Final_Code_For_Model_Training_Weight_Initialization.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "3c0b54b4-64f2-4313-b4bc-10ca42558d6a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "%pylab is deprecated, use %matplotlib inline and import the required libraries.\n",
      "Populating the interactive namespace from numpy and matplotlib\n"
     ]
    }
   ],
   "source": [
    "%pylab inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "ef92803e-0e97-4c95-bf37-2408638a1701",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import snntorch as snn\n",
    "from snntorch import spikeplot as splt\n",
    "from snntorch import spikegen\n",
    "\n",
    "import numpy as np\n",
    "\n",
    "import scipy.io\n",
    "import scipy.signal \n",
    "\n",
    "from matplotlib import mlab\n",
    "import matplotlib.pyplot as plt\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "4e436c7c-eee1-4ad8-b199-5791afda45b7",
   "metadata": {},
   "outputs": [],
   "source": [
    "def psd(trials):\n",
    "    \n",
    "    ntrials = trials.shape[2]\n",
    "    trials_PSD = np.zeros((nchannels, 101, ntrials))\n",
    "\n",
    "    # Iterate over trials and channels\n",
    "    for trial in range(ntrials):\n",
    "        for ch in range(nchannels):\n",
    "            # Calculate the PSD\n",
    "            (PSD, freqs) = mlab.psd(trials[ch,:,trial], NFFT=int(nsamples), Fs=sample_rate)\n",
    "            trials_PSD[ch, :, trial] = PSD.ravel()\n",
    "                \n",
    "    return trials_PSD, freqs\n",
    "\n",
    "\n",
    "def plot_psd(trials_PSD, freqs, chan_ind, chan_lab=None, maxy=None):\n",
    "    plt.figure(figsize=(12,5))\n",
    "    \n",
    "    nchans = len(chan_ind)\n",
    "    \n",
    "    # Maximum of 3 plots per row\n",
    "    nrows = int(np.ceil(nchans / 3))\n",
    "    ncols = min(3, nchans)\n",
    "    \n",
    "    # Enumerate over the channels\n",
    "    for i,ch in enumerate(chan_ind):\n",
    "        # Figure out which subplot to draw to\n",
    "        plt.subplot(nrows,ncols,i+1)\n",
    "    \n",
    "        # Plot the PSD for each class\n",
    "        for cl in trials.keys():\n",
    "            plt.plot(freqs, np.mean(trials_PSD[cl][ch,:,:], axis=1), label=cl)\n",
    "    \n",
    "        # All plot decoration below...\n",
    "        \n",
    "        plt.xlim(1,30)\n",
    "        \n",
    "        if maxy != None:\n",
    "            plt.ylim(0,maxy)\n",
    "    \n",
    "        plt.grid()\n",
    "    \n",
    "        plt.xlabel('Frequency (Hz)')\n",
    "        \n",
    "        if chan_lab == None:\n",
    "            plt.title('Channel %d' % (ch+1))\n",
    "        else:\n",
    "            plt.title(chan_lab[i])\n",
    "\n",
    "        plt.legend()\n",
    "        \n",
    "    plt.tight_layout()\n",
    "\n",
    "def bandpass(trials, lo, hi, sample_rate):\n",
    " \n",
    "    a, b = scipy.signal.iirfilter(6, [lo/(sample_rate/2.0), hi/(sample_rate/2.0)])\n",
    "\n",
    "    ntrials = trials.shape[2]\n",
    "    trials_filt = np.zeros((nchannels, nsamples, ntrials))\n",
    "    for i in range(ntrials):\n",
    "        trials_filt[:,:,i] = scipy.signal.filtfilt(a, b, trials[:,:,i], axis=1)\n",
    "    \n",
    "    return trials_filt\n",
    "\n",
    "def logvar(trials):\n",
    "    return np.var(trials, axis=1)\n",
    "\n",
    "def plot_logvar(trials, trial_no = None):\n",
    "    plt.figure(figsize=(12,5))\n",
    "\n",
    "    print(trials[cl1].shape)\n",
    "    \n",
    "    x0 = np.arange(nchannels)\n",
    "    x1 = np.arange(nchannels) + 0.4\n",
    "\n",
    "    if trial_no == None:\n",
    "        y0 = np.mean(trials[cl1], axis=1)\n",
    "        y1 = np.mean(trials[cl2], axis=1)\n",
    "    else:\n",
    "        y0 = trials[cl1][:,trial_no]\n",
    "        y1 = trials[cl2][:,trial_no]\n",
    "\n",
    "    plt.bar(x0, y0, width=0.5, color='b')\n",
    "    plt.bar(x1, y1, width=0.4, color='r')\n",
    "\n",
    "    plt.xlim(-0.5, nchannels+0.5)\n",
    "\n",
    "    plt.gca().yaxis.grid(True)\n",
    "    plt.title('log-var of each channel/component')\n",
    "    plt.xlabel('channels/components')\n",
    "    plt.ylabel('log-var')\n",
    "    plt.legend(cl_lab)\n",
    "\n",
    "from numpy import linalg\n",
    "\n",
    "def cov(trials):\n",
    "    ntrials = trials.shape[2]\n",
    "    covs = [ trials[:,:,i].dot(trials[:,:,i].T) / nsamples for i in range(ntrials) ]\n",
    "    return np.mean(covs, axis=0)\n",
    "\n",
    "def whitening(sigma):\n",
    "    U, l, _ = linalg.svd(sigma)\n",
    "    return U.dot( np.diag(l ** -0.5) )\n",
    "\n",
    "def csp(trials_r, trials_f):\n",
    "    cov_r = cov(trials_r)\n",
    "    cov_f = cov(trials_f)\n",
    "    P = whitening(cov_r + cov_f)\n",
    "    B, _, _ = linalg.svd( P.T.dot(cov_f).dot(P) )\n",
    "    W = P.dot(B)\n",
    "    return W\n",
    "\n",
    "def apply_mix(W, trials):\n",
    "    ntrials = trials.shape[2]\n",
    "    trials_csp = np.zeros((nchannels, nsamples, ntrials))\n",
    "    for i in range(ntrials):\n",
    "        trials_csp[:,:,i] = W.T.dot(trials[:,:,i])\n",
    "    return trials_csp\n",
    "\n",
    "import torch\n",
    "import snntorch.spikeplot as splt\n",
    "from snntorch import spikegen\n",
    "\n",
    "def spike_latency(trials, trial_n, cls, num_steps):\n",
    "    input_data = trials[cl1] if cls == 1 else trials[cl2]\n",
    "    \n",
    "    tensor = torch.tensor(input_data[:,trial_n])\n",
    "    device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "    tensor = tensor.to(device)\n",
    "    \n",
    "    spike_data = spikegen.latency(tensor, num_steps=num_steps, threshold=(np.mean(input_data[:,trial_n])+1)*1/2, bypass=True)\n",
    "\n",
    "    return spike_data\n",
    "\n",
    "def spike_rate(trials, trial_n, cls, num_steps):\n",
    "    input_data = trials[cl1] if cls == 1 else trials[cl2]\n",
    "    \n",
    "    tensor = torch.tensor(input_data[:,trial_n])\n",
    "    device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "    tensor = tensor.to(device)\n",
    "    \n",
    "    spike_data = spikegen.rate(tensor, num_steps=num_steps)\n",
    "\n",
    "    return spike_data\n",
    "\n",
    "def plot_raster(spike_data):\n",
    "    \n",
