SfvCbYvR3JPxbo3vSZP8zAZelnO+pt0Bt8Fo7snnmvsC5beWD7Yz4DZY03tydc45DI13Ae7MOe/f7qDH0Gj+Lfky8NrmmOnAte0Lty1Gc08uZ8X/OUPjO8aT0f6fsyelG/p4NJp7chcrKkO3sKI7nKRRGrfrALXMtPJsyswph1D+0flDzvnCZqaVIyhJ4Idzzv8TQvgXyniXjYDbKTOt/K0jH2AMjOaetJx7MLDdOJ4Fbk1+Tq4DHgvc2lwm55zHzYQZo7wn21HGu9SU5OdtOeffduQDjIFH83enOf/WnPOkNoc9Zkb5MzKFMntiBfyNMvvVnzryAcbAKO/J5pTEcCvKuI4355wXdiL+sTDavzchhIuB9+ecf9mZyMfOKH9OnkHpUrtOc86snPMvOvIBpHFi3CZAkiRJkjTceO4CJ0mSJEkPYwIkSZIkqW+YAEmSJEnqGyZAkiRJkvqGCZAkSZKkvmECJEmSJKlvmABJkiRJ6hsmQJIkSZL6xv8HF62dED+cc2UAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x121337ba8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model.plotIndividualFeatureImportance(sess, dl_test, plotPath='model_char_shift/model_FFN/plot', top=46, figsize=(12, 9))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Model Performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n"
     ]
    }
   ],
   "source": [
    "Ftrain = model.getSDFFactor(sess, dl_train, normalized=True, norm='l1')\n",
    "Fvalid = model.getSDFFactor(sess, dl_valid, normalized=True, norm='l1')\n",
    "Ftest = model.getSDFFactor(sess, dl_test, normalized=True, norm='l1')\n",
    "# np.savez('model_char_shift/model_FFN/SDF_normalized.npz', train=Ftrain, valid=Fvalid, test=Ftest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">==================> SDF Portfolio Sharpe Ratio: Train 0.45\tValid 0.42\tTest 0.44\n",
      ">==================> Max 1 Month Loss: Train -3.30\tValid -4.61\tTest -3.37\n",
      ">==================> Max Drawdown: Train 6\tValid 3\tTest 5\n"
     ]
    }
   ],
   "source": [
    "SR_train = sharpe(Ftrain)\n",
    "SR_valid = sharpe(Fvalid)\n",
    "SR_test = sharpe(Ftest)\n",
    "deco_print('SDF Portfolio Sharpe Ratio: Train %0.2f\\tValid %0.2f\\tTest %0.2f' %(SR_train, SR_valid, SR_test))\n",
    "loss_train = max_1_month_loss(Ftrain / Ftrain.std())\n",
    "loss_valid = max_1_month_loss(Fvalid / Fvalid.std())\n",
    "loss_test = max_1_month_loss(Ftest / Ftest.std())\n",
    "deco_print('Max 1 Month Loss: Train %0.2f\\tValid %0.2f\\tTest %0.2f' %(loss_train, loss_valid, loss_test))\n",
    "drawdown_train = max_drawdown(Ftrain)\n",
    "drawdown_valid = max_drawdown(Fvalid)\n",
    "drawdown_test = max_drawdown(Ftest)\n",
    "deco_print('Max Drawdown: Train %d\\tValid %d\\tTest %d' %(drawdown_train, drawdown_valid, drawdown_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n"
     ]
    }
   ],
   "source": [
    "Ftrain = model.getSDFFactor(sess, dl_train)\n",
    "Fvalid = model.getSDFFactor(sess, dl_valid)\n",
    "Ftest = model.getSDFFactor(sess, dl_test)\n",
    "# np.savez('model_char_shift/model_FFN/SDF.npz', train=Ftrain, valid=Fvalid, test=Ftest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">==================> SDF Portfolio Sharpe Ratio: Train 0.48\tValid 0.42\tTest 0.47\n",
      ">==================> Max 1 Month Loss: Train -4.60\tValid -4.72\tTest -3.28\n",
      ">==================> Max Drawdown: Train 6\tValid 3\tTest 5\n"
     ]
    }
   ],
   "source": [
    "SR_train = sharpe(Ftrain)\n",
    "SR_valid = sharpe(Fvalid)\n",
    "SR_test = sharpe(Ftest)\n",
    "deco_print('SDF Portfolio Sharpe Ratio: Train %0.2f\\tValid %0.2f\\tTest %0.2f' %(SR_train, SR_valid, SR_test))\n",
    "loss_train = max_1_month_loss(Ftrain / Ftrain.std())\n",
    "loss_valid = max_1_month_loss(Fvalid / Fvalid.std())\n",
    "loss_test = max_1_month_loss(Ftest / Ftest.std())\n",
    "deco_print('Max 1 Month Loss: Train %0.2f\\tValid %0.2f\\tTest %0.2f' %(loss_train, loss_valid, loss_test))\n",
    "drawdown_train = max_drawdown(Ftrain)\n",
    "drawdown_valid = max_drawdown(Fvalid)\n",
    "drawdown_test = max_drawdown(Ftest)\n",
    "deco_print('Max Drawdown: Train %d\\tValid %d\\tTest %d' %(drawdown_train, drawdown_valid, drawdown_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n"
     ]
    }
   ],
   "source": [
    "w_train = model.getPrediction(sess, dl_train)\n",
    "w_valid = model.getPrediction(sess, dl_valid)\n",
    "w_test = model.getPrediction(sess, dl_test)\n",
    "# np.savez('model_char_shift/model_FFN/forecasts.npz', train=w_train, valid=w_valid, test=w_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "### Large Cap\n",
    "Ftrain = getFactorWithLargeCap(w_train, dl_train, mask_train, normalized=True, norm='l1')\n",
    "Fvalid = getFactorWithLargeCap(w_valid, dl_valid, mask_valid, normalized=True, norm='l1')\n",
    "Ftest = getFactorWithLargeCap(w_test, dl_test, mask_test, normalized=True, norm='l1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">==================> SDF Portfolio Sharpe Ratio: Train 0.32\tValid 0.17\tTest 0.24\n",
      ">==================> Max 1 Month Loss: Train -3.30\tValid -4.53\tTest -5.08\n",
      ">==================> Max Drawdown: Train 7\tValid 5\tTest 5\n"
     ]
    }
   ],
   "source": [
    "SR_train = sharpe(Ftrain)\n",
    "SR_valid = sharpe(Fvalid)\n",
    "SR_test = sharpe(Ftest)\n",
    "deco_print('SDF Portfolio Sharpe Ratio: Train %0.2f\\tValid %0.2f\\tTest %0.2f' %(SR_train, SR_valid, SR_test))\n",
    "loss_train = max_1_month_loss(Ftrain / Ftrain.std())\n",
    "loss_valid = max_1_month_loss(Fvalid / Fvalid.std())\n",
    "loss_test = max_1_month_loss(Ftest / Ftest.std())\n",
    "deco_print('Max 1 Month Loss: Train %0.2f\\tValid %0.2f\\tTest %0.2f' %(loss_train, loss_valid, loss_test))\n",
    "drawdown_train = max_drawdown(Ftrain)\n",
    "drawdown_valid = max_drawdown(Fvalid)\n",
    "drawdown_test = max_drawdown(Ftest)\n",
