example_neuromorphic_torch_layers (1).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "ba4d0c20-47c5-4edc-9a46-e96ce270b9e4",
   "metadata": {},
   "source": [
    "# Turning Neuromorphic Dynex Chips into Torch Layers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "ea16bae0-d582-44c2-a32b-b0733e6ecf29",
   "metadata": {},
   "outputs": [],
   "source": [
    "import math\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.metrics import classification_report\n",
    "from torchvision.transforms import ToTensor\n",
    "from torch.utils.data import random_split\n",
    "from torch.utils.data import DataLoader\n",
    "from torch.nn import Module\n",
    "from torch.nn import Conv2d\n",
    "from torch.nn import Linear\n",
    "from torch.nn import MaxPool2d\n",
    "from torch.nn import ReLU\n",
    "from torch.nn import LogSoftmax\n",
    "from torch import flatten\n",
    "from torch.optim import Adam"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eba0e774-0668-4bfe-bda6-3e22cf8ebe2a",
   "metadata": {},
   "source": [
    "## Import dnx layer classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "86fc9d73-965c-4d24-872b-142e2fdc9e74",
   "metadata": {},
   "outputs": [],
   "source": [
    "from HybridQRBM.pytorchdnx import dnx\n",
    "from HybridQRBM.optimizers import RBMOptimizer\n",
    "from HybridQRBM.samplers import DynexSampler"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "db156edb-1a7b-4013-a0a3-025e8f05564a",
   "metadata": {},
   "source": [
    "## Parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "ac5d5952-4df2-4d80-9be2-40b3d4372ab4",
   "metadata": {},
   "outputs": [],
   "source": [
    "# define training hyperparameters\n",
    "INIT_LR    = 1e-3\n",
    "BATCH_SIZE = 12000 \n",
    "EPOCHS     = 2 \n",
    "device = \"cpu\" # no GPU used for Dynex only"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b9f814a6-8875-402c-b21d-f664c4185ce6",
   "metadata": {},
   "source": [
    "## Define Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "9b919152-4c82-4086-963c-5b1534e2f2a7",
   "metadata": {},
   "outputs": [],
   "source": [
    "optimizer = RBMOptimizer(\n",
    "                learning_rate=0.05,\n",
    "                momentum=0.9,\n",
    "                decay_factor=1.00005,\n",
    "                regularizers=()\n",
    "            );\n",
    "\n",
    "sampler = DynexSampler(mainnet=True, #False, \n",
    "               num_reads=25000, #100, \n",
    "               annealing_time=1000, \n",
    "               debugging=False, \n",
    "               logging=True, \n",
    "               num_gibbs_updates=1, \n",
    "               minimum_stepsize=0.002);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "b520b594-ea68-4c88-962d-840c7f1ffe93",
   "metadata": {},
   "outputs": [],
   "source": [
    "class QModel(nn.Module):\n",
    "    def __init__(self, n_hidden, sampler, optimizer, trainSteps):\n",
    "        super().__init__();\n",
    "        # Dynex Neuromporphic layer\n",
    "        self.dnxlayer = dnx(n_hidden, sampler=sampler, optimizer=optimizer, steps_per_epoch=trainSteps); \n",
    "\n",
    "    def forward(self, x):\n",
    "        x = self.dnxlayer(x);\n",
    "        return x;"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5f207539-7361-4953-a373-57120ff140b5",
   "metadata": {},
   "source": [
    "## Load Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "506946c2-b543-4d28-8576-1b72ef645665",
   "metadata": {},
   "outputs": [],
   "source": [
    "# conversion function from grayscale to black & white\n",
    "from torchvision import transforms\n",
    "class ThresholdTransform(object):\n",
    "  def __init__(self, thr_255):\n",
    "    self.thr = thr_255 / 255.  # input threshold for [0..255] gray level, convert to [0..1]\n",
    "\n",
    "  def __call__(self, x):\n",
    "    return (x > self.thr).to(x.dtype)  # do not change the data type\n",
    "\n",
    "data_transformer = transforms.Compose([\n",
    "        #transforms.Grayscale(),\n",
    "        transforms.ToTensor(),\n",
    "        ThresholdTransform(thr_255=240)\n",
    "    ])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "09636e29-8e5c-4572-854b-7d6169fac19d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[INFO] MNIST dataset lodaed\n"
     ]
    }
   ],
   "source": [
    "# load the MNIST dataset\n",
    "from torchvision.datasets import MNIST\n",
    "trainData = MNIST(root=\"data\", train=True, download=True,\n",
    "\ttransform=data_transformer)\n",