    "    fig = plt.figure(facecolor=\"w\", figsize=(10, 5))\n",
    "    ax = fig.add_subplot(111)\n",
    "    \n",
    "    spike_data_sample = spike_data\n",
    "    splt.raster(spike_data_sample, ax, s=1.5, c=\"black\")\n",
    "    plt.ylim(0, 60)\n",
    "    \n",
    "    plt.title(\"Input Layer\")\n",
    "    plt.xlabel(\"Time step\")\n",
    "    plt.ylabel(\"Neuron Number\")\n",
    "    plt.show()\n",
    "\n",
    "def plot_snn_spikes(spk_in, spk1_rec, spk2_rec, num_steps, title):\n",
    "  # Generate Plots\n",
    "  fig, ax = plt.subplots(3, figsize=(8,7), sharex=True, \n",
    "                        gridspec_kw = {'height_ratios': [1, 1, 0.4]})\n",
    "\n",
    "  # Plot input spikes\n",
    "  splt.raster(spk_in[:,0], ax[0], s=0.1, c=\"black\")\n",
    "  ax[0].set_ylabel(\"Input Spikes\")\n",
    "  ax[0].set_title(title)\n",
    "\n",
    "  # Plot hidden layer spikes\n",
    "  splt.raster(spk1_rec.reshape(num_steps, -1), ax[1], s = 0.1, c=\"black\")\n",
    "  ax[1].set_ylabel(\"Hidden Layer\")\n",
    "\n",
    "  # Plot output spikes\n",
    "  splt.raster(spk2_rec.reshape(num_steps, -1), ax[2], c=\"black\", marker=\"|\")\n",
    "  ax[2].set_ylabel(\"Output Spikes\")\n",
    "  ax[2].set_ylim([0, 10])\n",
    "\n",
    "  plt.show()\n",
    "\n",
    "def plot_snn_spikes(spk_in, spk1_rec, spk2_rec, num_steps, title):\n",
    "  # Generate Plots\n",
    "  fig, ax = plt.subplots(3, figsize=(8,7), sharex=True, \n",
    "                        gridspec_kw = {'height_ratios': [1, 1, 0.4]})\n",
    "\n",
    "  # Plot input spikes\n",
    "  splt.raster(spk_in[:,0], ax[0], s=0.1, c=\"black\")\n",
    "  ax[0].set_ylabel(\"Input Spikes\")\n",
    "  ax[0].set_title(title)\n",
    "\n",
    "  # Plot hidden layer spikes\n",
    "  splt.raster(spk1_rec.reshape(num_steps, -1), ax[1], s = 0.1, c=\"black\")\n",
    "  ax[1].set_ylabel(\"Hidden Layer\")\n",
    "\n",
    "  # Plot output spikes\n",
    "  splt.raster(spk2_rec.reshape(num_steps, -1), ax[2], c=\"black\", marker=\"|\")\n",
    "  ax[2].set_ylabel(\"Output Spikes\")\n",
    "  ax[2].set_yticks([0, 1], ['left', 'right'])\n",
    "  ax[2].set_ylim([-0.5, 1.5])\n",
    "\n",
    "\n",
    "  plt.show()\n",
    "\n",
    "def plot_snn_spikes_modified(spk_in, spk1_rec, spk2_rec, num_steps, title):\n",
    "  # Generate Plots\n",
    "  fig, ax = plt.subplots(3, figsize=(8,7), sharex=False, \n",
    "                        gridspec_kw = {'height_ratios': [1, 1, 0.4]})\n",
    "\n",
    "  # Plot input spikes\n",
    "\n",
    "  neuron_numbers = list(range(59))\n",
    "  currents = spk_in.tolist()  # Currents fed to each neuron\n",
    "    \n",
    "  # Create bar chart\n",
    "  ax[0].bar(neuron_numbers, currents, color='skyblue')\n",
    "  ax[0].set_ylabel(\"Input Currents\")\n",
    "  ax[0].set_title(title)\n",
    "\n",
    "  # Plot hidden layer spikes\n",
    "  splt.raster(spk1_rec.reshape(num_steps, -1), ax[1], s = 0.1, c=\"black\")\n",
    "  ax[1].set_ylabel(\"Hidden Layer\")\n",
    "\n",
    "  # Plot output spikes\n",
    "  splt.raster(spk2_rec.reshape(num_steps, -1), ax[2], c=\"black\", marker=\"|\")\n",
    "  ax[2].set_ylabel(\"Output Spikes\")\n",
    "  ax[2].set_yticks([0, 1], ['left', 'right'])\n",
    "  ax[2].set_ylim([-0.5, 1.5])\n",
    "\n",
    "\n",
    "  plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "0f9d2b61-e4b2-441b-830d-bf478b3c9d23",
   "metadata": {},
   "outputs": [],
   "source": [
    "m = scipy.io.loadmat('BCICIV_1_mat\\BCICIV_calib_ds1d.mat', struct_as_record=True)\n",
    "\n",
    "sample_rate = m['nfo']['fs'][0][0][0][0]\n",
    "EEG = m['cnt'].T\n",
    "nchannels, nsamples = EEG.shape\n",
    "\n",
    "channel_names = [s[0] for s in m['nfo']['clab'][0][0][0]]\n",
    "event_onsets = m['mrk'][0][0][0]\n",
    "event_codes = m['mrk'][0][0][1]\n",
    "labels = np.zeros((1, nsamples), int)\n",
    "labels[0, event_onsets] = event_codes\n",
    "\n",
    "cl_lab = [s[0] for s in m['nfo']['classes'][0][0][0]]\n",
    "cl1 = cl_lab[0]\n",
    "cl2 = cl_lab[1]\n",
    "nclasses = len(cl_lab)\n",
    "nevents = len(event_onsets)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "a4b8e6b1-c762-49b1-a3c1-2834cde6d88b",
   "metadata": {},
   "outputs": [],
   "source": [
    "trials = {}\n",
    "\n",
    "win = np.arange(int(0.5*sample_rate), int(2.5*sample_rate))\n",
    "\n",
    "nsamples = len(win)\n",
    "\n",
    "for cl, code in zip(cl_lab, np.unique(event_codes)):\n",
    "    \n",
    "    cl_onsets = event_onsets[event_codes == code]\n",
    "\n",
    "    trials[cl] = np.zeros((nchannels, nsamples, len(cl_onsets)))\n",
    "    \n",
    "    for i, onset in enumerate(cl_onsets):\n",
    "        trials[cl][:,:,i] = EEG[:, win+onset]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "02cab55d-2ee6-4ea2-b173-78322d7b334c",
   "metadata": {},
   "outputs": [],
   "source": [
    "trials_filt = {cl1: bandpass(trials[cl1], 2, 30, sample_rate),\n",
    "               cl2: bandpass(trials[cl2], 2, 30, sample_rate)}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "886326e2-8579-4eb8-ac3b-66b19466e54b",
   "metadata": {},
   "outputs": [],
   "source": [
    "W = csp(trials_filt[cl1], trials_filt[cl2])\n",
    "trials_csp = {cl1: apply_mix(W, trials_filt[cl1]),\n",
    "              cl2: apply_mix(W, trials_filt[cl2])}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "1eddca94-1501-4c61-9109-f65f9705ff21",
   "metadata": {},
   "outputs": [],
   "source": [
    "trials_logvar = {cl1: logvar(trials_csp[cl1]),\n",
    "                 cl2: logvar(trials_csp[cl2])}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "6c51458d-aa0f-4ac3-b3b8-ba3cfb439822",
   "metadata": {},
   "outputs": [],
   "source": [
    "normalized_within_trial = {cl1: ((trials_logvar[cl1]-np.min(trials_logvar[cl1],axis=0))/(np.max(trials_logvar[cl1],axis=0)-np.min(trials_logvar[cl1],axis=0))),\n",
    "                 cl2: ((trials_logvar[cl2]-np.min(trials_logvar[cl2],axis=0))/(np.max(trials_logvar[cl2],axis=0)-np.min(trials_logvar[cl2],axis=0)))}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "7084c644-e665-4a10-8844-28f80b9e62e6",
   "metadata": {},
   "outputs": [],
   "source": [
    "max_value = np.max(trials_logvar[cl1]) or np.max(trials_logvar[cl2])\n",
    "min_value = np.min(trials_logvar[cl1]) or np.min(trials_logvar[cl2])\n",