    "deco_print('Max Drawdown: Train %d\\tValid %d\\tTest %d' %(drawdown_train, drawdown_valid, drawdown_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "### Large Cap 0.001%\n",
    "Ftrain = getFactorWithLargeCap(w_train, dl_train, mask_train_0001, normalized=True, norm='l1')\n",
    "Fvalid = getFactorWithLargeCap(w_valid, dl_valid, mask_valid_0001, normalized=True, norm='l1')\n",
    "Ftest = getFactorWithLargeCap(w_test, dl_test, mask_test_0001, normalized=True, norm='l1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">==================> SDF Portfolio Sharpe Ratio: Train 0.42\tValid 0.20\tTest 0.30\n",
      ">==================> Max 1 Month Loss: Train -3.89\tValid -4.66\tTest -4.33\n",
      ">==================> Max Drawdown: Train 6\tValid 4\tTest 5\n"
     ]
    }
   ],
   "source": [
    "SR_train = sharpe(Ftrain)\n",
    "SR_valid = sharpe(Fvalid)\n",
    "SR_test = sharpe(Ftest)\n",
    "deco_print('SDF Portfolio Sharpe Ratio: Train %0.2f\\tValid %0.2f\\tTest %0.2f' %(SR_train, SR_valid, SR_test))\n",
    "loss_train = max_1_month_loss(Ftrain / Ftrain.std())\n",
    "loss_valid = max_1_month_loss(Fvalid / Fvalid.std())\n",
    "loss_test = max_1_month_loss(Ftest / Ftest.std())\n",
    "deco_print('Max 1 Month Loss: Train %0.2f\\tValid %0.2f\\tTest %0.2f' %(loss_train, loss_valid, loss_test))\n",
    "drawdown_train = max_drawdown(Ftrain)\n",
    "drawdown_valid = max_drawdown(Fvalid)\n",
    "drawdown_test = max_drawdown(Ftest)\n",
    "deco_print('Max Drawdown: Train %d\\tValid %d\\tTest %d' %(drawdown_train, drawdown_valid, drawdown_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n"
     ]
    }
   ],
   "source": [
    "# EV, XS-R2\n",
    "EV_train, XSR2_train, WXSR2_train = model.calculateStatistics(sess, dl_train)\n",
    "EV_valid, XSR2_valid, WXSR2_valid = model.calculateStatistics(sess, dl_valid)\n",
    "EV_test, XSR2_test, WXSR2_test = model.calculateStatistics(sess, dl_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">==================> Explained Variation: Train 0.11\tValid 0.04\tTest 0.04\n",
      ">==================> XS-R2: Train -0.01\tValid 0.02\tTest 0.03\n",
      ">==================> (Weighted) XS-R2: Train 0.14\tValid -0.00\tTest 0.15\n"
     ]
    }
   ],
   "source": [
    "deco_print('Explained Variation: Train %0.2f\\tValid %0.2f\\tTest %0.2f' %(EV_train, EV_valid, EV_test))\n",
    "deco_print('XS-R2: Train %0.2f\\tValid %0.2f\\tTest %0.2f' %(XSR2_train, XSR2_valid, XSR2_test))\n",
    "deco_print('(Weighted) XS-R2: Train %0.2f\\tValid %0.2f\\tTest %0.2f' %(WXSR2_train, WXSR2_valid, WXSR2_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "### Large Cap\n",
    "EV_train, XSR2_train, WXSR2_train = calculateStatistics(w_train, dl_train, mask_train)\n",
    "EV_valid, XSR2_valid, WXSR2_valid = calculateStatistics(w_valid, dl_valid, mask_valid)\n",
    "EV_test, XSR2_test, WXSR2_test = calculateStatistics(w_test, dl_test, mask_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">==================> Explained Variation: Train 0.13\tValid 0.22\tTest 0.09\n",
      ">==================> XS-R2: Train -0.02\tValid 0.06\tTest 0.03\n",
      ">==================> (Weighted) XS-R2: Train 0.22\tValid 0.15\tTest 0.26\n"
     ]
    }
   ],
   "source": [
    "deco_print('Explained Variation: Train %0.2f\\tValid %0.2f\\tTest %0.2f' %(EV_train, EV_valid, EV_test))\n",
    "deco_print('XS-R2: Train %0.2f\\tValid %0.2f\\tTest %0.2f' %(XSR2_train, XSR2_valid, XSR2_test))\n",
    "deco_print('(Weighted) XS-R2: Train %0.2f\\tValid %0.2f\\tTest %0.2f' %(WXSR2_train, WXSR2_valid, WXSR2_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "### Large Cap 0.001%\n",
    "EV_train, XSR2_train, WXSR2_train = calculateStatistics(w_train, dl_train, mask_train_0001)\n",
    "EV_valid, XSR2_valid, WXSR2_valid = calculateStatistics(w_valid, dl_valid, mask_valid_0001)\n",
    "EV_test, XSR2_test, WXSR2_test = calculateStatistics(w_test, dl_test, mask_test_0001)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">==================> Explained Variation: Train 0.11\tValid 0.10\tTest 0.05\n",
      ">==================> XS-R2: Train 0.01\tValid 0.08\tTest -0.02\n",
      ">==================> (Weighted) XS-R2: Train 0.19\tValid 0.08\tTest 0.18\n"
     ]
    }
   ],
   "source": [
    "deco_print('Explained Variation: Train %0.2f\\tValid %0.2f\\tTest %0.2f' %(EV_train, EV_valid, EV_test))\n",
    "deco_print('XS-R2: Train %0.2f\\tValid %0.2f\\tTest %0.2f' %(XSR2_train, XSR2_valid, XSR2_test))\n",
    "deco_print('(Weighted) XS-R2: Train %0.2f\\tValid %0.2f\\tTest %0.2f' %(WXSR2_train, WXSR2_valid, WXSR2_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n"
     ]
    }
   ],
   "source": [
    "### Turnover\n",
    "w_train = model.getNormalizedWeightWithData(sess, dl_train)\n",
    "w_valid = model.getNormalizedWeightWithData(sess, dl_valid)\n",
    "w_test = model.getNormalizedWeightWithData(sess, dl_test)\n",
    "\n",
    "_, _, T_train_plus_mean, T_train_minus_mean = calculateTurnover_with_dl(dl_train, w_train, rf_train)\n",
    "_, _, T_valid_plus_mean, T_valid_minus_mean = calculateTurnover_with_dl(dl_valid, w_valid, rf_valid)\n",
    "_, _, T_test_plus_mean, T_test_minus_mean = calculateTurnover_with_dl(dl_test, w_test, rf_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">==================> Turnover (Long): Train 0.69\tValid 0.63\tTest 0.65\n",
      ">==================> Turnover (Short): Train 1.38\tValid 1.29\tTest 1.27\n"
     ]
    }
   ],
   "source": [