    "testData = MNIST(root=\"data\", train=False, download=True,\n",
    "\ttransform=data_transformer) \n",
    "print(\"[INFO] MNIST dataset lodaed\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "03cdfcce-6001-45ef-b3bf-694455b1b6f3",
   "metadata": {},
   "outputs": [],
   "source": [
    "# initialize the train, validation, and test data loaders\n",
    "trainDataLoader = DataLoader(trainData, shuffle=True, batch_size=BATCH_SIZE)\n",
    "testDataLoader = DataLoader(testData, batch_size=BATCH_SIZE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "be186985-a278-44fb-8976-48c35fcc9642",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5 train steps per epoch\n"
     ]
    }
   ],
   "source": [
    "# calculate steps per epoch for training and validation set\n",
    "trainSteps = len(trainDataLoader.dataset) // BATCH_SIZE\n",
    "print(trainSteps,'train steps per epoch');"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eecbde1b-f124-4df5-a6e1-31befc31dc54",
   "metadata": {},
   "source": [
    "## Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "653ac4f3-89b3-409f-ba17-55584ebbd90a",
   "metadata": {},
   "outputs": [],
   "source": [
    "#img_dim = len(trainData.data[0]);\n",
    "n_hidden = 60;\n",
    "model = QModel(n_hidden, sampler, optimizer, trainSteps);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "86f70235-caf3-4773-9d8b-45c2fc8a4836",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "╭────────────┬───────────┬───────────┬─────────┬───────┬─────────┬───────┬──────┬──────────┬────────────╮\n",
      "│   DYNEXJOB │   ELAPSED │   WORKERS │   CHIPS │     ✔ │   STEPS │   LOC │    ✔ │   ENERGY │          ✔ │\n",
      "├────────────┼───────────┼───────────┼─────────┼───────┼─────────┼───────┼──────┼──────────┼────────────┤\n",
      "│       5024 │      0.45 │         2 │  455104 │ 49984 │    1000 │  7583 │ 7743 │  5743084 │ 5759890.00 │\n",
      "╰────────────┴───────────┴───────────┴─────────┴───────┴─────────┴───────┴──────┴──────────┴────────────╯\n",
      "╭───────────┬───────────────┬─────────┬───────┬────────────┬───────────┬───────────────┬──────────╮\n",
      "│    WORKER │       VERSION │   CHIPS │   LOC │     ENERGY │   RUNTIME │   LAST UPDATE │   STATUS │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959484 │ 2.3.0.SRB.234 │   25000 │  8258 │ 5743084.00 │  0.58 min │    15.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959398 │ 2.3.0.SRB.235 │   25000 │  8260 │ 5746740.00 │  1.28 min │     3.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959388 │ 2.3.0.OZM.123 │   24992 │  8434 │ 5757046.00 │  1.05 min │    19.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959404 │ 2.3.0.OZM.123 │   24992 │  8393 │ 5758237.00 │  1.07 min │    15.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959389 │ 2.3.0.OZM.124 │   24992 │  7743 │ 5759890.00 │  1.05 min │    19.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959403 │ 2.3.0.OZM.124 │   24992 │  7717 │ 5762077.00 │  1.05 min │    16.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959434 │ 2.3.0.OZM.124 │   24992 │  7908 │ 5764308.00 │  1.08 min │     2.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959409 │ 2.3.0.OZM.124 │   24992 │  7775 │ 5765150.00 │  1.07 min │    14.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959385 │ 2.3.0.OZM.124 │   17056 │  7583 │ 5765405.00 │  1.05 min │    20.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959396 │ 2.3.0.OZM.123 │   24992 │  8403 │ 5767310.00 │  1.05 min │    17.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959411 │ 2.3.0.OZM.123 │   24992 │  8431 │ 5768056.00 │  1.05 min │    14.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959393 │ 2.3.0.OZM.124 │   24992 │  7865 │ 5768848.00 │  0.05 min │    78.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959395 │ 2.3.0.OZM.124 │   24992 │  7996 │ 5768893.00 │  1.05 min │    17.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959437 │ 2.3.0.SRB.235 │   25000 │  8250 │ 5768996.00 │  0.35 min │    45.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959418 │ 2.3.0.SRB.235 │   25000 │  8274 │ 5769116.00 │  0.27 min │    60.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959397 │ 2.3.0.OZM.124 │   24992 │  7968 │ 5769758.00 │  1.05 min │    17.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959387 │ 2.3.0.OZM.123 │    5504 │  8465 │ 5773971.00 │  1.05 min │    20.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959410 │ 2.3.0.OZM.124 │    7648 │  7756 │ 5774034.00 │  1.07 min │    13.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959394 │ 2.3.0.OZM.124 │   24992 │  7689 │ 5775024.00 │  0.05 min │    78.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "├───────────┼───────────────┼─────────┼───────┼────────────┼───────────┼───────────────┼──────────┤\n",