    "\n",
    "\n",
    "normalized_trials_logvar = {cl1: ((trials_logvar[cl1]-min_value)/(max_value-min_value)),\n",
    "                 cl2: ((trials_logvar[cl2]-min_value)/(max_value-min_value))}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "533f8a99-a0c9-43a2-a338-dca47d0c6e27",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtype = torch.float\n",
    "device = torch.device(\"cuda\") if torch.cuda.is_available() else torch.device(\"cpu\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "24797234-e175-42b5-99e8-c09d4fe08257",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import torch\n",
    "from torch.utils.data import Dataset, DataLoader\n",
    "\n",
    "class CustomDataset(Dataset):\n",
    "    def __init__(self, data):\n",
    "        self.data = data\n",
    "\n",
    "    def __len__(self):\n",
    "        if len(self.data.shape) > 1:\n",
    "            return self.data.shape[1]\n",
    "        else:\n",
    "            return 1\n",
    "\n",
    "    def __getitem__(self, idx):\n",
    "        sample = self.data[1:,idx]\n",
    "        label = self.data[0,idx]\n",
    "        return torch.tensor(sample, dtype=torch.float32), torch.tensor(label, dtype=torch.long)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "475f6d0f-346c-4e1d-adfb-702767c659d4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([32, 59]) torch.Size([32])\n",
      "torch.Size([32, 59]) torch.Size([32])\n",
      "torch.Size([32, 59]) torch.Size([32])\n",
      "torch.Size([32, 59]) torch.Size([32])\n",
      "torch.Size([32, 59]) torch.Size([32])\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\USER\\AppData\\Local\\Temp\\ipykernel_11440\\1880285545.py:18: DeprecationWarning: an integer is required (got type numpy.float64).  Implicit conversion to integers using __int__ is deprecated, and may be removed in a future version of Python.\n",
      "  return torch.tensor(sample, dtype=torch.float32), torch.tensor(label, dtype=torch.long)\n"
     ]
    }
   ],
   "source": [
    "data = normalized_within_trial\n",
    "\n",
    "# Additional column for class identifiers\n",
    "class_column_left = np.zeros((1, 100))  # Assuming all instances belong to class 0\n",
    "class_column_right = np.ones((1, 100))  # Assuming all instances belong to class 0\n",
    "\n",
    "# Concatenate class column along with left and right arrays\n",
    "left_with_class = np.concatenate([class_column_left, data['left']], axis=0)\n",
    "right_with_class = np.concatenate([class_column_right, data['right']], axis=0)\n",
    "\n",
    "# Concatenate left and right arrays along axis 1\n",
    "result = np.concatenate([left_with_class, right_with_class], axis=1)\n",
    "\n",
    "cl1_split_idx = int(result.shape[1]*0.5*0.8)\n",
    "cl2_split_idx = int(result.shape[1]*0.5*1.8)\n",
    "\n",
    "train_data = np.concatenate([result[:, :cl1_split_idx],result[:, 100:cl2_split_idx]], axis=1)\n",
    "test_data = np.concatenate([result[:, cl1_split_idx:100],result[:, cl2_split_idx:]], axis=1)\n",
    "\n",
    "# Create CustomDataset instance\n",
    "train_dataset = CustomDataset(train_data)\n",
    "test_dataset = CustomDataset(test_data)\n",
    "\n",
    "# Create DataLoader\n",
    "batch_size = 32\n",
    "train_loader = DataLoader(train_dataset, batch_size=batch_size, shuffle=True)\n",
    "test_loader = DataLoader(test_dataset, batch_size=batch_size, shuffle=False)\n",
    "\n",
    "for samples, labels in train_loader:\n",
    "    # Your training code here\n",
    "    print(samples.shape, labels.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "83aa766e-32c4-4758-aff4-43f1d056ac83",
   "metadata": {},
   "outputs": [],
   "source": [
    "num_inputs = 59\n",
    "num_hidden = 1000\n",
    "num_outputs = 2\n",
    "beta = 0.99"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "029343f6-1daa-4052-831f-e38b989eca38",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "        0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "        0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.])\n",
      "torch.Size([500, 59])\n"
     ]
    }
   ],
   "source": [
    "# Define Network\n",
    "class Net(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "\n",
    "        # Initialize layers\n",
    "        self.fc1 = nn.Linear(num_inputs, num_hidden)\n",
    "        self.lif1 = snn.Leaky(beta=beta)\n",
    "        self.fc2 = nn.Linear(num_hidden, num_outputs)\n",
    "        self.lif2 = snn.Leaky(beta=beta)\n",
    "\n",
    "        self.spkIn = []\n",
    "        self.spk1 = []\n",
    "        self.spk2 = []\n",
    "\n",
    "        self.initialize_weights()\n",
    "\n",
    "    def forward(self, x):\n",
    "\n",
    "        # Initialize hidden states at t=0\n",
    "        mem1 = self.lif1.init_leaky()\n",
    "        mem2 = self.lif2.init_leaky()\n",
    "\n",
    "        self.spkIn = []\n",
    "        self.spk1 = []\n",
    "        self.spk2 = []\n",
    "        \n",
    "        # Record the final layer\n",
    "        spk2_rec = []\n",
    "        mem2_rec = []\n",
    "\n",
    "        for step in range(num_steps):\n",
    "            cur1 = self.fc1(x)\n",
    "            spk1, mem1 = self.lif1(cur1, mem1)\n",
    "            cur2 = self.fc2(spk1)\n",
    "            spk2, mem2 = self.lif2(cur2, mem2)\n",
    "\n",
    "            self.spkIn.append(x)\n",
    "            self.spk1.append(spk1)\n",
    "            self.spk2.append(spk2)\n",
    "            \n",
    "            spk2_rec.append(spk2)\n",
    "            mem2_rec.append(mem2)\n",
    "\n",
    "        return torch.stack(spk2_rec, dim=0), torch.stack(mem2_rec, dim=0)\n",
    "\n",
    "    def initialize_weights(self):\n",
    "        for m in self.modules():\n",
    "            if isinstance(m, nn.Linear) and m.weight.shape[0] == 1000:\n",
    "                m.weight.data[0] = torch.zeros(1, 59)\n",
    "                print(m.weight.data[0])\n",
    "                print(m.weight[:500,:].shape)\n",
    "        \n",
    "# Load the network onto CUDA if available\n",
    "net = Net().to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "4e0f2158-9ae7-4c1a-bfaf-db14c9fac710",
   "metadata": {},
   "outputs": [],
   "source": [