    "deco_print('Turnover (Long): Train %0.2f\\tValid %0.2f\\tTest %0.2f' %(T_train_plus_mean, T_valid_plus_mean, T_test_plus_mean))\n",
    "deco_print('Turnover (Short): Train %0.2f\\tValid %0.2f\\tTest %0.2f' %(T_train_minus_mean, T_valid_minus_mean, T_test_minus_mean))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Performance of Extreme Portfolio"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "deciles = [0.01, 0.05, 0.1, 0.25, 0.5]\n",
    "SR = pd.DataFrame(np.zeros((len(deciles), 3)), index=deciles, columns=['train','valid','test'])\n",
    "SR_value = pd.DataFrame(np.zeros((len(deciles), 3)), index=deciles, columns=['train','valid','test'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n"
     ]
    }
   ],
   "source": [
    "for decile in deciles:\n",
    "    SR.loc[decile, 'train'] = eval_RtnFcst(model, sess, dl_train, quantile=decile)[1]\n",
    "    SR.loc[decile, 'valid'] = eval_RtnFcst(model, sess, dl_valid, quantile=decile)[1]\n",
    "    SR.loc[decile, 'test'] = eval_RtnFcst(model, sess, dl_test, quantile=decile)[1]\n",
    "for decile in deciles:\n",
    "    SR_value.loc[decile, 'train'] = eval_RtnFcst(model, sess, dl_train, value=value, quantile=decile)[1]\n",
    "    SR_value.loc[decile, 'valid'] = eval_RtnFcst(model, sess, dl_valid, value=value_valid, quantile=decile)[1]\n",
    "    SR_value.loc[decile, 'test'] = eval_RtnFcst(model, sess, dl_test, value=value_test, quantile=decile)[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>train</th>\n",
       "      <th>valid</th>\n",
       "      <th>test</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0.01</th>\n",
       "      <td>1.241270</td>\n",
       "      <td>0.648566</td>\n",
       "      <td>0.655466</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.05</th>\n",
       "      <td>1.364297</td>\n",
       "      <td>1.098866</td>\n",
       "      <td>0.706342</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.10</th>\n",
       "      <td>1.295585</td>\n",
       "      <td>1.314723</td>\n",
       "      <td>0.672129</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.25</th>\n",
       "      <td>1.190140</td>\n",
       "      <td>1.199377</td>\n",
       "      <td>0.571438</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.50</th>\n",
       "      <td>1.090735</td>\n",
       "      <td>1.256454</td>\n",
       "      <td>0.521496</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         train     valid      test\n",
       "0.01  1.241270  0.648566  0.655466\n",
       "0.05  1.364297  1.098866  0.706342\n",
       "0.10  1.295585  1.314723  0.672129\n",
       "0.25  1.190140  1.199377  0.571438\n",
       "0.50  1.090735  1.256454  0.521496"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "SR"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>train</th>\n",
       "      <th>valid</th>\n",
       "      <th>test</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0.01</th>\n",
       "      <td>0.977004</td>\n",
       "      <td>0.349197</td>\n",
       "      <td>0.385756</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.05</th>\n",
       "      <td>0.887942</td>\n",
       "      <td>0.712945</td>\n",
       "      <td>0.419876</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.10</th>\n",
       "      <td>0.700880</td>\n",
       "      <td>0.449859</td>\n",
       "      <td>0.318788</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.25</th>\n",
       "      <td>0.545985</td>\n",
       "      <td>0.284607</td>\n",
       "      <td>0.174412</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.50</th>\n",
       "      <td>0.427917</td>\n",
       "      <td>0.199222</td>\n",
       "      <td>0.149572</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         train     valid      test\n",
       "0.01  0.977004  0.349197  0.385756\n",
       "0.05  0.887942  0.712945  0.419876\n",
       "0.10  0.700880  0.449859  0.318788\n",
       "0.25  0.545985  0.284607  0.174412\n",
       "0.50  0.427917  0.199222  0.149572"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "SR_value"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Predictive Performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "decile = 10\n",
    "for _, (I_macro, I, R, mask) in enumerate(dl_train.iterateOneEpoch(subEpoch=False)):\n",
    "    portfolios_train = construct_decile_portfolios(w_train, R, mask, decile=decile)\n",
    "for _, (I_macro, I, R, mask) in enumerate(dl_valid.iterateOneEpoch(subEpoch=False)):\n",
    "    portfolios_valid = construct_decile_portfolios(w_valid, R, mask, decile=decile)\n",
    "for _, (I_macro, I, R, mask) in enumerate(dl_test.iterateOneEpoch(subEpoch=False)):\n",
    "    portfolios_test = construct_decile_portfolios(w_test, R, mask, decile=decile)\n",
    "df = pd.DataFrame(np.concatenate([portfolios_train, portfolios_valid, portfolios_test]), columns=['decile %d' %(k+1) for k in range(decile)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x16b0d25f8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_decile_portfolios(df, plotPath='model_char_shift/model_FFN/plot')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "decile = 10\n",
    "for _, (I_macro, I, R, mask) in enumerate(dl_train.iterateOneEpoch(subEpoch=False)):\n",
    "    portfolios_train = construct_decile_portfolios(w_train, R, mask, value=value, decile=decile)\n",
    "for _, (I_macro, I, R, mask) in enumerate(dl_valid.iterateOneEpoch(subEpoch=False)):\n",