      "│ 133959480 │ 2.3.0.OZM.124 │   24992 │  8089 │ 5775762.00 │  0.27 min │    35.00s ago │  \u001b[1;31mSTOPPED\u001b[0m │\n",
      "╰───────────┴───────────────┴─────────┴───────┴────────────┴───────────┴───────────────┴──────────╯\n",
      "[DYNEX] FINISHED READ AFTER 0.45 SECONDS\n",
      "[DYNEX] COMPUTING FILE c643091ab8cc04a7080baf2ad0f637c9.bin REMOVED\n",
      "[DYNEX] SAMPLESET LOADED\n",
      "DynexQRBM PyTorch Layer | applying sampling result... 60000 x 784\n",
      "DynexQRBM PyTorch Layer | SME: 0.007722 ACCURACY: 99.227823%\n"
     ]
    }
   ],
   "source": [
    "for e in range(0, EPOCHS):\n",
    "\tprint('EPOCH',e,'of',EPOCHS);\n",
    "\t# set the model in training mode\n",
    "\tmodel.train()\n",
    "\t# loop over the training set\n",
    "\tfor (x, y) in trainDataLoader:\n",
    "\t\t# send the input to the device\n",
    "\t\t(x, y) = (x.to(device), y.to(device))\n",
    "\t\t# perform a forward pass and calculate the training loss\n",
    "\t\tpred = model(x);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "de55dc1a-313f-4bb9-99b1-983569fad054",
   "metadata": {},
   "outputs": [],
   "source": [
    "dnxModel = model.dnxlayer.errors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "d6a2894e-4c7a-4a68-9114-8f4f73a57a2d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjcAAAGwCAYAAABVdURTAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8pXeV/AAAACXBIWXMAAA9hAAAPYQGoP6dpAABJD0lEQVR4nO3dd1xT5/4H8M9JIAFkicjSIE5cEOtCtF71iuKstnXUtmrVtvdaO6x1oXW34uzUW63Vau+vda866qKirdJ66wAH4kJBlOFiykrO7w9qagqokfEk4fN+vfLSnDznnO/DCeZjvicnkizLMoiIiIishEJ0AURERETlieGGiIiIrArDDREREVkVhhsiIiKyKgw3REREZFUYboiIiMiqMNwQERGRVbERXUBl0+v1uHHjBpycnCBJkuhyiIiI6AnIsozMzEz4+PhAoXj0ezNVLtzcuHEDGo1GdBlERET0FBITE1G7du1Hjqly4cbJyQlA0Q/H2dlZcDVERET0JDIyMqDRaAyv449S5cLNg1aUs7Mzww0REZGFeZJTSnhCMREREVkVhhsiIiKyKgw3REREZFWq3Dk3RERkXfR6PfLz80WXQeVApVI99mPeT4LhhoiILFZ+fj7i4+Oh1+tFl0LlQKFQoG7dulCpVGXaDsMNERFZJFmWcfPmTSiVSmg0mnL5Hz+J8+Aiuzdv3oSvr2+ZLrTLcENERBapsLAQOTk58PHxgYODg+hyqBzUrFkTN27cQGFhIWxtbZ96O4y5RERkkXQ6HQCUuYVB5uPBsXxwbJ8Www0REVk0fk+g9SivY8lwQ0RERFZFaLg5fPgw+vbtCx8fH0iShG3btj12ncjISLRs2RJqtRoNGjTA6tWrK7xOIiIishxCw012dja0Wi2WLl36ROPj4+PRu3dvdOnSBadOncLYsWPx+uuvY+/evRVcKRERUcXq3Lkzxo4dK7oMk6xevRqurq4mrfOkb2aUhdBw07NnT3z00Ud4/vnnn2j8smXLULduXSxevBhNmjTB22+/jQEDBuDTTz+t4EqfzOW0LMTfyhZdBhERETp37gxJkjBv3rxij/Xu3RuSJGHmzJmVX1glsKhzbqKiohASEmK0LDQ0FFFRUaWuk5eXh4yMDKNbRdh7Nhm9Pv8FY9efQqGOF5MiIiLxNBpNsdM3kpKSEBERAW9vbzFFVQKLCjfJycnw9PQ0Wubp6YmMjAzcv3+/xHXCw8Ph4uJiuGk0mgqpLbC2C1Q2CkQn3sOKX+IrZB9ERFQ6WZaRk18o5CbLskm1ZmdnY9iwYXB0dIS3tzcWL15s9Lifnx/mzp2LkSNHwsnJCb6+vvj6668Nj3/33XdwdHTExYsXDcveeustNG7cGDk5OYZlffr0wa1bt3DkyBHDsjVr1qB79+7w8PAw2ufdu3cxbNgwVK9eHQ4ODujZs6fR9oGiNpSvry8cHBzw/PPP4/bt28Xmtn37drRs2RJ2dnaoV68eZs2ahcLCQpN+PmVl9RfxCwsLw7hx4wz3MzIyKiTgeLvYY3qfppiwKQaf7r+Ark080MjTqdz3Q0REJbtfoEPT6WLOwTw3OxQOqid/SZ0wYQIOHTqE7du3w8PDA1OmTMGJEyfQokULw5jFixdjzpw5mDJlCjZt2oTRo0ejU6dO8Pf3x7Bhw7Bz50688sorOHr0KPbu3YtvvvkGUVFRRhc0VKlUeOWVV/Dtt9+iQ4cOAIoCyoIFC4q1pF577TVcvHgRP/74I5ydnTFp0iT06tUL586dg62tLX7//XeMGjUK4eHh6N+/P/bs2YMZM2YYbeOXX37BsGHD8MUXX6Bjx464fPky3nzzTQAoNrYiWdQ7N15eXkhJSTFalpKSAmdnZ9jb25e4jlqthrOzs9GtogxoVRtdG3sgX6fHBxuiUcD2FBER/U1WVhZWrlyJRYsWoWvXrggICMCaNWuKvbvRq1cvvPXWW2jQoAEmTZoEd3d3HDx40PD48uXLcfPmTbz77rsYNWoUZs6ciVatWhXb38iRI7FhwwZkZ2fj8OHDSE9PR58+fYzGPAg133zzDTp27AitVovvv/8eSUlJhpN/P//8c/To0QMTJ05Eo0aN8O677yI0NNRoO7NmzcLkyZMxfPhw1KtXD926dcOcOXOwfPnycvrpPRmLeucmODgYu3fvNlq2f/9+BAcHC6rImCRJmPtCALp/ehink9Kx/NBlvP3PhqLLIiKqEuxtlTg3O/TxAyto30/q8uXLyM/PR1BQkGGZm5sb/P39jcYFBgYa/i5JEry8vJCammpYVr16daxcuRKhoaFo3749Jk+eXOL+tFotGjZsiE2bNuHgwYMYOnQobGyMX/5jY2NhY2NjVFONGjXg7++P2NhYw5i/fwAoODgYe/bsMdyPjo7GkSNH8PHHHxuW6XQ65ObmIicnp9K+JkNouMnKysKlS5cM9+Pj43Hq1Cm4ubnB19cXYWFhSEpKwnfffQcA+Pe//40lS5Zg4sSJGDlyJH7++Wds2LABu3btEjWFYjyd7TDruWYYu/4UPo+4iK5NPNHEu+LeLSIioiKSJJnUGjJ3f/9uJUmSin37+eHDh6FUKnHz5k1kZ2fDyank0yFGjhyJpUuX4ty5czh27FiF1ZyVlYVZs2bhhRdeKPaYnZ1dhe3374S2pf744w8888wzeOaZZwAA48aNwzPPPIPp06cDAG7evImEhATD+Lp162LXrl3Yv38/tFotFi9ejG+++abY22Ki9Wvhg+5NPVGgk9meIiIiI/Xr1zecw/LA3bt3ceHCBZO2c/ToUcyfPx87duyAo6Mj3n777VLHvvzyyzh9+jSaN2+Opk2bFnu8SZMmKCwsNKrp9u3biIuLM4xv0qSJ0eMA8Ntvvxndb9myJeLi4tCgQYNit8r81nahEbdz586PPMO8pKsPd+7cGSdPnqzAqspOkiR8/HwA/nf1Ds7dzMDSg5cwNqSR6LKIiMgMODo6YtSoUZgwYQJq1KgBDw8PTJ061aQX/8zMTAwdOhTvvvsuevbsidq1a6NNmzbo27cvBgwYUGx89