    "def print_batch_accuracy(data, targets, train=False):\n",
    "    output, _ = net(data.view(batch_size, -1))\n",
    "    _, idx = output.sum(dim=0).max(1)\n",
    "    acc = np.mean((targets == idx).detach().cpu().numpy())\n",
    "\n",
    "    if train:\n",
    "        print(f\"Train set accuracy for a single minibatch: {acc*100:.2f}%\")\n",
    "    else:\n",
    "        print(f\"Test set accuracy for a single minibatch: {acc*100:.2f}%\")\n",
    "\n",
    "def train_printer(\n",
    "    data, targets, epoch,\n",
    "    counter, iter_counter,\n",
    "        loss_hist, test_loss_hist, test_data, test_targets):\n",
    "    print(f\"Epoch {epoch}, Iteration {iter_counter}\")\n",
    "    print(f\"Train Set Loss: {loss_hist[counter]:.2f}\")\n",
    "    print(f\"Test Set Loss: {test_loss_hist[counter]:.2f}\")\n",
    "    print_batch_accuracy(data, targets, train=True)\n",
    "    print_batch_accuracy(test_data, test_targets, train=False)\n",
    "    print(\"\\n\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "c58a6092-5d1f-4e70-8e99-360890edfd2f",
   "metadata": {},
   "outputs": [],
   "source": [
    "loss = nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.Adam(net.parameters(), lr=5e-4, betas=(0.9, 0.999))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "130b0c22-0b24-4d3a-9b4d-fef18633182e",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\USER\\AppData\\Local\\Temp\\ipykernel_2204\\1880285545.py:18: DeprecationWarning: an integer is required (got type numpy.float64).  Implicit conversion to integers using __int__ is deprecated, and may be removed in a future version of Python.\n",
      "  return torch.tensor(sample, dtype=torch.float32), torch.tensor(label, dtype=torch.long)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1\n",
      "Epoch 0, Iteration 0\n",
      "Train Set Loss: 22.08\n",
      "Test Set Loss: 15.73\n",
      "Train set accuracy for a single minibatch: 68.75%\n",
      "Test set accuracy for a single minibatch: 56.25%\n",
      "\n",
      "\n",
      "2\n",
      "Epoch 0, Iteration 1\n",
      "Train Set Loss: 15.08\n",
      "Test Set Loss: 18.34\n",
      "Train set accuracy for a single minibatch: 68.75%\n",
      "Test set accuracy for a single minibatch: 62.50%\n",
      "\n",
      "\n",
      "3\n",
      "Epoch 0, Iteration 2\n",
      "Train Set Loss: 19.49\n",
      "Test Set Loss: 15.52\n",
      "Train set accuracy for a single minibatch: 62.50%\n",
      "Test set accuracy for a single minibatch: 68.75%\n",
      "\n",
      "\n",
      "4\n",
      "Epoch 0, Iteration 3\n",
      "Train Set Loss: 15.11\n",
      "Test Set Loss: 12.98\n",
      "Train set accuracy for a single minibatch: 90.62%\n",
      "Test set accuracy for a single minibatch: 81.25%\n",
      "\n",
      "\n",
      "5\n",
      "Epoch 0, Iteration 4\n",
      "Train Set Loss: 12.91\n",
      "Test Set Loss: 11.86\n",
      "Train set accuracy for a single minibatch: 81.25%\n",
      "Test set accuracy for a single minibatch: 90.62%\n",
      "\n",
      "\n",
      "6\n",
      "Epoch 1, Iteration 0\n",
      "Train Set Loss: 10.83\n",
      "Test Set Loss: 11.02\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 75.00%\n",
      "\n",
      "\n",
      "7\n",
      "Epoch 1, Iteration 1\n",
      "Train Set Loss: 11.11\n",
      "Test Set Loss: 10.41\n",
      "Train set accuracy for a single minibatch: 71.88%\n",
      "Test set accuracy for a single minibatch: 71.88%\n",
      "\n",
      "\n",
      "8\n",
      "Epoch 1, Iteration 2\n",
      "Train Set Loss: 10.84\n",
      "Test Set Loss: 9.90\n",
      "Train set accuracy for a single minibatch: 75.00%\n",
      "Test set accuracy for a single minibatch: 78.12%\n",
      "\n",
      "\n",
      "9\n",
      "Epoch 1, Iteration 3\n",
      "Train Set Loss: 12.34\n",
      "Test Set Loss: 9.05\n",
      "Train set accuracy for a single minibatch: 81.25%\n",
      "Test set accuracy for a single minibatch: 84.38%\n",
      "\n",
      "\n",
      "10\n",
      "Epoch 1, Iteration 4\n",
      "Train Set Loss: 9.81\n",
      "Test Set Loss: 8.27\n",
      "Train set accuracy for a single minibatch: 87.50%\n",
      "Test set accuracy for a single minibatch: 90.62%\n",
      "\n",
      "\n",
      "11\n",
      "Epoch 2, Iteration 0\n",
      "Train Set Loss: 8.10\n",
      "Test Set Loss: 7.87\n",
      "Train set accuracy for a single minibatch: 93.75%\n",
      "Test set accuracy for a single minibatch: 96.88%\n",
      "\n",
      "\n",
      "12\n",
      "Epoch 2, Iteration 1\n",
      "Train Set Loss: 7.88\n",
      "Test Set Loss: 7.55\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "13\n",
      "Epoch 2, Iteration 2\n",
      "Train Set Loss: 8.36\n",
      "Test Set Loss: 7.46\n",
      "Train set accuracy for a single minibatch: 96.88%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "14\n",
      "Epoch 2, Iteration 3\n",
      "Train Set Loss: 7.85\n",
      "Test Set Loss: 7.14\n",
      "Train set accuracy for a single minibatch: 96.88%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "15\n",
      "Epoch 2, Iteration 4\n",
      "Train Set Loss: 6.43\n",
      "Test Set Loss: 6.89\n",
      "Train set accuracy for a single minibatch: 96.88%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "16\n",
      "Epoch 3, Iteration 0\n",
      "Train Set Loss: 6.03\n",
      "Test Set Loss: 6.55\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "17\n",
      "Epoch 3, Iteration 1\n",
      "Train Set Loss: 6.83\n",
      "Test Set Loss: 6.08\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "18\n",
      "Epoch 3, Iteration 2\n",
      "Train Set Loss: 6.17\n",
      "Test Set Loss: 5.65\n",
      "Train set accuracy for a single minibatch: 96.88%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "19\n",
      "Epoch 3, Iteration 3\n",
      "Train Set Loss: 6.96\n",
      "Test Set Loss: 5.22\n",
      "Train set accuracy for a single minibatch: 96.88%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "20\n",
      "Epoch 3, Iteration 4\n",
      "Train Set Loss: 5.91\n",
      "Test Set Loss: 4.89\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "21\n",
      "Epoch 4, Iteration 0\n",
      "Train Set Loss: 5.08\n",
      "Test Set Loss: 4.65\n",