    "    portfolios_valid = construct_decile_portfolios(w_valid, R, mask, value=value_valid, decile=decile)\n",
    "for _, (I_macro, I, R, mask) in enumerate(dl_test.iterateOneEpoch(subEpoch=False)):\n",
    "    portfolios_test = construct_decile_portfolios(w_test, R, mask, value=value_test, decile=decile)\n",
    "df = pd.DataFrame(np.concatenate([portfolios_train, portfolios_valid, portfolios_test]), columns=['decile %d' %(k+1) for k in range(decile)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1b98bed30>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_decile_portfolios(df, plotPath='model_char_shift/model_FFN/plot')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">==================> Plotting Variable: A2ME\tElapse / Estimate: 1.29s / 59.51s\n",
      ">==================> Plotting Variable: AC\tElapse / Estimate: 2.67s / 61.39s\n",
      ">==================> Plotting Variable: AT\tElapse / Estimate: 3.97s / 60.93s\n",
      ">==================> Plotting Variable: ATO\tElapse / Estimate: 5.38s / 61.84s\n",
      ">==================> Plotting Variable: BEME\tElapse / Estimate: 6.89s / 63.35s\n",
      ">==================> Plotting Variable: Beta\tElapse / Estimate: 8.19s / 62.82s\n",
      ">==================> Plotting Variable: C\tElapse / Estimate: 9.52s / 62.56s\n",
      ">==================> Plotting Variable: CF\tElapse / Estimate: 10.84s / 62.33s\n",
      ">==================> Plotting Variable: CF2P\tElapse / Estimate: 12.16s / 62.15s\n",
      ">==================> Plotting Variable: CTO\tElapse / Estimate: 13.48s / 62.01s\n",
      ">==================> Plotting Variable: D2A\tElapse / Estimate: 14.79s / 61.86s\n",
      ">==================> Plotting Variable: D2P\tElapse / Estimate: 16.13s / 61.82s\n",
      ">==================> Plotting Variable: DPI2A\tElapse / Estimate: 17.49s / 61.89s\n",
      ">==================> Plotting Variable: E2P\tElapse / Estimate: 18.76s / 61.62s\n",
      ">==================> Plotting Variable: FC2Y\tElapse / Estimate: 20.07s / 61.54s\n",
      ">==================> Plotting Variable: IdioVol\tElapse / Estimate: 21.40s / 61.53s\n",
      ">==================> Plotting Variable: Investment\tElapse / Estimate: 22.76s / 61.58s\n",
      ">==================> Plotting Variable: Lev\tElapse / Estimate: 24.12s / 61.64s\n",
      ">==================> Plotting Variable: LME\tElapse / Estimate: 25.45s / 61.61s\n",
      ">==================> Plotting Variable: LT_Rev\tElapse / Estimate: 26.83s / 61.71s\n",
      ">==================> Plotting Variable: LTurnover\tElapse / Estimate: 28.21s / 61.78s\n",
      ">==================> Plotting Variable: MktBeta\tElapse / Estimate: 29.63s / 61.96s\n",
      ">==================> Plotting Variable: NI\tElapse / Estimate: 30.97s / 61.94s\n",
      ">==================> Plotting Variable: NOA\tElapse / Estimate: 32.28s / 61.87s\n",
      ">==================> Plotting Variable: OA\tElapse / Estimate: 33.55s / 61.73s\n",
      ">==================> Plotting Variable: OL\tElapse / Estimate: 35.08s / 62.06s\n",
      ">==================> Plotting Variable: OP\tElapse / Estimate: 36.47s / 62.14s\n",
      ">==================> Plotting Variable: PCM\tElapse / Estimate: 37.78s / 62.06s\n",
      ">==================> Plotting Variable: PM\tElapse / Estimate: 39.07s / 61.97s\n",
      ">==================> Plotting Variable: PROF\tElapse / Estimate: 40.37s / 61.91s\n",
      ">==================> Plotting Variable: Q\tElapse / Estimate: 41.65s / 61.81s\n",
      ">==================> Plotting Variable: r2_1\tElapse / Estimate: 42.93s / 61.72s\n",
      ">==================> Plotting Variable: r12_2\tElapse / Estimate: 44.21s / 61.63s\n",
      ">==================> Plotting Variable: r12_7\tElapse / Estimate: 45.58s / 61.67s\n",
      ">==================> Plotting Variable: r36_13\tElapse / Estimate: 46.87s / 61.60s\n",
      ">==================> Plotting Variable: Rel2High\tElapse / Estimate: 48.20s / 61.59s\n",
      ">==================> Plotting Variable: Resid_Var\tElapse / Estimate: 49.52s / 61.57s\n",
      ">==================> Plotting Variable: RNA\tElapse / Estimate: 50.94s / 61.66s\n",
      ">==================> Plotting Variable: ROA\tElapse / Estimate: 52.31s / 61.70s\n",
      ">==================> Plotting Variable: ROE\tElapse / Estimate: 53.61s / 61.65s\n",
      ">==================> Plotting Variable: S2P\tElapse / Estimate: 54.89s / 61.58s\n",
      ">==================> Plotting Variable: SGA2S\tElapse / Estimate: 56.18s / 61.53s\n",
      ">==================> Plotting Variable: Spread\tElapse / Estimate: 57.47s / 61.48s\n",
      ">==================> Plotting Variable: ST_REV\tElapse / Estimate: 58.75s / 61.42s\n",
      ">==================> Plotting Variable: SUV\tElapse / Estimate: 60.04s / 61.37s\n",
      ">==================> Plotting Variable: Variance\tElapse / Estimate: 61.32s / 61.32s\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x16b175908>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plotReturnDecile(dl_test, w_test, plotPath='model_char_shift/model_FFN/plot')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">==================> Plotting Variable: A2ME\tElapse / Estimate: 1.42s / 65.13s\n",
      ">==================> Plotting Variable: AC\tElapse / Estimate: 2.93s / 67.38s\n",
      ">==================> Plotting Variable: AT\tElapse / Estimate: 4.38s / 67.15s\n",
      ">==================> Plotting Variable: ATO\tElapse / Estimate: 6.11s / 70.29s\n",