erVcfPmzVK/abthw4bo168f3njjDSxfvhxOTk6YPHkyatWqhX79+gEA3n33XXTo0AGLFi1Cv379sHfvXqOWFABMnz4dffr0ga+vLwYMGACFQoHo6GicOXMGH330kQk/pbKxqBOKLUlNJzVm92sOAFjy8yWcSUoXXBEREZmLhQsXomPHjujbty9CQkLw7LPPlngycGnee+89VKtWDXPnzgUABAQEYO7cufjXv/6FpKSkEtdxdXVFtWrVSt3mt99+i1atWqFPnz4IDg6GLMvYvXu3IRC1a9cOK1aswOeffw6tVot9+/bhww8/NNpGaGgodu7ciX379qFNmzZo164dPv30U9SpU+eJ51YeJNnUD+dbuIyMDLi4uCA9Pb1CPzkFFF1zYcwPJ7D7dDIaeznhx7efhcqGeZKIqDzk5uYiPj4edevWrdTzOajiPOqYmvL6zVfaCiRJEub0a44a1VQ4n5yJL3+++PiViIiIqEwYbipYDUc1Pupf1J76T+RlRCfeE1sQERGRlWO4qQQ9A7zRV+sDnV7G+I3RyC3QiS6JiIjIajHcVJLZzzWDu6MaF1Oz8NkBtqeIiMpLFTt11KqV17FkuKkk1aupMPf5ovbU14cv40TCXcEVERFZNqWy6KrA+fn5giuh8vLgWD44tk/Lei7laAG6N/PC88/UwtaTSRi/MRq73+0IOxMu2U1ERH+xsbGBg4MD0tLSYGtrW6kXiaPyp9frkZaWBgcHh2JfD2EqhptKNrNvMxy5dAtX0rKxeF8cpvYufqVIIiJ6PEmS4O3tjfj4eFy7dk10OVQOFAoFfH19IUlSmbbD69wI8PP5FIxc/QckCdj4r2C09nMTUgcRkTXQ6/VsTVkJlUpV6jtwprx+850bAf7Z2BMDW9XGxuPXMX5jNH567x+wV7E9RUT0NBQKBS/iR0bYoBTkwz5N4e1ih6u3c7Bg73nR5RAREVkNhhtBXOxtMe/FQADAt0eu4rcrtwVXREREZB0YbgTq1KgmhrTVAAAmbopBdl6h4IqIiIgsH8ONYFN6NUEtV3sk3MnB/D1sTxEREZUVw41gTna2mP9ne+q7qGs4eumW4IqIiIgsG8ONGXi2oTtebecLAJiwKQaZuQWCKyIiIrJcDDdmIqxnE9Subo+ke/cxdzfbU0RERE+L4cZMVFPbYOEALQBg7bEEHL6QJrgiIiIiy8RwY0aC69fAa+39AACTNscgg+0pIiIikzHcmJmJPfzhV8MBN9Nz8dHOc6LLISIisjgMN2bGQWWDhQO1kCRgwx/XcfB8quiSiIiILArDjRlq4+eGUR3qAgAmb4lBeg7bU0RERE+K4cZMjQ/1Rz33akjJyMOsnWdFl0NERGQxGG7MlJ2tEosGaaGQgC0nkrD/XIrokoiIiCwCw40Za+lbHW/8ox4AYMrW07ibnS+4IiIiIvPHcGPm3g9phAYejkjLzMPMHWxPERERPQ7DjZmzs1Vi8UAtlAoJ20/dwJ4zN0WXREREZNYYbiyAVuOKf3cqak9N3XoGt7PyBFdERERkvhhuLMS7XRvC39MJt7PzMX0721NERESlYbixEGobJRYPKmpP7Tp9EztjboguiYiIyCwx3FiQ5rVcMKZLAwDAtG1nkJbJ9hQREdHfMdxYmLe7NEBTb2fczSnAh9tOQ5Zl0SURERGZFYYbC6OyUWDRQC1slRL2nk3Bj9FsTxERET2M4cYCNfVxxrv/bAgAmL79LFIzcgVXREREZD4YbizUvzvXR0AtF6TfL8CUrWxPERERPcBwY6FslUXtKZVSgQOxqdhyIkl0SURERGaB4caC+Xs5YWy3ovbUzB1nkZzO9hQRERHDjYV7s2M9aDWuyMwtxOQtMWxPERFRlcdwY+FslAosHhgIlY0CkXFp2PjHddElERERCcVwYwUaeDhhfPdGAIDZO88h6d59wRURERGJw3BjJUY9Ww8tfV2RlVeISZvYniIioqqL4cZKKBUSFg3UQm2jwK+XbuGHYwmiSyIiIhKC4caK1KvpiIk9GgMAPt4Vi8Q7OYIrIiIiqnwMN1ZmRHs/tPVzQ06+DhM3xUCvZ3uKiIiqFoYbK6NQSFg4MBD2tkpEXbmN//v9muiSiIiIKhXDjRWqU6MawnoVtafCd5/HtdvZgisiIiKqPAw3VurVoDoIrlcD9wt0mLCR7SkiIqo6GG6slEIhYcGAQFRTKXHs6h2sPnpVdElERESVguHGimncHDCldxMAwIK953ElLUtwRURERBWP4cbKvdzWF882cEdugR4TNsVAx/YUERFZOYYbKydJEuYPCISj2gbHr93Fql/jRZdERERUoRhuqoBarvaY1qeoPbVwXxwupbI9RURE1ovhpooY1FqDTo1qIr9Qjw82RqNQpxddEhERUYVguKkiJEnCvBcD4GRng+jEe/j6lyuiSyIiIqoQDDdViLeLPWb0bQYA+Gz/RcQlZwquiIiIqPwx3FQxL7asha6NPZCv02P8xmgUsD1FRERWRni4Wbp0Kfz8/GBnZ4egoCAcO3bskeM/++wz+Pv7w97eHhqNBu+//z5yc3MrqVrLJ0kSwl8IgIu9LU4npWNZ5GXRJREREZUroeFm/fr1GDduHGbMmIETJ05Aq9UiNDQUqampJY7/4YcfMHnyZMyYMQOxsbFYuXIl1q9fjylTplRy5ZbNw9kOs/sVtae++Pkizt3IEFwRERFR+REabj755BO88cYbGDFiBJo2bYply5bBwcEBq1atKnH80aNH0aFDB7z88svw8/ND9+7dMWTIkMe+20PFPaf1QWgzTxToZIzfGI38QraniIjIOggLN/n5+Th+/DhCQkL+KkahQEhICKKiokpcp3379jh+/LghzFy5cgW7d+9Gr169St1PXl4eMjIyjG5U1J76qH8AqjvY4tzNDCw9eEl0SUREROVCWLi5desWdDodPD09jZZ7enoiOTm5xHVefvllzJ49G88++yxsbW1Rv359dO7c+ZFtqfDwcLi4uBhuGo2mXOdhyWo6qTGnf3MAwNKDl3AmKV1wRURERGUn/IRiU0RGRmLu3Ln4z3/+gxMnTmDLli3YtWsX5syZU+o6YWFhSE9PN9wSExMrsWLz1yfQB70DvFGoL2pP5RXqRJdERERUJjaiduzu7g6lUomUlBSj5SkpKfDy8ipxnWnTpmHo0KF4/fXXAQABAQHIzs7Gm2++ialTp0KhKJ7V1Go11Gp1+U/Aiszu1wy/XbmN88mZ+DLiEsaH+osuiYiI6KkJe+dGpVKhVatWiIiIMCzT6/WIiIhAcHBwievk5OQUCzBKpRIAIMv8tuunVcNRjY/+bE99degyohPviS2IiIioDIS2pcaNG4cVK1ZgzZo1iI2NxejRo5GdnY0RI0YAAIYNG4awsDDD+L59++Krr77CunXrEB8fj/3792PatGno27evIeTQ0+kZ4I3ntD7Q6WV8sDEauQVsTxERkWUS1pYCgMGDByMtLQ3Tp09HcnIyWrRogT179hhOMk5ISDB6p+bDDz+EJEn48MMPkZSUhJo1a6Jv3774+OOPRU3Bqsx6rhmOXr6NS6lZ+PTABYT1bCK6JCIiIpNJchXr52RkZMDFxQXp6elwdnYWXY7Z2Xc2GW/+9zgUErDx3+3Rqk510SURERGZ9PptUZ+WoorXvZkXXnimFvQyMIHtKSIiskAMN1TMjL7N4OmsxpVb2Vi0N050OURERCZhuKFiXBxsMe+FQADAyiPx+N/VO4IrIiIienIMN1SiLo09MKh1bch/tqdy8gtFl0RERPREGG6oVB/2aQpvFztcvZ2DBXvYniIiIsvAcEOlcrazxfwXi9pTq49exW9XbguuiIiI6PEYbuiR/tGoJoa09QUATNgUjew8tqeIiMi8MdzQY03t3QS1XO2ReOc+5v10XnQ5REREj8RwQ4/lqLbBggFF7an//nYNRy7dElwRERFR6Rhu6Il0aOCOoe3qAAAmbopBZm6B4IqIiIhKxnBDT2xyz8bQuNkj6d59zN0dK7ocIiKiEjHc0BOrprbBwgFaAMDaY4k4dCFNcEVERETFMdyQSdrVq4HX2vsBACZtikH6fbaniIjIvDDckMkm9vCHXw0HJGfk4qOd50SXQ0REZIThhkzmoLLBooFaSBKw8fh1/Hw+RXRJREREBgw39FRa+7nh9WfrAgAmbz6N9By2p4iIyDww3NBT+6C7P+rVrIbUzDzM2nFWdDlEREQAGG6oDOxslVg0UAuFBGw5mYR9Z5NFl0RERMRwQ2XT0rc63vxHfQDAlK1ncDc7X3BFRERU1THcUJmNDWmIhh6OuJWVhxk/sj1FRERiMdxQmT1oTykVEn6MvoGfTt8UXRIREVVhDDdULrQaV4zuVNSe+nDbGdzOyhNcERERVVUMN1Ru3unaAI29nHA7Ox/Ttp+BLMuiSyIioiqI4YbKjdqmqD1lo5Cw+3QydsawPUVERJWP4YbKVfNaLhjTpQEAYNr2M0jNzBVcERERVTUMN1TuxnRpgKbezriXU4CpW9meIiKiysVwQ+VOZaPA4kFa2Col7D+Xgu2nboguiYiIqhCGG6oQTbyd8V7XhgCAGT+eRUoG21NERFQ5GG6owvy7U30E1HJB+v0CTNlymu0pIiKqFAw3VGFslEXtKZVSgYjzqdh8Ikl0SUREVAUw3FCFauTphPe7NQIAzNpxFjfT7wuuiIiIrB3DDVW4NzrWRQuNKzJzCzF5M9tTRERUsRhuqMLZKBVYNFALlY0Chy6kYcMfiaJLIiIiK8ZwQ5WigYcjJnT3BwDM2RmLpHtsTxERUcVguKFKM/LZumhVpzqy8goxaVMM21NERFQhGG6o0igVEhYOCISdrQK/XrqF739PEF0SERFZIYYbqlT1ajpiYmhjAMDc3bFIvJMjuCIiIrI2DDdU6V5r74e2dd2Qk6/DhE3R0OvZniIiovLDcEOVTqGQsGiAFg4qJX67cgf//e2a6JKIiMiKMNyQEL41HBDWs6g9Ne+n87h6K1twRUREZC0YbkiYV4LqoH39GrhfwPYUERGVH4YbEkahkDD/xUBUUynxv6t38e3Rq6JLIiIiK8BwQ0Jp3BwwtXdTAMCCPedxJS1LcEVERGTpGG5IuCFtNejY0B15hXqM3xgNHdtTRERUBgw3JJwkFbWnnNQ2OJFwDyt/vSK6JCIismAMN2QWfFztMa1PUXtq0b4LuJSaKbgiIiKyVAw3ZDYGtq6Nzv41kV+oxwcbY1Co04suiYiILBDDDZkNSZIw74VAONnZIDrxHpYfZnuKiIhMx3BDZsXLxQ4z+zYDAHx24ALOJ2cIroiIiCwNww2ZnRda1kJIEw8U6GSM3xiNAraniIjIBAw3ZHYkScLc5wPgYm+LM0kZ+CrysuiSiIjIgjDckFnycLbD7H5F7akvIi7i7I10wRUREZGlYLghs/Wc1gc9mnmhUC9j/MYY5BeyPUVERI/HcENmS5IkfPR8c7hVUyH2ZgaWHLwkuiQiIrIADDdk1twd1ZjTrzkAYOnBSziTxPYUERE9GsMNmb3egd7oHegNnV7GBxuikVeoE10SERGZMYYbsghz+jWHu6MKcSmZ+CLiouhyiIjIjAkPN0uXLoWfnx/s7OwQFBSEY8eOPXL8vXv3MGbMGHh7e0OtVqNRo0bYvXt3JVVLorhVU+Gj/gEAgK8iLyM68Z7YgoiIyGwJDTfr16/HuHHjMGPGDJw4cQJarRahoaFITU0tcXx+fj66deuGq1evYtOmTYiLi8OKFStQq1atSq6cROjR3Av9WvhALwMfbIxGbgHbU0REVJwky7IsaudBQUFo06YNlixZAgDQ6/XQaDR45513MHny5GLjly1bhoULF+L8+fOwtbV9on3k5eUhLy/PcD8jIwMajQbp6elwdnYun4lQpbmbnY/unx1GWmYe/tWpHsJ6NhFdEhERVYKMjAy4uLg80eu3sHdu8vPzcfz4cYSEhPxVjEKBkJAQREVFlbjOjz/+iODgYIwZMwaenp5o3rw55s6dC52u9P/Bh4eHw8XFxXDTaDTlPheqPNWrqTD3+aL21IrDV3D82l3BFRERkbkRFm5u3boFnU4HT09Po+Wenp5ITk4ucZ0rV65g06ZN0Ol02L17N6ZNm4bFixfjo48+KnU/YWFhSE9PN9wSExPLdR5U+bo19cQLLWtBLwPjN0bjfj7bU0RE9BfhJxSbQq/Xw8PDA19//TVatWqFwYMHY+rUqVi2bFmp66jVajg7OxvdyPLN6NMMns5qxN/KxqJ9caLLISIiMyIs3Li7u0OpVCIlJcVoeUpKCry8vEpcx9vbG40aNYJSqTQsa9KkCZKTk5Gfn1+h9ZJ5cXGwxbwXAwEAq47E41j8HcEVERGRuRAWblQqFVq1aoWIiAjDMr1ej4iICAQHB5e4TocOHXDp0iXo9X99x9CFCxfg7e0NlUpV4TWTeeni74HBrTWQZWDCpmjk5BeKLomIiMyA0LbUuHHjsGLFCqxZswaxsbEYPXo0srOzMWLECADAsGHDEBYWZhg/evRo3LlzB++99x4uXLiAXbt2Ye7cuRgzZoyoKZBgU/s0gY+LHa7dzsGCPWxPERERYCNy54MHD0ZaWhqmT5+O5ORktGjRAnv27DGcZJyQkACF4q/8pdFosHfvXrz//vsIDAxErVq18N5772HSpEmipkCCOdvZYv6AQAxdeQyrj15FaDMvBNevIbosIiISSOh1bkQw5XPyZDmmbD2NH35PQO3q9tg79h+ophaa24mIqJxZxHVuiMrTlF5NUMvVHtfv3kf4T7GiyyEiIoEYbsgqOKptsHBA0aen/u+3BPx68ZbgioiISBSGG7Ia7Ru4Y1hwHQDApM0xyMwtEFwRERGJwHBDVmVSj8bwdXNA0r37mLub7SkioqqI4YasSrWH2lNrjyUiMq7kb5gnIiLrxXBDVieoXg2M6OAHAJi8+TTS77M9RURUlTDckFWaGNoYdd2rITkjF3N2nhNdDhERVSKGG7JK9iolFg0MhCQBm45fR0RsyuNXIiIiq8BwQ1arVR03vNGxHgAgbMtp3Mvhl6sSEVUFDDdk1cZ1a4T6NashNTMPs3awPUVEVBUw3JBVs7NVYtFALRQSsPVkEvaeTRZdEhERVTCGG7J6z/hWx7861QcATN16Gney2Z4iIrJmDDdUJYwNaYhGno64lZWPGT+eFV0OERFVIIYbqhLUNkXtKaVCwo7oG9h9+qbokoiIqIIw3FCVEVjbFW91LmpPfbjtDG5l5QmuiIiIKoJJ4WbBggW4f/++4f6RI0eQl/fXC0RmZibeeuut8quOqJy988+GaOzlhDvZ+Zi27QxkWRZdEhERlTOTwk1YWBgyMzMN93v27ImkpCTD/ZycHCxfvrz8qiMqZyobBRYN1MJGIeGnM8nYGcP2FBGRtTEp3Pz9f7n8Xy9Zoua1XPD2PxsAAKZtP4PUzFzBFRERUXniOTdUJY3p0gBNvZ1xL6cAU7eyPUVEZE0YbqhKslUq8MlgLWyVEvafS8G2U0mPX4mIiCyCjakrfPPNN3B0dAQAFBYWYvXq1XB3dwcAo/NxiMxdYy9njA1phIV74zBj+1m0r+8OT2c70WUREVEZSbIJ78f7+flBkqTHjouPjy9TURUpIyMDLi4uSE9Ph7Ozs+hySLBCnR4vfHUUMdfT8c/GHlg5vPUTPceJiKhymfL6bdI7N1evXi1LXURmx0apwOKBWvT+4lf8fD4Vm45fx8DWGtFlERFRGfCcG6ryGno6YVz3RgCA2TvO4Wb6/cesQURE5sykcBMVFYWdO3caLfvuu+9Qt25deHh44M033zS6qB+RpXijYz084+uKzLxCTNp8mp+eIiKyYCaFm9mzZ+Ps2b++dPD06dMYNWoUQkJCMHnyZOzYsQPh4eHlXiRRRVMqJCwaqIXaRoHDF9Kw/n+JoksiIqKnZFK4OXXqFLp27Wq4v27dOgQFBWHFihUYN24cvvjiC2zYsKHciySqDPVrOmJCqD8A4KNdsbh+N0dwRURE9DRMCjd3796Fp6en4f6hQ4fQs2dPw/02bdogMZH/4yXLNaJDXbSuUx1ZeYWYtDmG7SkiIgtkUrjx9PQ0fMw7Pz8fJ06cQLt27QyPZ2ZmwtbWtnwrJKpESoWEhQO1sLNV4Mil2/j+9wTRJRERkYlMCje9evXC5MmT8csvvyAsLAwODg7o2LGj4fGYmBjUr1+/3Iskqkx13athUo/GAIC5u2ORcJvtKSIiS2JSuJkzZw5sbGzQqVMnrFixAl9//TVUKpXh8VWrVqF79+7lXiRRZRse7Ie2dd2Qk6/DhE3R0OvZniIishQmXaH4gfT0dDg6OkKpVBotv3PnDpycnMy6NcUrFNOTSridgx6fH0ZOvg4z+zbFax3qii6JiKjKqrArFI8cOfKJxq1atcqUzRKZJd8aDgjr1QTTtp3BvD3n0dnfA37u1USXRUREj2FSW2r16tU4ePAg7t27h7t375Z6I7IWr7T1RYcGNZBboMf4jdHQsT1FRGT2THrnZvTo0Vi7di3i4+MxYsQIvPrqq3Bzc6uo2oiEUygkzH8xEKGfHsYf1+7i2yPxeL1jPdFlERHRI5j0zs3SpUtx8+ZNTJw4ETt27IBGo8GgQYOwd+9eXg+ErFbt6g74sE9TAMDCvXG4nJYluCIiInoUk784U61WY8iQIdi/fz/OnTuHZs2a4a233oKfnx+ysviPPlmnl9po0LGhO/IK2Z4iIjJ3ZfpWcIVCAUmSIMsydDpdedVEZHYkqag95aS2wcmEe/jmlyuiSyIiolKYHG7y8vKwdu1adOvWDY0aNcLp06exZMkSJCQkwNHRsSJqJDILPq72mNa3qD21eP8FXEzJFFwRERGVxKRw89Zbb8Hb2xvz5s1Dnz59kJiYiI0bN6JXr15QKMr0JhCRRRjYqja6+NdE/p/tqUKdXnRJRET0NyZdxE+hUMDX1xfPPPMMJEkqddyWLVvKpbiKwIv4UVklp+ei+6eHkJFbiAmh/hjTpYHokoiIrF6FXcRv2LBhjww1RFWBl4sdZj7XDOM2ROOzAxfQtYkHGnsxKBMRmYun+voFS8Z3bqg8yLKMN747jgOxKWjm44xtYzrAVsnWLBFRRTHl9Zv/GhM9BUmSMPeF5nB1sMXZGxn4z8HLoksiIqI/MdwQPSUPJzvM7tccAPDlzxdx9ka64IqIiAhguCEqk76B3ujZ3AuFehkfbIhGfiE/PUVEJBrDDVEZSJKEOf2bw62aCueTM7Hk54uiSyIiqvIYbojKyN1RjTl/tqeWRl7G6etsTxERicRwQ1QOegd6o0+gN3R6GR9sPIW8Qn4dCRGRKAw3ROVkdr/mcHdU4UJKFj4/wPYUEZEoDDdE5cStmgof9Q8AACw7dBmnEu+JLYiIqIpiuCEqRz2ae6F/Cx/oZeCDDaeQW8D2FBFRZWO4ISpnM59rhppOalxOy8an+y+ILoeIqMphuCEqZ64OKoQ/X9Se+vqXKzh+7Y7gioiIqhaGG6IKENLUEy+2rA1ZBsZvjMH9fLaniIgqC8MNUQWZ3rcpvJztEH8rGwv3xokuh4ioyjCLcLN06VL4+fnBzs4OQUFBOHbs2BOtt27dOkiShP79+1dsgURPwcXeFvNeLGpPfXs0Hr9fuS24IiKiqkF4uFm/fj3GjRuHGTNm4MSJE9BqtQgNDUVqauoj17t69SrGjx+Pjh07VlKlRKbr7O+Bl9poIMvAhE0xyMkvFF0SEZHVEx5uPvnkE7zxxhsYMWIEmjZtimXLlsHBwQGrVq0qdR2dTodXXnkFs2bNQr169SqxWiLTTe3dBD4udki4k4P5P50XXQ4RkdUTGm7y8/Nx/PhxhISEGJYpFAqEhIQgKiqq1PVmz54NDw8PjBo16rH7yMvLQ0ZGhtGNqDI52dliwQAtAGBN1DUcvXxLcEVERNZNaLi5desWdDodPD09jZZ7enoiOTm5xHV+/fVXrFy5EitWrHiifYSHh8PFxcVw02g0Za6byFTPNnTHK0G+AICJm2KQlcf2FBFRRRHeljJFZmYmhg4dihUrVsDd3f2J1gkLC0N6errhlpiYWMFVEpUsrFcT1K5uj+t37yN8d6zocoiIrJaNyJ27u7tDqVQiJSXFaHlKSgq8vLyKjb98+TKuXr2Kvn37Gpbp9XoAgI2NDeLi4lC/fn2jddRqNdRqdQVUT2QaR7UNFgwIxMsrfsf3vyegR3MvdGxYU3RZRERWR+g7NyqVCq1atUJERIRhmV6vR0REBIKDg4uNb9y4MU6fPo1Tp04Zbs899xy6dOmCU6dOseVEZq99fXcMD64DAJi0KQaZuQWCKyIisj5C37kBgHHjxmH48OFo3bo12rZti88++wzZ2dkYMWIEAGDYsGGoVasWwsPDYWdnh+bNmxut7+rqCgDFlhOZq0k9G+NgXBoS7uTg412xmPdioOiSiIisivBwM3jwYKSlpWH69OlITk5GixYtsGfPHsNJxgkJCVAoLOrUIKJHclDZYNFALQZ/HYV1/0tEj+Ze6OzvIbosIiKrIcmyLIsuojJlZGTAxcUF6enpcHZ2Fl0OVWGzd5zDqiPx8HK2w973/wEXe1vRJRERmS1TXr/5lgiRIBNC/VHXvRqSM3Ixe8c50eUQEVkNhhsiQexVSiwaGAiFBGw+cR0HzqU8fiUiInoshhsigVrVccMbHYu+QiRs62ncy8kXXBERkeVjuCES7P1ujVC/ZjWkZeZh5o9nRZdDRGTxGG6IBLOzVWLxoBZQSMC2Uzew50zJXz1CRERPhuGGyAy00Lji352Krq794bbTuJPN9hQR0dNiuCEyE++FNEQjT0fcysrH9O1nRJdDRGSxGG6IzITaRonFA1tAqZCwM+YmdsXcFF0SEZFFYrghMiMBtV0wpnNRe2ra9jO4lZUnuCIiIsvDcENkZt7+Z0M09nLCnex8TNt2BlXsIuJERGXGcENkZlQ2CiwepIWNQsJPZ5Kxg+0pIiKTMNwQmaFmPi54558NAQDTt59Bamau4IqIiCwHww2RmXqrS30083HGvZwCTNnC9hQR0ZNiuCEyU7bKovaUrVLCgdgUbD2ZJLokIiKLwHBDZMYaezljbEgjAMDMH88iOZ3tKSKix2G4ITJz//pHPWhruyAjtxBhW2LYniIiegyGGyIzZ6NUYNFALVQ2ChyMS8PG49dFl0REZNYYbogsQENPJ3zQrag9NWfHOdy4d19wRURE5ovhhshCvN6xHp7xdUVmXiEmbWZ7ioioNAw3RBZCqZCwaKAWahsFfrl4C+v+lyi6JCIis8RwQ2RB6td0xIRQfwDARzvP4frdHMEVERGZH4YbIgszokNdtPGrjux8HSZuioFez/YUEdHDGG6ILIxSIWHhAC3sbBU4evk2vj+WILokIiKzwnBDZIH83Kthco/GAIDw3bFIuM32FBHRAww3RBZqWLAfguq6ISdfhwmbotmeIiL6E8MNkYVS/NmeclAp8Xv8HayJuiq6JCIis8BwQ2TBfGs4YEqvJgCA+XvOI/5WtuCKiIjEY7ghsnCvBPni2QbuyC3QY8LGaOjYniKiKo7hhsjCSZKEeS8GwFFtgz+u3cW3R+JFl0REJBTDDZEVqF3dAR/2LmpPLdwbh0upWYIrIiISh+GGyEoMbqPBPxrVRF6hHuPZniKiKozhhshKSJKE+S8GwMnOBqcS72HFL1dEl0REJATDDZEV8Xaxx/Q+TQEAn+y7gIspmYIrIiKqfAw3RFZmQKva+GdjD+Tr9PhgYzQKdXrRJRERVSqGGyIrI0kSwl8IgLOdDWKup2P5YbaniKhqYbghskKeznaY1a8ZAOCzAxdwPjlDcEVERJWH4YbISvVvUQvdmnqiQCfjgw3RKGB7ioiqCIYbIislSRI+fr45XB1scfZGBpYevCS6JCKiSsFwQ2TFPJzsMLtfcwDAkp8v4UxSuuCKiIgqHsMNkZXrG+iNXgFeKNTLGL8xGvmFbE8RkXVjuCGycpIkYU6/5qhRTYXzyZn48ueLoksiIqpQDDdEVUANRzU+6l/UnvpP5GXEXL8ntiAiogrEcENURfQM8EZfrQ90+qJPT+UV6kSXRERUIRhuiKqQ2c81g7ujGhdTs/DZAbaniMg6MdwQVSHVq6kw9/mi9tTyQ5dxMuGu4IqIiMofww1RFdO9mReef6YW9DIwfmM0cgvYniIi68JwQ1QFzejbFB5OalxOy8Yn+y+ILoeIqFwx3BBVQa4OKoS/EAAAWPHLFRy/dkdwRURE5YfhhqiK6trEEwNa1YYsA+M3xuB+PttTRGQdGG6IqrBpfZrCy9kO8beysXBvnOhyiIjKBcMNURXmYm+LeS8Wtae+PRqP36/cFlwREVHZMdwQVXGd/T0wpK0GsgxM2BSD7LxC0SUREZUJww0RYUqvJqjlao+EOzmYv+e86HKIiMqE4YaI4GRni/kvBgIAvou6hqOXbgmuiIjo6THcEBEA4NmG7ni1nS+AovZUFttTRGShGG6IyCCsZxPUrm6PpHv3MXd3rOhyiIieCsMNERlUU9tg4QAtAOCH3xNw+EKa4IqIiExnFuFm6dKl8PPzg52dHYKCgnDs2LFSx65YsQIdO3ZE9erVUb16dYSEhDxyPBGZJrh+DbzW3g8AMHlzDDJyC8QWRERkIuHhZv369Rg3bhxmzJiBEydOQKvVIjQ0FKmpqSWOj4yMxJAhQ3Dw4EFERUVBo9Gge/fuSEpKquTKiazXxB7+qFPDATfSc/HxTraniMiySLIsyyILCAoKQps2bbBkyRIAgF6vh0ajwTvvvIPJkyc/dn2dTofq1atjyZIlGDZs2GPHZ2RkwMXFBenp6XB2di5z/UTW6lj8HQz+OgqyDHw7og26+HuILomIqjBTXr+FvnOTn5+P48ePIyQkxLBMoVAgJCQEUVFRT7SNnJwcFBQUwM3NrcTH8/LykJGRYXQjosdrW9cNIzvUBVDUnkrPYXuKiCyD0HBz69Yt6HQ6eHp6Gi339PREcnLyE21j0qRJ8PHxMQpIDwsPD4eLi4vhptFoylw3UVUxvrs/6rlXQ0pGHmbvPCe6HCKiJyL8nJuymDdvHtatW4etW7fCzs6uxDFhYWFIT0833BITEyu5SiLLZa9SYuFALRQSsPnEdew/lyK6JCKixxIabtzd3aFUKpGSYvwPZkpKCry8vB657qJFizBv3jzs27cPgYGBpY5Tq9VwdnY2uhHRk2tVpzre+Ec9AMCUradxNztfcEVERI8mNNyoVCq0atUKERERhmV6vR4REREIDg4udb0FCxZgzpw52LNnD1q3bl0ZpRJVae+HNEIDD0ekZeZh5o6zosshInok4W2pcePGYcWKFVizZg1iY2MxevRoZGdnY8SIEQCAYcOGISwszDB+/vz5mDZtGlatWgU/Pz8kJycjOTkZWVlZoqZAZPXsbJVYPFALpULC9lM3sOfMTdElERGVSni4GTx4MBYtWoTp06ejRYsWOHXqFPbs2WM4yTghIQE3b/71D+lXX32F/Px8DBgwAN7e3obbokWLRE2BqErQalzx705F7ampW8/gdlae4IqIiEom/Do3lY3XuSF6enmFOjz35RHEpWSid6A3lr7cUnRJRFRFWMx1bojIsqhtlFg8qKg9tSvmJnbG3BBdEhFRMQw3RGSS5rVcMKZLAwDAtG1nkJbJ9hQRmReGGyIy2dtdGqCJtzPu