      "Train set accuracy for a single minibatch: 96.88%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "22\n",
      "Epoch 4, Iteration 1\n",
      "Train Set Loss: 5.17\n",
      "Test Set Loss: 4.47\n",
      "Train set accuracy for a single minibatch: 96.88%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "23\n",
      "Epoch 4, Iteration 2\n",
      "Train Set Loss: 5.06\n",
      "Test Set Loss: 4.38\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "24\n",
      "Epoch 4, Iteration 3\n",
      "Train Set Loss: 4.91\n",
      "Test Set Loss: 4.27\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "25\n",
      "Epoch 4, Iteration 4\n",
      "Train Set Loss: 4.83\n",
      "Test Set Loss: 4.19\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "26\n",
      "Epoch 5, Iteration 0\n",
      "Train Set Loss: 5.19\n",
      "Test Set Loss: 4.01\n",
      "Train set accuracy for a single minibatch: 96.88%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "27\n",
      "Epoch 5, Iteration 1\n",
      "Train Set Loss: 4.61\n",
      "Test Set Loss: 3.90\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "28\n",
      "Epoch 5, Iteration 2\n",
      "Train Set Loss: 3.86\n",
      "Test Set Loss: 3.76\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "29\n",
      "Epoch 5, Iteration 3\n",
      "Train Set Loss: 4.41\n",
      "Test Set Loss: 3.65\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "30\n",
      "Epoch 5, Iteration 4\n",
      "Train Set Loss: 3.96\n",
      "Test Set Loss: 3.56\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "31\n",
      "Epoch 6, Iteration 0\n",
      "Train Set Loss: 3.91\n",
      "Test Set Loss: 3.45\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "32\n",
      "Epoch 6, Iteration 1\n",
      "Train Set Loss: 3.72\n",
      "Test Set Loss: 3.38\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "33\n",
      "Epoch 6, Iteration 2\n",
      "Train Set Loss: 4.22\n",
      "Test Set Loss: 3.29\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "34\n",
      "Epoch 6, Iteration 3\n",
      "Train Set Loss: 3.74\n",
      "Test Set Loss: 3.22\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "35\n",
      "Epoch 6, Iteration 4\n",
      "Train Set Loss: 3.59\n",
      "Test Set Loss: 3.17\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "36\n",
      "Epoch 7, Iteration 0\n",
      "Train Set Loss: 3.48\n",
      "Test Set Loss: 3.12\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "37\n",
      "Epoch 7, Iteration 1\n",
      "Train Set Loss: 3.61\n",
      "Test Set Loss: 3.09\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "38\n",
      "Epoch 7, Iteration 2\n",
      "Train Set Loss: 3.37\n",
      "Test Set Loss: 3.04\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "39\n",
      "Epoch 7, Iteration 3\n",
      "Train Set Loss: 3.73\n",
      "Test Set Loss: 2.97\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "40\n",
      "Epoch 7, Iteration 4\n",
      "Train Set Loss: 3.13\n",
      "Test Set Loss: 2.94\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "41\n",
      "Epoch 8, Iteration 0\n",
      "Train Set Loss: 3.30\n",
      "Test Set Loss: 2.89\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "42\n",
      "Epoch 8, Iteration 1\n",
      "Train Set Loss: 3.16\n",
      "Test Set Loss: 2.84\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "43\n",
      "Epoch 8, Iteration 2\n",
      "Train Set Loss: 2.78\n",
      "Test Set Loss: 2.81\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "44\n",
      "Epoch 8, Iteration 3\n",
      "Train Set Loss: 3.57\n",
      "Test Set Loss: 2.79\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "45\n",
      "Epoch 8, Iteration 4\n",
      "Train Set Loss: 2.82\n",
      "Test Set Loss: 2.75\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "46\n",
      "Epoch 9, Iteration 0\n",
      "Train Set Loss: 2.88\n",
      "Test Set Loss: 2.72\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "47\n",
      "Epoch 9, Iteration 1\n",
      "Train Set Loss: 2.96\n",
      "Test Set Loss: 2.71\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "48\n",
      "Epoch 9, Iteration 2\n",
      "Train Set Loss: 2.78\n",
      "Test Set Loss: 2.68\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "49\n",
      "Epoch 9, Iteration 3\n",
      "Train Set Loss: 2.64\n",
      "Test Set Loss: 2.64\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "50\n",
      "Epoch 9, Iteration 4\n",
      "Train Set Loss: 3.25\n",
      "Test Set Loss: 2.60\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "51\n",
      "Epoch 10, Iteration 0\n",
      "Train Set Loss: 2.86\n",
      "Test Set Loss: 2.55\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "52\n",
      "Epoch 10, Iteration 1\n",
      "Train Set Loss: 2.80\n",
      "Test Set Loss: 2.53\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "53\n",
      "Epoch 10, Iteration 2\n",
      "Train Set Loss: 2.56\n",
      "Test Set Loss: 2.49\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "54\n",
      "Epoch 10, Iteration 3\n",
      "Train Set Loss: 2.45\n",
      "Test Set Loss: 2.46\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "55\n",
      "Epoch 10, Iteration 4\n",
      "Train Set Loss: 2.99\n",
      "Test Set Loss: 2.45\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "56\n",
      "Epoch 11, Iteration 0\n",
      "Train Set Loss: 3.01\n",
      "Test Set Loss: 2.43\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "57\n",
      "Epoch 11, Iteration 1\n",
      "Train Set Loss: 2.44\n",
      "Test Set Loss: 2.40\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "58\n",
      "Epoch 11, Iteration 2\n",
      "Train Set Loss: 2.75\n",
      "Test Set Loss: 2.39\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "59\n",
      "Epoch 11, Iteration 3\n",
      "Train Set Loss: 2.40\n",
      "Test Set Loss: 2.37\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "60\n",
      "Epoch 11, Iteration 4\n",
      "Train Set Loss: 2.14\n",
      "Test Set Loss: 2.35\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "61\n",
      "Epoch 12, Iteration 0\n",
      "Train Set Loss: 2.33\n",
      "Test Set Loss: 2.34\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "62\n",
      "Epoch 12, Iteration 1\n",
      "Train Set Loss: 2.57\n",