      ">==================> Plotting Variable: BEME\tElapse / Estimate: 7.98s / 73.38s\n",
      ">==================> Plotting Variable: Beta\tElapse / Estimate: 9.94s / 76.18s\n",
      ">==================> Plotting Variable: C\tElapse / Estimate: 12.21s / 80.21s\n",
      ">==================> Plotting Variable: CF\tElapse / Estimate: 14.66s / 84.28s\n",
      ">==================> Plotting Variable: CF2P\tElapse / Estimate: 17.01s / 86.95s\n",
      ">==================> Plotting Variable: CTO\tElapse / Estimate: 18.85s / 86.73s\n",
      ">==================> Plotting Variable: D2A\tElapse / Estimate: 20.66s / 86.41s\n",
      ">==================> Plotting Variable: D2P\tElapse / Estimate: 22.28s / 85.40s\n",
      ">==================> Plotting Variable: DPI2A\tElapse / Estimate: 23.99s / 84.90s\n",
      ">==================> Plotting Variable: E2P\tElapse / Estimate: 25.57s / 84.01s\n",
      ">==================> Plotting Variable: FC2Y\tElapse / Estimate: 27.11s / 83.15s\n",
      ">==================> Plotting Variable: IdioVol\tElapse / Estimate: 28.64s / 82.34s\n",
      ">==================> Plotting Variable: Investment\tElapse / Estimate: 30.21s / 81.74s\n",
      ">==================> Plotting Variable: Lev\tElapse / Estimate: 31.74s / 81.10s\n",
      ">==================> Plotting Variable: LME\tElapse / Estimate: 33.19s / 80.34s\n",
      ">==================> Plotting Variable: LT_Rev\tElapse / Estimate: 34.76s / 79.94s\n",
      ">==================> Plotting Variable: LTurnover\tElapse / Estimate: 36.37s / 79.67s\n",
      ">==================> Plotting Variable: MktBeta\tElapse / Estimate: 38.02s / 79.50s\n",
      ">==================> Plotting Variable: NI\tElapse / Estimate: 39.67s / 79.33s\n",
      ">==================> Plotting Variable: NOA\tElapse / Estimate: 41.36s / 79.28s\n",
      ">==================> Plotting Variable: OA\tElapse / Estimate: 43.03s / 79.18s\n",
      ">==================> Plotting Variable: OL\tElapse / Estimate: 44.71s / 79.10s\n",
      ">==================> Plotting Variable: OP\tElapse / Estimate: 46.28s / 78.85s\n",
      ">==================> Plotting Variable: PCM\tElapse / Estimate: 47.82s / 78.55s\n",
      ">==================> Plotting Variable: PM\tElapse / Estimate: 49.35s / 78.27s\n",
      ">==================> Plotting Variable: PROF\tElapse / Estimate: 51.00s / 78.19s\n",
      ">==================> Plotting Variable: Q\tElapse / Estimate: 52.71s / 78.22s\n",
      ">==================> Plotting Variable: r2_1\tElapse / Estimate: 54.39s / 78.19s\n",
      ">==================> Plotting Variable: r12_2\tElapse / Estimate: 55.99s / 78.05s\n",
      ">==================> Plotting Variable: r12_7\tElapse / Estimate: 57.62s / 77.96s\n",
      ">==================> Plotting Variable: r36_13\tElapse / Estimate: 59.31s / 77.96s\n",
      ">==================> Plotting Variable: Rel2High\tElapse / Estimate: 61.15s / 78.13s\n",
      ">==================> Plotting Variable: Resid_Var\tElapse / Estimate: 62.91s / 78.21s\n",
      ">==================> Plotting Variable: RNA\tElapse / Estimate: 64.64s / 78.25s\n",
      ">==================> Plotting Variable: ROA\tElapse / Estimate: 66.21s / 78.10s\n",
      ">==================> Plotting Variable: ROE\tElapse / Estimate: 67.75s / 77.91s\n",
      ">==================> Plotting Variable: S2P\tElapse / Estimate: 69.28s / 77.73s\n",
      ">==================> Plotting Variable: SGA2S\tElapse / Estimate: 70.79s / 77.53s\n",
      ">==================> Plotting Variable: Spread\tElapse / Estimate: 72.28s / 77.33s\n",
      ">==================> Plotting Variable: ST_REV\tElapse / Estimate: 73.82s / 77.18s\n",
      ">==================> Plotting Variable: SUV\tElapse / Estimate: 75.36s / 77.04s\n",
      ">==================> Plotting Variable: Variance\tElapse / Estimate: 76.88s / 76.88s\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1b304e470>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plotReturnDecile(dl_test, w_test, value=value_test, plotPath='model_char_shift/model_FFN/plot')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Plot Weight As Function of Firm Characteristics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      ">==================> Plotting Variable: ST_REV\tElapse / Estimate: 0.95s / 0.95s\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAh0AAAGACAYAAAAJcNDnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAIABJREFUeJzt3Xu0XVV96PFvRJCImAg+gKRK8OpPi1EDFS3YgIKl6K1FtIJFAbUMfFYK4sXH8CJeJEDx0dZqoVXKQ9ErGrxwvQjh0QqKQhCt0p8PopQAipWIQECBc/9Y6+DhZJ999mPttffa+/sZ44xF1lxznXnWOGT/MufvN9eCqakpJEmSBu0Rwx6AJEmaDAYdkiSpFgYdkiSpFgYdkiSpFgYdkiSpFo8c9gDGWUQ8CngecCvwwJCHI0lSHTYDtge+lZn3zWww6Bis5wH/NuxBSJI0BH8EfG3mCYOOwboV4JxzzmG77bYb9lgkSRq42267jYMPPhjKz8CZDDoG6wGA7bbbjqVLlw57LJIk1WmTtAITSSVJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0a8+6ViFgGnArsVZ66ADg6M2+vol9E7Au8D9gVeBD4BvC+zPxGFeOQJGnSNWKmIyK2BS4DXgCcRPGh/3Lg4ojYot9+EbEn8BVgMfBe4APAU4ErImK3fschSZKaM9NxFLAUWJ6ZNwBExNXAxcChwOl99vso8J/A8zPznvK6M4EbgBOAl/Q5DkmSJl4jZjqAg4DLpz/oATLzEiDLtp77RcTjgOcAn58OOMrrfgZcAexewTgkSZp4Ix90lEHBTsC1LZrXUuRg9NPvTiCAj7S47vHA/f2MQ5IkFUY+6ACWlMf1LdpuBRZFxKJe+2XmA5n5w8y8ZeYFEfFsYA/gqj7HIUmSaEbQsXV5vKdF28byuFWF/YiIxwBnln9c1e/9JElSM4KOBeVxqs01rdp66hcRjwa+TJHnsSozr+hzHJIkiWYEHXeVx4Ut2qbP3VlFv4hYDHwVeBHwKYry2X7HIUmSaEbJ7E3lcfsWbTsAGzLz7n77RcQTgYuA5wKnAW/KzJkzF72OQ5Ik0YCZjszcAKwDdmnRvAK4pt9+EbE1vws4PpKZR8wKOHoehyRJKox80FE6D9gnIp4xfSIi9qEodT23gn4fpwg4PpaZRw1gHJIkTbwmLK8AnAwcAqyJiFOBLYF3UeyZcTZAROxEsZHXVZl5Yxf9ngm8DtgAfDsiXjv7m2fm2Z3eT5IktdaImY7yZWorgeuB44EjgdXAfpl5X3nZSuCs8thNvz3L42Lg0+U9Zn91cz9JktTCgqkpqzwHJSJ2BNatWbOGpUuXDns4kiQN3M0338zee+8NsCwzfzKzrREzHZIkqfkMOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0MOiRJUi0eOewBqHurr1vPKRclt2zYyA6LF3LMvsH+K5YMe1iSJLVl0NEwq69bz7u/+F02/vYBANZv2Mi7v/hdAAMPSdJIc3mlYU65KB8KOKZt/O0DnHJRDmlEkiR1xqCjYW7ZsLGr85IkjQqDjobZYfHCrs5LkjQqDDoa5ph9g4Wbb/awcws334xj9o2H/rz6uvXssepSlh17IXusupTV162ve5iSJG3CRNKGmU4Wnat6xURTSdKoMuhooP1XLJkzgGiXaGrQIUkaJpdXxoyJppKkUWXQMWZMNJUkjSqDjjFjoqkkaVQ1JqcjIpYBpwJ7lacuAI7OzNur7hcRpwFPz8y9WrR9E3hei27nZear2v8Ug2eiqSRpVDUi6IiIbYHLgC2AkyjGfQzw7IjYLTN/U1W/iHgjcDhwRYu2BcDvA6uB82Y1/7S3n656JppKkkZRI4IO4ChgKbA8M28AiIirgYuBQ4HT++0XEZsB7wWOazOOHYGtgPMz8+zef5zhMdFUkjQsTcnpOAi4fDpwAMjMS4As2/rqFxFbAmuBDwBnAXMlOexcHm+Yo33kmWgqSRqWkQ86IuJxwE7AtS2a1wK7VtBvS+CxwIGZeShw/xzDeVjQERFbzTf+UWOiqSRpWEY+6ACmEw1affLdCiyKiEV99rsTeFpmfn6esTwL+DXw4Yj4NXBXRPw4ItrNtoyU/Vcs4cQDlrNk8UIWAEsWL+TEA5Zvkmi6fsNGpvhdoqmBhySpX03I6di6PN7Tom06EWEr4Fe99svMB4EHOxjLzuV9FwOHlMd3AJ+NiM0z86wO7jF0JppKkoahCUHHgvI41eaaVm299mvnNGCzzPz49ImIOBf4d+CUiPhMZj4wZ+8GMNFUkjQoTVheuas8tsp0nD53Z4X95pSZn5wZcJTnNlIknz6Jopy20Uw0lSQNShOCjpvK4/Yt2nYANmTm3RX268XPy+NjKrrf0HSSaCpJUi9GPujIzA3AOmCXFs0rgGuq7DeXiFgSEd+LiPe3aH5GeVzXzT1H0XyJpmB1iySpNyMfdJTOA/aJiOkPdyJiHyCAcwfQbxOZuZ4icfTwiHjsjPs9GTgMuCwzb+vmnqNq/xVLuPLYF7Nu1cu48tgXbxJwWN0iSepFExJJAU6mqBZZExGnUuyr8S6KPTjOBoiInYDdgasy88ZO+3XprcCXgKsi4nSKSpa3Uezr8dbefrRmsbpFktSrRsx0lC9nWwlcDxwPHEnx/pP9MvO+8rKVFAmdK7vs1804VgP7A3dTvMvlaODrwO4zdz0dZ1a3SJJ61ZSZDjIzgZe2aT8DOKPbfnPca8c2becD53dzv3Gyw+KFrG8RYFjdIkmaTyNmOjQ6rG6RJPWqMTMdGg3TeRunXJTcsmEjOyxeyDH7xibJpu3aJUmTyaBDXWu3jfp0dct0sul0dct0P0nS5HJ5RZVqV90iSZpsBh2qlNUtkqS5GHSoUr67RZI0F4MOVcrqFknSXEwkVaWsbpEkzcWgQ5WzukWS1IrLK6qV1S2SNLkMOlQrq1skaXIZdKhWVrdI0uQy6FCtOqluWX3devZYdSnLjr2QPVZdyurr1tc9TEnSAJhIqlrNV91ioqkkjS+DDtWuXXVLu0RTgw5JajaXVzRSTDSVpPFl0KGRYqKpJI0vgw6NFBNNJWl8mdOhkWKiqSSNL4MOjRwTTSVpPBl0qFE6STT1hXKSNJrM6VCjzJdoOr38sn7DRqb43fKLeR+SNHwGHWqU+RJNfaGcJI0ul1fUKPMlmrr8Ikmjy6BDjdMu0XSHxQtZ3yLwmL38YvWLJNXP5RWNFZdfJGl0OdOhseLyiySNLoMOjR2XXyRpNLm8ooni8oskDY8zHZooVSy/SJJ6Y9ChidPP8guY8yFJvXJ5RZphvuUXdzyVpN4ZdEgz7L9iCScesJwlixeyAFiyeCEnHrD8Ycsy5nxIUm9cXpFmabf8YsmtJPXOmQ6pC75wTpJ6Z9AhdaGKktvV161nj1WXsuzYC9lj1aUGJJImhssrUhf6Lbl18zFJk8ygQ+pSPyW37WZCpu9pToikceXyilSh+ZZfOp0JaZcT4vKMpKZqzExHRCwDTgX2Kk9dABydmbdX3S8iTgOenpl7tWjraRyaDPMtv/Q7E+LyjKQma0TQERHbApcBWwAnUYz7GODZEbFbZv6mqn4R8UbgcOCKqsahydJu+eWYfeNhQQN0NxPi8oykJmtE0AEcBSwFlmfmDQARcTVwMXAocHq//SJiM+C9wHEDGIcE9D8TYqKqpCZrSk7HQcDl0x/0AJl5CZBlW1/9ImJLYC3wAeAsYK5F8l7HIT1k/xVLuPLYF7Nu1cu48tgXPywYmC8nZL59QizZlTTKRj7oiIjHATsB17ZoXgvsWkG/LYHHAgdm5qHA/VWNQ+rGfNuwm6gqqcmasLwy/c/AVn/z3QosiohFmfmrPvrdCTwtMzcJNioYh9SVdjkhJqpKarImBB1bl8d7WrRN/+26FTD7w77jfpn5IPDggMYhVcpEVUlN1YSgY0F5nGpzTau2XvtVPQ6pNqOQqGpQImkuTQg67iqPrTLops/dWWG/qsch1aqfmZA6lmcMSqTJNfKJpMBN5XH7Fm07ABsy8+4K+1U9DmlkDDpRdb7qGd/CK022kZ/pyMwNEbEO2KVF8wrgmir7VT0OadQMMlHVnBFJ7TRhpgPgPGCfiHjG9ImI2AcI4NwB9KvrftLIGeQ+Ipb0SpNt5Gc6SicDhwBrIuJUin013kWxZ8bZABGxE7A7cFVm3thpv6rHIY2z+WZCRiFnRNLoakTQkZm3R8RK4CPA8RRlq6uBYzLzvvKylcCngdcDN3bRr+pxSGOtn+UZS3qlydaIoAMgMxN4aZv2M4Azuu03x7127HUc0qQbZs6IMyHSaGtM0CFpPAyzpBecCZGGqbJE0oh4REQsq+p+kibPKLx7RtLgdDzTEREPAK/NzM/OcclhFLkOiyoYl6QJNcx3z4AzIdIgzRl0RMQOwD4zTi0AVkbE5i0ufwRwMG4DLmnABvnuGXNCpMFqN9NxO/Ae4Onln6eAI8qvufxtReOSpK45EyKNtjmDjsz8bUT8MbCMYpbjUuBDwMUtLn8AuL2s7JCkoXEmRBpdbXM6MvMmyneORMTrgX/NzHV1DEySquZMiDRcHSeSZua/DHIgklQHZ0Kk4emmemULYBVwALAd0CqhdCoz3ftDUiPVMRMiTbJuAoRVwJHAjyi2/m4d8ktSgw1yJgRcftFk6yboeA3wZeAVmWlprKSJ0+9MiMsvmnTdBB2LgAsNOCRNsn5mQlx+0aTrJuj4JvCcQQ1EkppuvpmQTpZfpHHWTdBxNLAmIn4AnJuZPx/QmCSpsdrNhMy3/ALmfGi8tdsG/UE23dZ8AcX7VT4SEa26Wb0iSXOYb/nFnA+Nu3YBwpn4LhVJqsx8yy/mfGjctdsG/bAaxyFJE6Hd8os5Hxp3LoVI0ogw50PjrpsdSVvleMx2H/Bz4Grg+Mz8Xh9jk6SJYs6Hxt0jurj2OOAOisDjIuBjwEnAlyh2J70XOA/4DvBS4OqIWF7lYCVpnO2/YgknHrCcJYsXsgBYsnghJx6wvKOcD6kJulleWVB+7ZqZ189siIhlwNeB72fmiRHxJODfgA9QvKtFktQBcz40zrqZ6Tgc+NjsgAOgfN393wFvKf/8M+B04IVVDFKS9PDcjk7OS6Omm6DjscDdbdrvBbad8ec7AP9PkKSKHLNvsHDzzR52bmbOBxR5H3usupRlx17IHqsuZfV16+sepjSnboKOa4A3R8Q2sxsiYhHwJuC6GadfBPy4v+FJkqbNl/MxnWi6fsNGpvhdoqmBh0ZFNzkd7wYuBX4QEWdRvOL+PuDpFG+gfQLwlwAR8RXgj4G/qnS0kjTh2uV8uLmYRl3HQUdmXh0RK4GTgbfz8FmSq4A/z8xvRMQTgWcCH8rMj1c6WknSnEw01ajranOwzLwW2DsiFgM7AZsDN2bm7TOu+TmwY5WDlCTNr5PNxaRh6mlH0szcAKyteCySpD7Mt7kYuKOphqvdW2ZvBI7MzC/P+PN8pjLzqVUNTpLUufleKOeOphq2djMdP+XhJbI34VtnJWmkmWiqUdbuLbMvmvXnvQY+GknSwJhoqmHrKacjIrYHngz8B8V7V+7PzAerHJgkqVq+xVbD1s3mYETEHhFxLXAzRZnsrsBewE0R8erqhydJqsp8O5q6uZgGreOgIyKeB1wCbA18dEbTL4HfAp+JiP2qHZ4kqSq+xVbD1s3yyv8C1lHMbmwF/DVAZl4TEc8BrgTeA3yl6kFKkqrhW2w1TN0sr/wh8OnM3MisKpbMvBM4DXhWhWOTJNXIt9hq0LrK6aB418pctuzhfpKkEeFbbDVo3QQJVwN/0aohIraieNnbt6oYlCSpfr7FVoPWTU7H+4HLI+IK4HyKJZbnR8SzKN4m+xSK19tLkhrKzcU0SB3PdGTm14GXAUuBvwEWACdQVLIsBA7KzMsGMUhJ0vCZaKp+dTzTEREXABcAewPbAE8FNgN+AlyTmfcPYoAzvv8y4FSKfUEox3L0zDfc9tOvi+u+CTyvxbc6LzNf1eGPI0mN41ts1a9ulleWAh8v//v7wP8FLgS+OejdSCNiW+AyYAvgJIpxHwM8OyJ2y8zf9NOvi+sWAL8PrAbOm/XtflrdTyxJo6eTt9hK7XQcdGTmcyPiScC+5dfrKT6YN0TEVykCkK9k5i8GMM6jKIKe5Zl5A0BEXA1cDBwKnN5nv06v25Fij5LzM/PsCn8+SRp5873FVppPV+9eycyfAWcCZ5b/6t+FIgB5A/DnwIPA5lUPEjgIuHw6ICjHcklEZNk2V9DRab9Or9u5PD50nSRNknaJptN8f4vm0tO+GhHxDOBw4EjgCGCnsuk/KhrXzO/1uPL+17ZoXkuxQ2rP/bq8/8OCjrJUWJJUsqxW7XTz7pUjI+K8iPgZ8D3gE8ByivyGVwJPzMzlAxjjdHjc6jf2VmBRRCzqo183938W8GvgwxHxa+CuiPhxRBzUwc8hSWPP97eonW6WVz5MsTfHeopcjk9n5h0DGdXDbV0e72nRNp1GvRXwqx77dXP/ncvrFwOHlMd3AJ+NiM0z86y5fwxJGn+W1aqdboKOv6IoJ90TOAV4Z0T8K3AFcEVmfr/64QHFfiAw630vs7Rq67RfN/c/DdgsM6ereIiIc4F/B06JiM9k5gOtbiBJk8CyWrXTzeZgf5+Zr8rMJ1AkkJ5M8b6VDwLfjYifR8QXBjDGu8pjq9/Y6XN39tGv4/tn5idnBhzluY3AWcCTKMppJWlidfL+Fk2urqpXpmXm9cD15b/yXwK8jWLDrFdUOLZpN5XH7Vu07QBsyMy7e+0XEb3ef6afl8fHzHOdJI21TspqrW6ZXF0FHWWlx4uAF5df06Hr9cCJwFcqHR2QmRsiYh3F7MpsK4Br+unX6XURsQT4KvC5zDx+1nXPKI/r5vlxJGnstSurna5umU42na5ume6n8dZN9cpa4HbgC8DBFHkMbwSWZOYumfnezPzaYIbJecA+Zanu9Hj2oQh6zq2g37zXZeZ6isTRwyPisTOuezJwGHBZZt7W6w8oSZPA6pbJ1u3yykkUsxlfrzlh8mSKapE1EXEqRS7Juyj21jgbICJ2AnYHrsrMGzvt1+V1bwW+BFwVEadTVLK8Dbi/bJMktWF1y2TrJpH0odmMuis0ypeuraRYxjmeYlOy1cB+mXlfedlKioTOlV326+a61cD+wN0UAdjRwNeB3WfuZipJam2uKharWybDgqmpdpWi6kdE7AisW7NmDUuXLh32cCRp6GbndEBR3XLiAcvN6RgTN998M3vvvTfAssz8ycy2nqpXJEnqhdUtk82gQ5JUK6tbJldPL3yTJGkQrG4ZbwYdkqSRYXXLeDPokCSNDKtbxptBhyRpZPjulvFmIqkkaWRY3TLeDDokSSPF6pbx5fKKJKkxrG5pNoMOSVJjWN3SbAYdkqTGsLql2Qw6JEmNYXVLs5lIKklqjE6qWzS6DDokSY3SrroFLKkdZQYdkqSxYUntaDOnQ5I0NiypHW0GHZKksWFJ7Wgz6JAkjQ1LakebQYckaWxYUjvaTCSVJI0NXxg32gw6JEljxRfGjS6XVyRJE8PqluEy6JAkTQyrW4bLoEOSNDGsbhkugw5J0sSwumW4TCSVJE0Mq1uGy6BDkjRRrG4ZHpdXJEkqWd0yWAYdkiSVrG4ZLIMOSZJKVrcMlkGHJEklq1sGy0RSSZJKnVS3qHcGHZIkzdCuugUsqe2HQYckSR2ypLY/5nRIktQhS2r7Y9AhSVKHLKntj0GHJEkdsqS2PwYdkiR1yJLa/phIKklSh3xhXH8aE3RExDLgVGCv8tQFwNGZeXsV/aq+TpI0nnxhXO8asbwSEdsClwEvAE6i+NB/OXBxRGzRb7+qr5MkTSarW9prykzHUcBSYHlm3gAQEVcDFwOHAqf32a/q6yRJE8jqlvYaMdMBHARcPv1BD5CZlwBZtvXbr+rrJEkTyOqW9kY+6IiIxwE7Ade2aF4L7NpPv6qvkyRNLqtb2hv5oAOYzrxZ36LtVmBRRCzqo1/V10mSJtT+K5Zw4gHLWbJ4IQuAJYsXcuIByzepbtlj1aUsO/ZC9lh1Kauva/WxMp6akNOxdXm8p0Xb9CLZVsCveuxX9XWzxyFJmiBWt8ytCTMdC8rjVJtrWrV12q/q6yRJamnSq1uaEHTcVR5bZeFMn7uzj35VXydJUkuTXt3ShKDjpvK4fYu2HYANmXl3H/2qvk6SpJYmvbpl5IOOzNwArAN2adG8Arimn35VXydJ0lwmvbpl5IOO0nnAPhHxjOkTEbEPEMC5FfSr+jpJkjbRSXXLOGtC9QrAycAhwJqIOBXYEngXxZ4ZZwNExE7A7sBVmXljp/0GdJ0kSS21q26B8X5hXCNmOsqXqa0ErgeOB44EVgP7ZeZ95WUrgbPKYzf9Kr9OkqReTJfUrt+wkSl+V1I7Lnt5LJiasspzUCJiR2DdmjVrWLp06bCHI0kacXusupT1LSpZlixeyJXHvngII+rezTffzN577w2wLDN/MrOtETMdkiRNgnEvqTXokCRpRIx7Sa1BhyRJI2LcS2qbUr0iSdLYm65SaVe90uTqFoMOSZJGyDi/MM7lFUmSGqLpL4wz6JAkqSGaXt1i0CFJUkM0vbrFoEOSpIZoenWLiaSSJDVEJ9Uto8ygQ5KkBmnyC+MMOiRJGhOjXlJrTockSWNi1EtqDTokSRoTo15Sa9AhSdKYGPWSWoMOSZLGxKiX1JpIKknSmBj1klqDDkmSxsgol9QadEiSNCGGXVJrTockSRNi2CW1Bh2SJE2IYZfUGnRIkjQhhl1Sa9AhSdKEGHZJrYmkkiRNiGGX1Bp0SJI0QeYrqR0kl1ckSVItDDokSVItDDokSVItDDokSVItDDokSVItDDokSVItDDokSVItDDokSVItDDokSVItDDokSVItDDokSVItDDokSVItDDokSVItDDokSVItGvNq+4h4M/AO4CnAj4ATMvPcKvt2cl1EvBS4cI5vtTwz/72zn0iSpMnSiJmOiHgn8A/Ad4AjgfXAZyPiwKr6dvE9dgamgMOA1836+s/efkJJksbfgqmpqWGPoa2IWAzcDJyfmQeX5x4BXA7sCCzLzAf66dvN94iIfwH2zMwdOxj7jsC6NWvWsHTp0q5/dkmSmubmm29m7733huKz8ycz25ow0/GnwFbAJ6ZPZOaDFLMSvwfsXkHfbr7HzsANPf80kiRNqCYEHX9QHtfOOr92Vns/fTu6LiIWAM8Evl/+ecuIaExejCRJw9SEoGMJcEdm3jPr/K3l8ckV9O30up2ARwNPjYjvAvcAd0fEuRHx+Pl/FEmSJtfQ/pUeEdvNc8ldmXkXsDXFh/tsG8vjVm3u0WnfTq/buTy+AFgFrANeSJF4unNE7JaZG5EkSZsY5tLArfO0nwC8D1hAUS0yl3Ztnfbt9LobgeOBczLzB+W58yPiR8AngTcCf9/mPpIkTaxhBh2Hz9N+XXm8C1jYon363J1t7tFp346uK/fgaLUPx6cogo0XYdAhSVJLQws6MvOfOrz0JmCbiHhUZt434/wO5XF9BX37+R5k5m8j4g7gMe2ukyRpkjUhkXQtxfLHc2edX1Eev1VB346ui4gPRsSNEbH1zIsiYhvgCRQ5HpIkqYUmBB0XAvcCb58+UW7c9Rbgp8A3Kujb6XU3AcuAv5z1fd5fHj/T+Y8lSdJkGfk9JjLzvyJiFXBcGQhcCrySomrkwJm7kUbE/mWf1d307eJ7nEGRi3JyRDwN+C6wD3AA8I+Z+a8DfBSSJDXayAcdpeOBu4G3Aq8AfgC8OjO/MOu6j5bH1T30nfe6MnfjT4APUQQab6SoaDkK+FifP6MkSWNt5N+90mS+e0WSNGma/u4VSZI0Bgw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLR457AGMuc0AbrvttmGPQ5KkWsz4zNtsdptBx2BtD3DwwQcPexySJNVte+DHM08YdAzWt4A/Am4FHhjyWCRJqsNmFAHHt2Y3LJiamqp/OJIkaeKYSCpJkmrh8ooqERFvBt4BPAX4EXBCZp7bw30+B7w6MxdUPMSR1+szjIjHAh8EXgk8EbgFOAf4QGb+ZnAjHp6IWAacCuxVnroAODozbx9Ev3HTx/PbF3gfsCvwIPAN4H2Z+Y3BjXb0VPF7FBHPBq4BPpSZx1U9xlHlTIf6FhHvBP4B+A5wJLAe+GxEHNjlfQ4AXl39CEdfr88wIhYAXwTeCpwP/BVwBfBu4HODHPOwRMS2wGXAC4CTKP7yfzlwcURsUXW/cdPH89sT+AqwGHgv8AHgqcAVEbHboMc9Kqr4PYqIRwJnAJsPaJgjy5kO9SUiFgPHAZ/JzIPLc6cDlwOnRMQXMnPeJNryf+RPAL8BJuYDAPp+hi8H9gbelpkfL899MiJuBt4TEXtk5pUD/QHqdxSwFFiemTcARMTVwMXAocDpFfcbN70+h48C/wk8PzPvKfudCdwAnAC8ZMDjHhVV/B69G9h5YCMcYc50qF9/CmxFETAAkJkPUvyr/feA3Tu8z98DvwDOq3qADdDPM9yrPJ4x6/zny2Onz79JDgIun/4LHyAzLwGybKu637jp+jlExOOA5wCfnw44yn4/o5hZG8ffs7n09XsUEcsplqg+OLARjjCDDvXrD8rj2lnn185qn1NE7E+xrPIGipmOSdPPM/wgsCIz7551/vHl8f4+xzZSyg+/nYBrWzSvpcg1qKzfuOmVKRvQAAAIUUlEQVTjOdwJBPCRFm2PZ8x+z+bS7+9RuazyaYpZkbMrH2ADGHSoX0uAO2b+66d0a3l8crvOEbEN8Engo5l59QDG1wQ9P8PM/GVmfrtF05vL41UVjG+ULCmP61u03QosiohFFfYbNz09h8x8IDN/mJm3zDxfJkPuwfj9ns2l39+j/wE8DXhT1QNrCnM61FJEbDfPJXdl5l3A1sDsD0uAjeVxq3nu83fArymmG8dKjc9w9vd9HUUly2VjGMhtXR7ne16/qqjfuKnsOUTEY4Azyz+u6n9ojdDz84uInYH3U+Rf3RwROw5khCPOoENzuXWe9hMoAoUFQLsd5uZsi4iXA68BXpSZG+e6rsEG/gxnK5/pPwO3AYd12q9Bpkupu31evfYbN5U8h4h4NPBlijyPEzPzigrG1gQ9Pb+I2Iwi7+prmTkpCcstGXRoLofP035debwLWNiiffrcna06l2ujnwQ+A3wvIqZzEB5Vtj8euLecCWiqgT7D2SLiNcC/UMwc7ZeZN3XSr2Gmfx+6fV699hs3fT+HstrqAopllU9RlM9Oil6f3zHAs4EXzvi77nHl8dHluV+WCeRjzaBDLWXmP3V46U3ANhHxqMy8b8b5Hcpjq7VPKP6FtD1wcPk12+0UH6CHdTiOkVPDM3xIRBxBUe3yS2CfzLy+q8E2x3QgtX2Lth2ADS2SavvpN276eg4R8UTgIuC5wGnAmzJzEmaIpvX6/P6EYiuAb7ZoO6b8Wgb8pIIxjjQTSdWvtRRTjs+ddX5FedzkhT+l6ynq+md/fbVsfwlwcqUjHV29PkMAIuIQilmj24CVYxxwkJkbgHXALi2aV1Ds8FhZv3HTz3OIiK35XcDxkcw8YsICjn6e39Fs+nfda8u2s8o/39a663hxpkP9uhC4F3g7cDVARDwCeAvwU4ptkjeRmXcAl8w+HxGvLds3aRtjPT3D8rrfp/gX5+3AXpn5w4GPdvjOA46MiGdk5n8ARMQ+FCWdpwyg37jp9Tl8nCLg+FhmHjX4YY6srp9fZm5SYjsjkfTGSfr7zqBDfcnM/4qIVcBx5QflpRSVEy8EDpy5k2a5HweZuXoogx1RfT7D/0mRB/P/gOdHxPNn3f47mfmdQf8MNTsZOARYExGnAlsC76LYO+FsgIjYiWLDqqsy88ZO+02Irp9fRDwTeB2wAfj29D8OZsrMSXmGvf7+CYMOVeN44G6K93+8AvgBxUvbvjDruo+WR4OOTfX6DPcsj68rv2b7AMX7XMZGZt4eESspNqo6nqJ8cTVwzIycmJUUmzC9Hrixi35jr8fnN/17trg838pEBB29/v6psGBqaqKW5CRJ0pCYSCpJkmph0CFJkmph0CFJkmph0CFJkmph0CFJkmph0CFJkmph0CFJkmrh5mCSBiIiHgX8NfAa4L8BDwA/BD5PsZX2vRFxHMWuqvO5IjP36uJ7H0brTazuB+6geJ/N/8zMa2b06WQs12fmcyPiUxQbPz0/M1u9xGv6debrgVsys9W7OqSJY9AhqXIR8UiKrdn/EDgT+EeKv2/+CDgReHlEvBj4IvCjGV2fCbwH+FLZNu1nPQ7lNODfZvz5URRvOD4C2DMinpOZP57V50PADXPc75fl8RyKoONVtH5zKMCLgCcxOS8ulOZl0CFpEF4N7AW8MjNnBg9/GxHHUHwQvyEzP8GMbdojYi+KoOM7Fb3L4+ut7hMRXwM+R/H2z7fMar44My+f576XAbdQBB3vmuOaAylmdz7bzYClcWZOh6RB2L08frVF2z8Av6WYBRmW/03xzowX9NI5Mx+kCCaWRcSus9sjYnPgAODSzLy1n4FK48SgQ9Ig/Lo8HjG7ITPvBh6TmYfUO6SHjWEK2Ags6OM255THV7VoewmwDRPyEjSpUy6vSBqEs4GjgL+JiNdT5GisoVjuuC8zfzPMwUXEHwDbAl9u0bwoIh4/R9c7p8eemddFxPcpgo53z7ruIIqZlC8i6SEGHZIql5nfi4gDgE8BO5df7wPujogvA8dl5g9qGMpjZgUQC4FdgFMpZjpWteizus39XjGr/RzghIh4bmZ+Gx6q2vkz4PzMvKufwUvjxuUVSQORmRcCT6FIqDwLuBXYiqKE9vqI2LOGYfwdcPuMr5soZh9uAXafI/B5J8XySKuvK2ddew4wxcOXWF4KPBaXVqRNONMhaWAy816KfTk+DxARu1B8qL8G+CRFiewgnUKRzPoIilLZd1PsFfLazLxpjj7XdlC9AkBm/jQirqQIOt5Xnj4I+Dmtk2iliWbQIalSEbEVRdnrtbPKZcnMtcBfRMRiYL+I2DYz/2uAw/l+Zl5S/vdXI+Iy4GvA5RGxW2b+ooLvcQ7wiYhYDtwI/HfgnzPz/gruLY0Vl1ckVe1eitmMt7e55nsUyxIbaxlRqdyB9FhgGfBPFd328xQlwAcA+wGPxqUVqSWDDkmVyswHKDbe2isiXju7PSK2oViOuCQz76l7fMDHKHIz/iwiDuz3Zpn5S+ArwMuAPwV+MNfW6NKkc3lF0iAcBewGnFUGHhcBv6J4B8thwBbAW4cxsMyciogjgOuAj0XEVzPzjhmXvCQilra5xZfKvUZmOpsi0Ho68OFqRyyND4MOSZXLzF+UO3X+NUX56Psplh1uoageOWGYO3WWJb0nA++lKJ99w4zm98zTfRkwO+j4PxQboi3CpRVpTgumpqaGPQZJkjQBzOmQJEm1cHlF0siLiM2AJ3R4+a8ys9aqGEmdMeiQ1AS/B6zr8NrXA2cMbiiSemXQIakJbqPYhrwT3xvkQCT1zkRSSZJUCxNJJUlSLQw6JElSLQw6JElSLQw6JElSLQw6JElSLf4/IeLYaV0mrn8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x121440d68>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model.plotConditionalReturn1DWithChar(sess, dl_test, 'ST_REV', label='weight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">==================> Please enter the idx for the plot you want! \n",
      "Please select from 1.1D 2.2D Contour 3.2D Contour Slice 4. 1D Curves: 4\n",
      "Enter variable idx for x (0 - 45, 46 - 45): 32\n",
      "Enter variable idx for y (0 - 45, 46 - 45): 43\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x16d433a90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model.plotConditionalReturn(sess, dl_test, plotPath='model_char_shift/model_FFN/plot/2D', label='weight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">==================> Please enter the idx for the plot you want! \n",
      "Please select from 1.1D, 2.2D Contour 3.2D Contour Slice 4.1D Curves (2 Variables) 5.1D Curves (3 Variables): 5\n",
      "Enter variable idx for x (0 - 45, 46 - 45): 43\n",
      "Enter variable idx for y (0 - 45, 46 - 45): 18\n",
      "Enter variable idx for z (0 - 45, 46 - 45): 4\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x174d5db38>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model.plotConditionalReturn(sess, dl_test, plotPath='model_char_shift/model_FFN/plot/3D', label='weight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# model.plotConditionalReturn1DWithChar(sess, dl_test, 'All', plotPath='model_char_shift/model_FFN/plot/1D', label='weight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">==================> Please enter the idx for the plot you want! \n",
      "Please select from 1.1D, 2.2D Contour 3.2D Contour Slice 4.1D Curves (2 Variables) 5.1D Curves (3 Variables): 2\n",
      "Enter variable idx for x (0 - 45, 46 - 45): 18\n",
      "Enter variable idx for y (0 - 45, 46 - 45): 11\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x16a09c9e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model.plotConditionalReturn(sess, dl_test, plotPath='model_char_shift/model_FFN/plot/2D', label='weight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">==================> Please enter the idx for the plot you want! \n",
      "Please select from 1.1D, 2.2D Contour 3.2D Contour Slice 4.1D Curves (2 Variables) 5.1D Curves (3 Variables): 3\n",
      "Enter variable idx for x (0 - 45, 46 - 45): 18\n",
      "Enter variable idx for y (0 - 45, 46 - 45): 11\n",
      "Enter variable idx for z (0 - 45, 46 - 45): 43\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x16ac9ca20>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model.plotConditionalReturn(sess, dl_test, plotPath='model_char_shift/model_FFN/plot/3D', label='weight')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Plot Beta As Function of Firm Characteristics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">==================> Please enter the idx for the plot you want! \n",
      "Please select from 1.1D, 2.2D Contour 3.2D Contour Slice 4.1D Curves (2 Variables) 5.1D Curves (3 Variables): 3\n",
      "Enter variable idx for x (0 - 45, 46 - 45): 18\n",
      "Enter variable idx for y (0 - 45, 46 - 45): 11\n",
      "Enter variable idx for z (0 - 45, 46 - 45): 43\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/1/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/2/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/3/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/4/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/5/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/6/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/7/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/8/model-best\n",
      ">==================> Restored checkpoint\n",
      "INFO:tensorflow:Restoring parameters from model_char_shift/model_FFN/9/model-best\n",
      ">==================> Restored checkpoint\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x16bc45240>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model.plotConditionalReturn(sess, dl_test, plotPath='model_char_shift/model_FFN/plot/beta', label='beta')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}