5hTgw22nUcW620Rk5hhuiMhkKhsFFg/UwkYhYe/ZFPwYzfYUEZkPhhsieipNfZzxbteGAIDp288iNSNXcEVEREUYbojoqY3uXB/Nazkj/X4Bpmxle4qIzAPDDRE9NVulAosHtoCtUsKB2FRsOZEkuiQiIoYbIiobfy8njA1pBACYueMsktPZniIisRhuiKjM/vWPetBqXJGZW4jJW2LYniIioRhuiKjMbJQKLB4YCJWNApFxadj4x3XRJRFRFcZwQ0TlooGHE8Z3L2pPzdl5Dkn37guuiIiqKoYbIio3o56th5a+rsjMK8TkzWxPEZEYDDdEVG6UCgmLBmqhtlHgl4u3sPZYouiSiKgKYrghonJVr6YjJvZoDAD4eNc5JN7JEVwREVU1DDdEVO5GtPdDWz83ZOfrMGlzDPR6tqeIqPIw3BBRuVMoJCwYEAh7WyWOXr6N73+/JrokIqpCGG6IqEL4uVfD5J5F7am5u88j4TbbU0RUORhuiKjCDG1XB+3queF+gQ7jN0WzPUVElYLhhogqjEIhYeEALRxUShyLv4M1UVdFl0REVQDDDRFVKI2bA6b0agIAmL/nPK6kZQmuiIisHcMNEVW4V4J88WwDd+QW6DFhUwx0bE8RUQViuCGiCidJEuYPCISj2gbHr93Fql/jRZdERFaM4YaIKkUtV3tM61PUnlq4Lw6XUtmeIqKKwXBDRJVmUGsNOjWqifxCPT7YGI1CnV50SURkhRhuiKjSSJKEeS8GwMnOBtGJ97DiF7aniKj8MdwQUaXydrHHjL7NAACf7r+ACymZgisiImvDcENEle7FlrXQtbEH8nV6fLAhGgVsTxFROWK4IaJKJ0kS5r4QABd7W5xOSsfyQ5dFl0REVoThhoiE8HS2w6znitpTn0dcROzNDMEVEZG1YLghImH6tfBB96aeKNDJbE8RUblhuCEiYSRJwsfPB6C6gy3O3czA0oOXRJdERFaA4YaIhKrppMbsfs0BAEt+voQzSemCKyIiS8dwQ0TC9dX6oHeANwr1MsZvjEZeoU50SURkwRhuiMgszO7XDDWqqXA+ORNfRrA9RURPj+GGiMxCDUc1Pupf1J766tBlRCfeE1sQEVkshhsiMhs9A7zxnNYHuj/bU7kFbE8RkekYbojIrMx6rhncHdW4mJqFzw5cFF0OEVkghhsiMivVq6kw9/mi9tTXhy/jRMJdwRURkaVhuCEis9O9mRdeeKYW9DLYniIikzHcEJFZmtG3GTyc1LiSlo3F++JEl0NEFoThhojMkouDLea9GAAA+ObXePxx9Y7giojIUjDcEJHZ+mdjTwxsVRvyn+2p+/lsTxHR4zHcEJFZ+7BPU3i72OHq7Rws2HtedDlEZAEYbojIrLnY22Lei4EAgG+PXMVvV24LroiIzB3DDRGZvU6NamJIWw0AYMKmaGTnFQquiIjMGcMNEVmEqb2boparPRLv3Me8n9ieIqLSMdwQkUVwVNtgwYCi9tR/f7uGI5duCa6IiMwVww0RWYwODdwxtF0dAMDETTHIzC0QXBERmSOGGyKyKJN7NobGzR5J9+5j7m62p4ioOIYbIrIo1dQ2WDhACwBYeywBhy+kCa6IiMwNww0RWZx29WrgtfZ+AIBJm2OQwfYUET2E4YaILNLEHv7wq+GAm+m5+GjnOdHlEJEZYbghIovkoLLBwoFaSBKw4Y/rOHg+VXRJRGQmGG6IyGK18XPDqA51AQCTt8QgPYftKSJiuCEiCzc+1B/13KshJSMPs3aeFV0OEZkBhhsismh2tkosGqSFQgK2nEjC/nMpoksiIsFsRBdARFRWLX2r441/1MPyQ1cwYVM0tLVdAQCSBEh/jpEk6aG/Aw8eMR4DSA8v//MBCUWDSttWScvx9209Zh+A9ND+jPf9932gtG09vPzPcQ9N1bCt0vbx13ip1DEl7QOP2FZJyx+eh/GYh5Y/tC/8+XN9kn2gpG09POZvP7OH6yn1GBmN/fu+H72PRz9vSpprSfsu+RiX/BwufR8oZfnjfq4P5mzKPtS2Cng42UEUswg3S5cuxcKFC5GcnAytVosvv/wSbdu2LXX8xo0bMW3aNFy9ehUNGzbE/Pnz0atXr0qsmIjMzfshjfBzbCoupmbhEK99QyRUS19XbHmrg7D9Cw8369evx7hx47Bs2TIEBQXhs88+Q2hoKOLi4uDh4VFs/NGjRzFkyBCEh4ejT58++OGHH9C/f3+cOHECzZs3FzADIjIHdrZKfP9GEI5cugWdvmiZLMuQHwyQgQf3ZBmG5XIpy/HnurJcfFvG6xvWKHVbf1+Oh9Y1bP/hfcF4+YM7JY0paR/4W32P28fD83hUHX+Nf2h5CXN+eH0YrV/SPoyXo6R9P+LnihKXFz9Wfz9OT7qP0p83JT8fUGz5Q+uWMOdHz+/vz9/S9l18Hw/G/bW/EuaMh8f/ffnf9lHiMSp9H2obJUSS5Id/wgIEBQWhTZs2WLJkCQBAr9dDo9HgnXfeweTJk4uNHzx4MLKzs7Fz507Dsnbt2qFFixZYtmzZY/eXkZEBFxcXpKenw9nZufwmQkRERBXGlNdvoScU5+fn4/jx4wgJCTEsUygUCAkJQVRUVInrREVFGY0HgNDQ0FLH5+XlISMjw+hGRERE1ktouLl16xZ0Oh08PT2Nlnt6eiI5ObnEdZKTk00aHx4eDhcXF8NNo9GUT/FERERklqz+o+BhYWFIT0833BITE0WXRERERBVI6AnF7u7uUCqVSEkxvi5FSkoKvLy8SlzHy8vLpPFqtRpqtbp8CiYiIiKzJ/SdG5VKhVatWiEiIsKwTK/XIyIiAsHBwSWuExwcbDQeAPbv31/qeCIiIqpahH8UfNy4cRg+fDhat26Ntm3b4rPPPkN2djZGjBgBABg2bBhq1aqF8PBwAMB7772HTp06YfHixejduzfWrVuHP/74A19//bXIaRAREZGZEB5uBg8ejLS0NEyfPh3Jyclo0aIF9uzZYzhpOCEhAQrFX28wtW/fHj/88AM+/PBDTJkyBQ0bNsS2bdt4jRsiIiICYAbXualsvM4NERGR5bGY69wQERERlTeGGyIiIrIqDDdERERkVRhuiIiIyKow3BAREZFVYbghIiIiqyL8OjeV7cEn3/nt4ERERJbjwev2k1zBpsqFm8zMTADgt4MTERFZoMzMTLi4uDxyTJW7iJ9er8eNGzfg5OQESZLKddsZGRnQaDRITEy0ygsEWvv8AOufI+dn+ax9jpyf5auoOcqyjMzMTPj4+Bh9c0FJqtw7NwqFArVr167QfTg7O1vtkxaw/vkB1j9Hzs/yWfscOT/LVxFzfNw7Ng/whGIiIiKyKgw3REREZFUYbsqRWq3GjBkzoFarRZdSIax9foD1z5Hzs3zWPkfOz/KZwxyr3AnFREREZN34zg0RERFZFYYbIiIisioMN0RERGRVGG6IiIjIqjDcPMLSpUvh5+cHOzs7BAUF4dixY48cv3HjRjRu3Bh2dnYICAjA7t27jR6XZRnTp0+Ht7c37O3tERISgosXL1bkFB7LlDmuWLECHTt2RPXq1VG9enWEhIQUG//aa69BkiSjW48ePSp6GqUyZX6rV68uVrudnZ3RGHM7hqbMr3PnzsXmJ0kSevfubRhjTsfv8OHD6Nu3L3x8fCBJErZt2/bYdSIjI9GyZUuo1Wo0aNAAq1evLjbG1N/rimTqHLds2YJu3bqhZs2acHZ2RnBwMPbu3Ws0ZubMmcWOYePGjStwFqUzdX6RkZElPkeTk5ONxlnyMSzpd0ySJDRr1swwxlyOYXh4ONq0aQMnJyd4eHigf//+iIuLe+x65vBayHBTivXr12PcuHGYMWMGTpw4Aa1Wi9DQUKSmppY4/ujRoxgyZAhGjRqFkydPon///ujfvz/OnDljGLNgwQJ88cUXWLZsGX7//XdUq1YNoaGhyM3NraxpGTF1jpGRkRgyZAgOHjyIqKgoaDQadO/eHUlJSUbjevTogZs3bxpua9eurYzpFGPq/ICiK2o+XPu1a9eMHjenY2jq/LZs2WI0tzNnzkCpVGLgwIFG48zl+GVnZ0Or1WLp0qVPND4+Ph69e/dGly5dcOrUKYwdOxavv/660Yv/0zwnKpKpczx8+DC6deuG3bt34/jx4+jSpQv69u2LkydPGo1r1qyZ0TH89ddfK6L8xzJ1fg/ExcUZ1e/h4WF4zNKP4eeff240t8TERLi5uRX7PTSHY3jo0CGMGTMGv/32G/bv34+CggJ0794d2dnZpa5jNq+FMpWobdu28pgxYwz3dTqd7OPjI4eHh5c4ftCgQXLv3r2NlgUFBcn/+te/ZFmWZb1eL3t5eckLFy40PH7v3j1ZrVbLa9eurYAZPJ6pc/y7wsJC2cnJSV6zZo1h2fDhw+V+/fqVd6lPxdT5ffvtt7KLi0up2zO3Y1jW4/fpp5/KTk5OclZWlmGZOR2/hwGQt27d+sgxEydOlJs1a2a0bPDgwXJoaKjhfll/ZhXpSeZYkqZNm8qzZs0y3J8xY4as1WrLr7By8iTzO3jwoAxAvnv3bqljrO0Ybt26VZYkSb569aphmbkew9TUVBmAfOjQoVLHmMtrId+5KUF+fj6OHz+OkJAQwzKFQoGQkBBERUWVuE5UVJTReAAIDQ01jI+Pj0dycrLRGBcXFwQFBZW6zYr0NHP8u5ycHBQUFMDNzc1oeWRkJDw8PODv74/Ro0fj9u3b5Vr7k3ja+WVlZaFOnTrQaDTo168fzp49a3jMnI5heRy/lStX4qWXXkK1atWMlpvD8Xsaj/sdLI+fmbnR6/XIzMws9jt48eJF+Pj4oF69enjllVeQkJAgqMKn06JFC3h7e6Nbt244cuSIYbk1HsOVK1ciJCQEderUMVpujscwPT0dAIo93x5mLq+FDDcluHXrFnQ6HTw9PY2We3p6Fuv9PpCcnPzI8Q/+NGWbFelp5vh3kyZNgo+Pj9GTtEePHvjuu+8QERGB+fPn49ChQ+jZsyd0Ol251v84TzM/f39/rFq1Ctu3b8f//d//Qa/Xo3379rh+/ToA8zqGZT1+x44dw5kzZ/D6668bLTeX4/c0SvsdzMjIwP3798vlOW9uFi1ahKysLAwaNMiwLCgoCKtXr8aePXvw1VdfIT4+Hh07dkRmZqbASp+Mt7c3li1bhs2bN2Pz5s3QaDTo3LkzTpw4AaB8/t0yJzdu3MBPP/1U7PfQHI+hXq/H2LFj0aFDBzRv3rzUcebyWljlvhWcyse8efOwbt06REZGGp10+9JLLxn+HhAQgMDAQNSvXx+RkZHo2rWriFKfWHBwMIKDgw3327dvjyZNmmD58uWYM2eOwMrK38qVKxEQEIC2bdsaLbfk41fV/PDDD5g1axa2b99udE5Kz549DX8PDAxEUFAQ6tSpgw0bNmDUqFEiSn1i/v7+8Pf3N9xv3749Ll++jE8//RT//e9/BVZWMdasWQNXV1f079/faLk5HsMxY8bgzJkzws7fMhXfuSmBu7s7lEolUlJSjJanpKTAy8urxHW8vLweOf7Bn6ZssyI9zRwfWLRoEebNm4d9+/YhMDDwkWPr1asHd3d3XLp0qcw1m6Is83vA1tYWzzzzjKF2czqGZZlfdnY21q1b90T/SIo6fk+jtN9BZ2dn2Nvbl8tzwlysW7cOr7/+OjZs2FCsBfB3rq6uaNSokUUcw5K0bdvWULs1HUNZlrFq1SoMHToUKpXqkWNFH8O3334bO3fuxMGDB1G7du1HjjWX10KGmxKoVCq0atUKERERhmV6vR4RERFG/7N/WHBwsNF4ANi/f79hfN26deHl5WU0JiMjA7///nup26xITzNHoOgs9zlz5mDPnj1o3br1Y/dz/fp13L59G97e3uVS95N62vk9TKfT4fTp04bazekYlmV+GzduRF5eHl599dXH7kfU8Xsaj/sdLI/nhDlYu3YtRowYgbVr1xp9jL80WVlZuHz5skUcw5KcOnXKULu1HEOg6JNIly5deqL/ZIg6hrIs4+2338bWrVvx888/o27duo9dx2xeC8vt1GQrs27dOlmtVsurV6+Wz507J7/55puyq6urnJycLMuyLA8dOlSePHmyYfyRI0dkGxsbedGiRXJsbKw8Y8YM2dbWVj59+rRhzLx582RXV1d5+/btckxMjNyvXz+5bt268v379yt9frJs+hznzZsnq1QqedOmTfLNmzcNt8zMTFmWZTkzM1MeP368HBUVJcfHx8sHDhyQW7ZsKTds2FDOzc01+/nNmjVL3rt3r3z58mX5+PHj8ksvvSTb2dnJZ8+eNYwxp2No6vweePbZZ+XBgwcXW25uxy8zM1M+efKkfPLkSRmA/Mknn8gnT56Ur127JsuyLE+ePFkeOnSoYfyVK1dkBwcHecKECXJsbKy8dOlSWalUynv27DGMedzPrLKZOsfvv/9etrGxkZcuXWr0O3jv3j3DmA8++ECOjIyU4+Pj5SNHjsghISGyu7u7nJqaavbz+/TTT+Vt27bJFy9elE+fPi2/9957skKhkA8cOGAYY+nH8IFXX31VDgoKKnGb5nIMR48eLbu4uMiRkZFGz7ecnBzDGHN9LWS4eYQvv/xS9vX1lVUqldy2bVv5t99+MzzWqVMnefjw4UbjN2zYIDdq1EhWqVRys2bN5F27dhk9rtfr5WnTpsmenp6yWq2Wu3btKsfFxVXGVEplyhzr1KkjAyh2mzFjhizLspyTkyN3795drlmzpmxrayvXqVNHfuONN4T9oyPLps1v7NixhrGenp5yr1695BMnThhtz9yOoanP0fPnz8sA5H379hXblrkdvwcfC/777cGchg8fLnfq1KnYOi1atJBVKpVcr149+dtvvy223Uf9zCqbqXPs1KnTI8fLctHH3729vWWVSiXXqlVLHjx4sHzp0qXKndifTJ3f/Pnz5fr168t2dnaym5ub3LlzZ/nnn38utl1LPoayXPTRZ3t7e/nrr78ucZvmcgxLmhcAo98rc30tlP6cABEREZFV4Dk3REREZFUYboiIiMiqMNwQERGRVWG4ISIiIqvCcENERERWheGGiIiIrArDDREREVkVhhsiIiKyKgw3RFTlRUZGQpIk3Lt3T3QpRFQOGG6IiIjIqjDcEBERkVVhuCEi4fR6PcLDw1G3bl3Y29tDq9Vi06ZNAP5qGe3atQuBgYGws7NDu3btcObMGaNtbN68Gc2aNYNarYafnx8WL15s9HheXh4mTZoEjUYDtVqNBg0aYOXKlUZjjh8/jtatW8PBwQHt27dHXFxcxU6ciCoEww0RCRceHo7vvvsOy5Ytw9mzZ/H+++/j1VdfxaFDhwxjJkyYgMWLF+N///sfatasib59+6KgoABAUSgZNGgQXnrpJZw+fRozZ87EtGnTsHr1asP6w4YNw9q1a/HFF18gNjYWy5cvh6Ojo1EdU6dOxeLFi/HHH3/AxsYGI0eOrJT5E1H54reCE5FQeXl5cHNzw4EDBxAcHGxY/vrrryMnJwdvvvkmunTpgnXr1mHw4MEAgDt37qB27dpYvXo1Bg0ahFdeeQVpaWnYt2+fYf2JEydi165dOHv2LC5cuAB/f3/s378fISEhxWqIjIxEly5dcODAAXTt2hUAsHv3bvTu3Rv379+HnZ1dBf8UiKg88Z0bIhLq0qVLyMnJQbdu3eDo6Gi4fffdd7h8+bJh3MPBx83NDf7+/oiNjQUAxMbGokOHDkbb7dChAy5evAidTodTp05BqVSiU6dOj6wlMDDQ8Hdvb28AQGpqapnnSESVy0Z0AURUtWVlZQEAdu3ahVq1ahk9plarjQLO07K3t3+icba2toa/S5IEoOh8ICKyLHznhoiEatq0KdRqNRISEtCgQQOjm0ajMYz77bffDH+/e/cuLly4gCZNmgAAmjRpgiNHjhht98iRI2jUqBGUSiUCAgKg1+uNzuEhIuvFd26ISCgnJyeMHz8e77//PvR6PZ599lmkp6fjyJEjcHZ2Rp06dQAAs2fPRo0aNeDp6YmpU6fC3d0d/fv3BwB88MEHaNOmDebMmYPBgwcjKioKS5YswX/+8x8AgJ+fH4YPH46RI0fiiy++gFarxbVr15CamopBgwaJmjoRVRCGGyISbs6cOahZsybCw8Nx5coVuLq6omXLlpgyZYqhLTRv3jy89957uHjxIlq0aIEdO3ZApVIBAFq2bIkNGzZg+vTpmDNnDry9vTF79my89tprhn189dVXmDJlCt566y3cvn0bvr6+mDJliojpElEF46eliMisPfgk0927d+Hq6iq6HCKyADznhoiIiKwKww0RERFZFbaliIiIyKrwnRsiIiKyKgw3REREZFUYboiIiMiqMNwQERGRVWG4ISIiIqvCcENERERWheGGiIiIrArDDREREVmV/wdYJPczDzOedgAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure()\n",
    "plt.plot(dnxModel, label='dnxModel')\n",
    "plt.xlabel('epoch')\n",
    "plt.ylabel('MSE')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "ab658319-f0a9-4a48-9104-e758aef71888",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1.0, 0.00948867974195667, 0.007721767276280092]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.dnxlayer.errors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ca1d028b-a028-4769-a1bc-3cb78dd82659",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "37dd0aeb-ebaa-4320-88a9-f01b7c874786",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.10.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}