      "Test Set Loss: 2.32\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "63\n",
      "Epoch 12, Iteration 2\n",
      "Train Set Loss: 2.17\n",
      "Test Set Loss: 2.28\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "64\n",
      "Epoch 12, Iteration 3\n",
      "Train Set Loss: 2.57\n",
      "Test Set Loss: 2.26\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "65\n",
      "Epoch 12, Iteration 4\n",
      "Train Set Loss: 2.32\n",
      "Test Set Loss: 2.24\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "66\n",
      "Epoch 13, Iteration 0\n",
      "Train Set Loss: 2.40\n",
      "Test Set Loss: 2.22\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "67\n",
      "Epoch 13, Iteration 1\n",
      "Train Set Loss: 2.43\n",
      "Test Set Loss: 2.21\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "68\n",
      "Epoch 13, Iteration 2\n",
      "Train Set Loss: 2.46\n",
      "Test Set Loss: 2.19\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "69\n",
      "Epoch 13, Iteration 3\n",
      "Train Set Loss: 2.18\n",
      "Test Set Loss: 2.19\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "70\n",
      "Epoch 13, Iteration 4\n",
      "Train Set Loss: 1.95\n",
      "Test Set Loss: 2.17\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "71\n",
      "Epoch 14, Iteration 0\n",
      "Train Set Loss: 2.28\n",
      "Test Set Loss: 2.15\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "72\n",
      "Epoch 14, Iteration 1\n",
      "Train Set Loss: 2.35\n",
      "Test Set Loss: 2.12\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "73\n",
      "Epoch 14, Iteration 2\n",
      "Train Set Loss: 2.12\n",
      "Test Set Loss: 2.10\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "74\n",
      "Epoch 14, Iteration 3\n",
      "Train Set Loss: 2.02\n",
      "Test Set Loss: 2.07\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "75\n",
      "Epoch 14, Iteration 4\n",
      "Train Set Loss: 2.16\n",
      "Test Set Loss: 2.07\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "76\n",
      "Epoch 15, Iteration 0\n",
      "Train Set Loss: 2.20\n",
      "Test Set Loss: 2.05\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "77\n",
      "Epoch 15, Iteration 1\n",
      "Train Set Loss: 2.10\n",
      "Test Set Loss: 2.02\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "78\n",
      "Epoch 15, Iteration 2\n",
      "Train Set Loss: 1.98\n",
      "Test Set Loss: 2.01\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "79\n",
      "Epoch 15, Iteration 3\n",
      "Train Set Loss: 1.88\n",
      "Test Set Loss: 2.00\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "80\n",
      "Epoch 15, Iteration 4\n",
      "Train Set Loss: 2.29\n",
      "Test Set Loss: 1.98\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "81\n",
      "Epoch 16, Iteration 0\n",
      "Train Set Loss: 1.85\n",
      "Test Set Loss: 1.98\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "82\n",
      "Epoch 16, Iteration 1\n",
      "Train Set Loss: 1.98\n",
      "Test Set Loss: 1.96\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "83\n",
      "Epoch 16, Iteration 2\n",
      "Train Set Loss: 2.22\n",
      "Test Set Loss: 1.95\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "84\n",
      "Epoch 16, Iteration 3\n",
      "Train Set Loss: 2.00\n",
      "Test Set Loss: 1.94\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "85\n",
      "Epoch 16, Iteration 4\n",
      "Train Set Loss: 1.99\n",
      "Test Set Loss: 1.92\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "86\n",
      "Epoch 17, Iteration 0\n",
      "Train Set Loss: 2.09\n",
      "Test Set Loss: 1.92\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "87\n",
      "Epoch 17, Iteration 1\n",
      "Train Set Loss: 1.80\n",
      "Test Set Loss: 1.91\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "88\n",
      "Epoch 17, Iteration 2\n",
      "Train Set Loss: 1.96\n",
      "Test Set Loss: 1.89\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "89\n",
      "Epoch 17, Iteration 3\n",
      "Train Set Loss: 1.81\n",
      "Test Set Loss: 1.88\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "90\n",
      "Epoch 17, Iteration 4\n",
      "Train Set Loss: 2.00\n",
      "Test Set Loss: 1.87\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "91\n",
      "Epoch 18, Iteration 0\n",
      "Train Set Loss: 1.79\n",
      "Test Set Loss: 1.85\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "92\n",
      "Epoch 18, Iteration 1\n",
      "Train Set Loss: 2.07\n",
      "Test Set Loss: 1.84\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "93\n",
      "Epoch 18, Iteration 2\n",
      "Train Set Loss: 1.86\n",
      "Test Set Loss: 1.82\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "94\n",
      "Epoch 18, Iteration 3\n",
      "Train Set Loss: 1.76\n",
      "Test Set Loss: 1.81\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "95\n",
      "Epoch 18, Iteration 4\n",
      "Train Set Loss: 1.80\n",
      "Test Set Loss: 1.81\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "96\n",
      "Epoch 19, Iteration 0\n",
      "Train Set Loss: 1.68\n",
      "Test Set Loss: 1.79\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "97\n",
      "Epoch 19, Iteration 1\n",
      "Train Set Loss: 1.79\n",
      "Test Set Loss: 1.78\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "98\n",
      "Epoch 19, Iteration 2\n",
      "Train Set Loss: 2.16\n",
      "Test Set Loss: 1.77\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "99\n",
      "Epoch 19, Iteration 3\n",
      "Train Set Loss: 1.68\n",
      "Test Set Loss: 1.76\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n",
      "100\n",
      "Epoch 19, Iteration 4\n",
      "Train Set Loss: 1.64\n",
      "Test Set Loss: 1.75\n",
      "Train set accuracy for a single minibatch: 100.00%\n",
      "Test set accuracy for a single minibatch: 100.00%\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "num_epochs = 20\n",
    "loss_hist = []\n",
    "test_loss_hist = []\n",
    "counter = 0\n",
    "i=0\n",
    "num_steps = 25\n",
    "\n",
    "# Outer training loop\n",
    "for epoch in range(num_epochs):\n",
    "    iter_counter = 0\n",
    "    train_batch = iter(train_loader)\n",
    "\n",
    "    # Minibatch training loop\n",
    "    for data, targets in train_batch:\n",
    "        data = data.to(device)\n",
    "        targets = targets.to(device)\n",
    "        i=i+1\n",
    "        print(i)\n",
    "\n",
    "        # forward pass\n",
    "        net.train()\n",
    "        spk_rec, mem_rec = net(data.view(batch_size, -1))\n",
    "\n",
    "        # initialize the loss & sum over time\n",
    "        loss_val = torch.zeros((1), dtype=dtype, device=device)\n",
    "        for step in range(num_steps):\n",
    "            loss_val += loss(mem_rec[step], targets)\n",
    "\n",
    "        # Gradient calculation + weight update\n",
    "        optimizer.zero_grad()\n",
    "        loss_val.backward()\n",
    "        optimizer.step()\n",
    "\n",
    "        # Store loss history for future plotting\n",
    "        loss_hist.append(loss_val.item())\n",
    "\n",
    "        # Test set\n",
    "        with torch.no_grad():\n",
    "            net.eval()\n",
    "            test_data, test_targets = next(iter(test_loader))\n",
    "            test_data = test_data.to(device)\n",
    "            test_targets = test_targets.to(device)\n",
    "\n",
    "            # Test set forward pass\n",
    "            test_spk, test_mem = net(test_data.view(batch_size, -1))\n",
    "\n",
    "            # Test set loss\n",
    "            test_loss = torch.zeros((1), dtype=dtype, device=device)\n",
    "            for step in range(num_steps):\n",
    "                test_loss += loss(test_mem[step], test_targets)\n",
    "            test_loss_hist.append(test_loss.item())\n",
    "\n",
    "            # Print train/test loss/accuracy\n",
    "            if counter % 1 == 0:\n",
    "                train_printer(\n",
    "                    data, targets, epoch,\n",
    "                    counter, iter_counter,\n",
    "                    loss_hist, test_loss_hist,\n",
    "                    test_data, test_targets)\n",
    "            counter += 1\n",
    "            iter_counter +=1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "ca2bfef5-d6a0-444c-b08c-3a28ce8eb7e1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot Loss\n",
    "fig = plt.figure(facecolor=\"w\", figsize=(10, 5))\n",
    "plt.plot(loss_hist)\n",
    "plt.plot(test_loss_hist)\n",
    "plt.title(\"Loss Curves\")\n",
    "plt.legend([\"Train Loss\", \"Test Loss\"])\n",
    "plt.xlabel(\"Iteration\")\n",
    "plt.ylabel(\"Loss\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4879dc5d-5ae1-4601-8800-1b6a3b784003",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "a9ece0ec-708d-4bbb-817a-7e3887ebe8cb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Net(\n",
       "  (fc1): Linear(in_features=59, out_features=1000, bias=True)\n",
       "  (lif1): Leaky()\n",
       "  (fc2): Linear(in_features=1000, out_features=2, bias=True)\n",
       "  (lif2): Leaky()\n",
       ")"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create an instance of the network\n",
    "net = Net()\n",
    "\n",
    "# Load the trained model\n",
    "net.load_state_dict(torch.load('model.pth'))\n",
    "\n",
    "# Set the model to evaluation mode\n",
    "net.eval()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "8d493019-224a-4dd3-b46a-946b143d757c",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\USER\\AppData\\Local\\Temp\\ipykernel_2204\\1880285545.py:18: DeprecationWarning: an integer is required (got type numpy.float64).  Implicit conversion to integers using __int__ is deprecated, and may be removed in a future version of Python.\n",
      "  return torch.tensor(sample, dtype=torch.float32), torch.tensor(label, dtype=torch.long)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "59\n",
      "25\n",
      "2\n",
      "tensor(0)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x700 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "59\n",
      "25\n",
      "2\n",
      "tensor(1)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAscAAAJbCAYAAAAIUjkWAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8pXeV/AAAACXBIWXMAAA9hAAAPYQGoP6dpAACwkUlEQVR4nOzde3zO9f8/8Mdl57GDOWymxZQkc2ojrCLTakpJByUhhxxjVMuhMGFFoXL6kPMhfSmVQo5NSDnMx/GjshWxmYmd2Mb2/v3Brl+XDdtcz+d2XXvcb7frduO6rr3er/fpdT3fr6PJMAwDRERERESECqWdASIiIiKisoLBMRERERHRNQyOiYiIiIiuYXBMRERERHQNg2MiIiIiomsYHBMRERERXcPgmIiIiIjoGgbHRERERETXMDgmIiIiIrqGwTGRgoULF8JkMhX6evPNN4uV1o8//giTyYQff/zR/N7YsWNhMpmsnGtgzZo16NChA3x9feHs7AwfHx+EhYVh2bJluHz5stW3V9rWrl2LsWPHiqTdpk0btGnT5pbfy8zMxAcffIDGjRvD09MTHh4euOuuu/DCCy8gNja2RNvu0aMHateubfGeyWTCoEGDbvp3f/75J0wmExYuXFii7d6O/G2bTCasWLGiwOf513xKSop63oqrqOe+TZs2MJlMePzxxwt8ln88Pvzww2Jv/+LFixg7dqxFmVEW1K5dG08++WRpZ4OoAMfSzgBRebJgwQLce++9Fu/5+/uXUm5uzDAM9OzZEwsXLkT79u0xZcoUBAQEIDU1FVu3bsWAAQOQkpKCIUOGlHZWrWrt2rWYMWOGWIB8K7m5uQgPD8fBgwfx1ltvoXnz5gCA33//HWvWrMFPP/2E1q1bFzvdd999t0TnqkaNGvj5559x1113FftvrWnUqFF49tln4eTkVKr50PLDDz9gy5YtaNu2rVXSu3jxIqKjowGgSEE6UXnH4JhIUVBQEEJCQko7G7c0efJkLFy4ENHR0Rg9erTFZx06dEBUVBT++OOPUsqd/dq2bRt27tyJ+fPn49VXXzW//9hjj2HQoEHIy8srUbolDW5dXFzQokWLEv2ttURERGDdunWYPXs2Xn/99VLNC3D1wTErKwtubm4i6d9zzz24cuUKoqKisHv3bpEWodJ28eJFuLu7l3Y2iG6I3SqIygiTyVRojWXt2rXRo0ePYqXVq1cv+Pj44OLFiwU+a9u2LRo0aHDDv718+TI++OAD3HvvvXj33XcL/Y6fnx8efPBB8///+ecfDBgwADVr1oSzszPq1KmDUaNGITs72+Lv8pvylyxZgvr168Pd3R2NGzfGd999Z/G9/Cbzw4cP46WXXoKXlxd8fX3Rs2dPpKamWnzXMAzMnDkTTZo0gZubGypXroznnnsO8fHxBfK9fv16hIWFwcvLC+7u7qhfvz5iYmIAXO16MGPGDHM+819//vlnsbZjGAYmTZqEWrVqwdXVFffffz/WrVt3w+P9b+fOnQNwtca2MBUq/P8iO7+rzsaNG/Hqq6/Cx8cHFStWRIcOHQrkqbBuFdczDAMjR46Ek5MT5s6dC6DwbhXFOTcXLlwwX4uVKlXCE088gfj4+Bte64Vp27YtHnvsMbz33ntIT0+/5fc3bdqEsLAweHp6wt3dHaGhodi8eXORjkdh3ZPyr9nZs2ejfv36cHFxwaJFiwAA0dHReOCBB+Dj4wNPT0/cf//9mDdvHgzDKNK+FcbJyQkTJkzA3r178cUXX9zy+0lJSejbty/uuOMOODs7IzAwENHR0bhy5QqAq+ewWrVq5vzmX9c9evTA4cOHYTKZsHLlSnN6e/fuhclkKlBGPPXUUwgODjb/Py8vD5MmTcK9994LFxcXVK9eHd26dcPff/9t8Xdt2rRBUFAQtm3bhlatWsHd3R09e/a84f7MnDkTjo6OGDNmzK0PFpEQBsdEinJzc3HlyhWLl4QhQ4bg/PnzWL58ucX7R44cwdatWzFw4MAb/u2ePXvwzz//4Omnny5SrVVWVhYeeeQRLF68GMOGDcP333+Prl27YtKkSejUqVOB73///feYPn06xo0bhy+//BI+Pj545plnCg1mn332Wdxzzz348ssvMXz4cCxfvhxDhw61+E7fvn0RGRmJdu3a4euvv8bMmTNx+PBhtGrVCmfOnDF/b968eWjfvj3y8vIwe/ZsrFmzBoMHDzb/mL/77rt47rnnAAA///yz+ZUfqBZ1O9HR0Xj77bfx6KOP4uuvv0b//v3Rp08fHDt27JbHMiQkBE5OThgyZAiWLVuGxMTEW/5Nr169UKFCBSxfvhzTpk3Dr7/+ijZt2uDChQu3/Nt82dnZ6NKlC6ZPn441a9agT58+t/ybW52bvLw8dOjQAcuXL8fbb7+N1atX44EHHii0P+2tfPDBB0hJScHkyZNv+r2lS5ciPDwcnp6eWLRoEf7v//4PPj4+eOyxxwoEyMXx9ddfY9asWRg9ejR++OEHPPTQQwCuBp59+/bF//3f/+Grr75Cp06d8Prrr+O9994r8bYAoHPnzggODsY777xz0779SUlJaN68OX744QeMHj0a69atQ69evRATE2M+hzVq1MD69esBXL1W8q/rd999Fw0aNECNGjWwadMmc5qbNm2Cm5sbjhw5gtOnTwMArly5gtjYWLRr1878vf79+5uv82+//Rbvvfce1q9fj1atWhXoB56YmIiuXbuiS5cuWLt2LQYMGFBgXwzDwJtvvonIyEh89tln5m4gRKXCICJxCxYsMAAU+rp8+bJhGIYBwBgzZkyBv61Vq5bRvXt38/+3bt1qADC2bt1qfm/MmDHG9bdz69atjSZNmli8179/f8PT09NIT0+/YV5XrFhhADBmz55dpH2bPXu2AcD4v//7P4v3P/jgAwOAsWHDBvN7AAxfX18jLS3N/F5SUpJRoUIFIyYmpsD+TJo0ySLNAQMGGK6urkZeXp5hGIbx888/GwCMjz76yOJ7J0+eNNzc3IyoqCjDMAwjPT3d8PT0NB588EHz3xZm4MCBBY5jcbZz/vx5w9XV1XjmmWcsvrdjxw4DgNG6desbbjvfvHnzjEqVKpmvjxo1ahjdunUztm3bZvG9/GvqRtsaP368+b3u3bsbtWrVsvgeAGPgwIHGuXPnjAcffNCoWbOmsX//fovvJCQkGACMBQsWmN8r6rn5/vvvDQDGrFmzLL4XExNzw2u9sG1PnjzZMAzDePnll42KFSsaiYmJFvk4e/asYRiGkZmZafj4+BgdOnSwSCc3N9do3Lix0bx585sej3+nef1x8vLyMv7555+b5jc3N9e4fPmyMW7cOKNKlSoW11nr1q2LdO5bt25tNGjQwDAMw9i0aZMBwPj0008LPR6GYRh9+/Y1KlWqZPz1118W6Xz44YcGAOPw4cOGYRjG2bNnb3jMu3btatSpU8f8/3bt2hl9+vQxKleubCxatMgwjP9/TeXfy0ePHjUAGAMGDLBI65dffjEAGCNHjrTYJwDG5s2bC2y7Vq1axhNPPGFcvHjRePbZZw0vLy9j06ZNtzxORNJYc0ykaPHixdi9e7fFy9FRpuv/kCFDsH//fuzYsQMAkJaWhiVLlqB79+6oVKmS1bazZcsWVKxY0Vzrmi+/K8j1NXaPPPIIPDw8zP/39fVF9erV8ddffxVI+6mnnrL4f6NGjZCVlYXk5GQAwHfffQeTyYSuXbta1Mb7+fmhcePG5tH5O3fuRFpaGgYMGFCiPpxF3c7PP/+MrKwsvPzyyxZ/36pVK9SqVatI2+rZsyf+/vtvLF++HIMHD0ZAQACWLl2K1q1bF1pzeqNtbd269ZbbSkhIQMuWLZGWloZdu3ahcePGRcojcOtzkz+zxgsvvGDxvZdeeqnI2/i38ePH4/LlyzesUdy5cyf++ecfdO/e3eIc5eXl4fHHH8fu3buRmZlZom23bdsWlStXLvD+li1b0K5dO3h5ecHBwQFOTk4YPXo0zp07Zz4OJRUWFobw8HCMGzfuht1JvvvuOzzyyCPw9/e32OeIiAgAKNLsJmFhYYiPj0dCQgKysrKwfft2PP7443jkkUewceNGAFdrk11cXMxdqfKvreu7ezVv3hz169cvcM9Xrlz5hoMLz507h7Zt2+LXX3/F9u3bERYWdss8E0njgDwiRfXr11cbkPf000+jdu3amDFjBkJDQ7Fw4UJkZmbetEsFANx5550ArgZORXHu3Dn4+fkVCDqrV68OR0dHcz/afFWqVCmQhouLCy5dulTg/eu/6+LiAgDm7545cwaGYcDX17fQvNWpUwcAcPbsWQDAHXfcUZRdKqCo28nfVz8/vwLfKey9G/Hy8sJLL71kDiQPHz6Mdu3aYdSoUejTpw+8vb1vmq6fn1+B416YX3/9FSkpKZgwYUKxj82tzs25c+fg6OgIHx8fi+/d6BjeSu3atTFgwABMnz4dw4YNK/B5fteW6x/S/u2ff/5BxYoVi73twvqA//rrrwgPD0ebNm0wd+5cc5/fr7/+GhMmTCj0ei6uDz74APfffz8+/PBDiwGa+c6cOYM1a9bccBaPokxzl99VYtOmTQgMDMTly5fRtm1bnDlzxtw9ZNOmTQgNDTUPQrxZ33h/f/8CD7o36kMPAL/99hvOnz+PPn36ICgo6Jb5JdLA4JiojHBxcSkwgA1AkYKcwlSoUAEDBw7EyJEj8dFHH2HmzJkICwtDvXr1bvp3ISEh8PHxwTfffIOYmJhb1rRWqVIFv/zyCwzDsPhucnIyrly5gqpVq5Yo/0VRtWpVmEwm/PTTT+bg7N/y38sfkHT9YCFrbyc/YExKSirwnaSkpFsOiruRBg0a4MUXX8S0adPw22+/mad4u9m27r777lum27lzZ/j5+WHUqFHIy8vDO++8U6L8FaZKlSq4cuUK/vnnH4sAubD8FtU777yD+fPnY+TIkQUGjOVfZ59++ukNZ9jID8xdXV0LvdduFEwWdg+sWLECTk5O+O677+Dq6mp+/+uvvy7SvhRFkyZN8NJLL2HKlClo3759gc+rVq2KRo0aYcKECYX+fVGmibzjjjtwzz33YNOmTahduzZCQkLg7e2NsLAwDBgwAL/88gt27dplUWOff50nJiYWeKg6ffp0gXv+ZmVIy5Yt8fzzz6NXr14AgFmzZlkMPCUqDbwCicqI2rVr48CBAxbvbdmyBRkZGSVOs3fv3nB2dsbLL7+MY8eO3XLRB+DqaPm3334b//vf/244sCg5OdncXSMsLAwZGRkFgoLFixebP5fy5JNPwjAMnDp1CiEhIQVeDRs2BHC1q4GXlxdmz55905kErq/9LO52WrRoAVdXVyxbtszi73fu3Flot5HrnTt3Djk5OYV+9r///Q9AwYDnRtsq6ny277zzDqZNm4bRo0djxIgRRfqbosifj/n6GRcKW9CjqKpUqYK3334bq1atwq+//mrxWWhoKLy9vXHkyJFCz1FISAicnZ0BXL3XkpOTLQZS5uTk4IcffihyXkwmExwdHeHg4GB+79KlS1iyZEmJ968w48ePR05OTqHdSZ588kkcOnQId911V6H7m3+t3Oi6zteuXTts2bIFGzduxKOPPgrg6pRyd955J0aPHo3Lly9bDMbL7yKxdOlSi3R2796No0ePFvue7969O1asWIEFCxagW7duyM3NLdbfE1kba46JyohXXnkF7777LkaPHo3WrVvjyJEjmD59Ory8vEqcpre3N7p164ZZs2ahVq1a6NChQ5H+7q233sLRo0cxZswY/Prrr+jSpYt5EZBt27Zhzpw5iI6ORmhoKLp164YZM2age/fu+PPPP9GwYUNs374dEydORPv27S1+VK0tNDQUr732Gl599VXs2bMHDz/8MCpWrIjExERs374dDRs2RP/+/VGpUiV89NFH6N27N9q1a4c+ffrA19cXf/zxB/773/9i+vTpAGAOcj/44ANERETAwcEBjRo1KvJ2KleujDfffBPjx49H79698fzzz+PkyZMYO3ZskbpVbN26FUOGDMHLL7+MVq1aoUqVKkhOTsbnn3+O9evXo1u3bgVq6vbs2WOxrVGjRqFmzZqFzghwI0OGDEGlSpXw2muvISMjA5988sltz6/7+OOPIzQ0FG+88QbS0tIQHByMn3/+2fzQVNLawcjISMyYMaPA9HiVKlXCp59+iu7du+Off/7Bc889h+rVq+Ps2bP473//i7Nnz2LWrFkArtaYjx49Gi+++CLeeustZGVl4ZNPPilWUPbEE09gypQp6NKlC1577TWcO3cOH374YaEtC7cjMDAQ/fv3x8cff1zgs3HjxmHjxo1o1aoVBg8ejHr16iErKwt//vkn1q5di9mzZ+OOO+6Ah4cHatWqhW+++QZhYWHw8fFB1apVzS0ZYWFhmDlzJlJSUjBt2jRz+mFhYViwYAEqV65sMY1bvXr18Nprr+HTTz9FhQoVEBERgT///BPvvvsuAgICCswoUxTPPfcc3N3d8dxzz+HSpUv4/PPPzQ8zROpKczQgUXmRP7PA7t27b/id7OxsIyoqyggICDDc3NyM1q1bG/v37y/xbBX5fvzxRwOA8f777xc73998843xxBNPGNWqVTMcHR2NypUrG4888ogxe/ZsIzs72/y9c+fOGf369TNq1KhhODo6GrVq1TJGjBhhZGVlWaSHazMkXO/6fbx+JoJ8+ccxISHB4v358+cbDzzwgFGxYkXDzc3NuOuuu4xu3boZe/bssfje2rVrjdatWxsVK1Y03N3djfvuu8/44IMPzJ9nZ2cbvXv3NqpVq2aYTKYC2yrKdvLy8oyYmBgjICDAcHZ2Nho1amSsWbOmSDMWnDx50njnnXeM0NBQw8/Pz3B0dDQ8PDyMBx54wPj000+NK1euFDgWGzZsMF555RXD29vbcHNzM9q3b2/8/vvvFunebLaKf/v8888NR0dH49VXXzVyc3NvOltFUc7NP//8Y7z66quGt7e34e7ubjz66KPGrl27DADGxx9/fNNjUdjsDPnmzJljns3j+nzExsYaTzzxhOHj42M4OTkZNWvWNJ544glj5cqVFt9bu3at0aRJE8PNzc2oU6eOMX369BvOVlHYNWsYV6+HevXqGS4uLkadOnWMmJgYY968eQWOQ0lmq/i3s2fPGp6enoUej7NnzxqDBw82AgMDDScnJ8PHx8cIDg42Ro0aZWRkZJi/t2nTJqNp06aGi4uLAcDifjt//rxRoUIFo2LFikZOTo75/WXLlhkAjE6dOhXIU25urvHBBx8Y99xzj+Hk5GRUrVrV6Nq1q3Hy5Mki7ZNh/P/ZKv5t69atRqVKlYzHH3/cuHjx4o0PFpEgk2HcxmzlRFTmvfHGG5g1axZOnjxZ6GA4sk0LFy7Eq6++it27d9vEqov5li9fjpdffhk7duxAq1atSjs7REQFsFsFkZ3atWsXfvvtN8ycORN9+/ZlYEzqPv/8c5w6dQoNGzZEhQoVsGvXLkyePBkPP/wwA2MiKrMYHBPZqZYtW8Ld3R1PPvkkxo8fX9rZoXLIw8MDK1aswPjx45GZmYkaNWqgR48evB6JqExjtwoiIiIioms4lRsRERER0TUMjomIiIiIrmFwTERERER0DQfkFVFeXh5Onz4NDw+P254cn4iIiIiszzAMpKenw9/fv8SLDTE4LqLTp08jICCgtLNBRERERLdw8uTJAiuKFhWD4yLy8PAAcPVge3p6lnJuiIiIiOh6aWlpCAgIMMdtJcHguIjyu1J4enoyOCYiIiIqw26nCywH5BERERERXcPgmIiIiIjoGpsLjrdt24YOHTrA398fJpMJX3/99S3/JjY2FsHBwXB1dUWdOnUwe/Zs+YwSERERkc2xueA4MzMTjRs3xvTp04v0/YSEBLRv3x4PPfQQ4uLiMHLkSAwePBhffvmlcE6JiIiIyNbY3IC8iIgIREREFPn7s2fPxp133olp06YBAOrXr489e/bgww8/xLPPPiuUSyIiIiKyRTZXc1xcP//8M8LDwy3ee+yxx7Bnzx5cvny5lHJFRERERGWRzdUcF1dSUhJ8fX0t3vP19cWVK1eQkpKCGjVqFPp32dnZyM7ONv8/LS1NNJ9EREREVPrsPjgGCs51ZxhGoe//W0xMDKKjo0XzRcD7cSk3/Xx406pKOSEiIiIqB90q/Pz8kJSUZPFecnIyHB0dUaVKlRv+3YgRI5Cammp+nTx5UjqrRERERFTK7L7muGXLllizZo3Fexs2bEBISAicnJxu+HcuLi5wcXGRzh4RERERlSE2V3OckZGB/fv3Y//+/QCuTtW2f/9+nDhxAsDVGt9u3bqZv9+vXz/89ddfGDZsGI4ePYr58+dj3rx5ePPNN0sj+0RERERUhtlczfGePXvwyCOPmP8/bNgwAED37t2xcOFCJCYmmgNlAAgMDMTatWsxdOhQzJgxA/7+/vjkk084jRsRERERFWAy8ken0U2lpaXBy8sLqamp8PT0LO3s2A0OyCMiIiJrsUa8ZnPdKoiIiIiIpDA4JiIiIiK6hsExEREREdE1DI6JiIiIiK5hcExEREREdA2DYyIiIiKiaxgcExERERFdw+CYiIiIiOgaBsdERERERNcwOCYiIiIiuobBMRERERHRNY6lnQG6sffjUm76+fCmVZVyUjK2nn8iIiIqf1hzTERERER0DYNjIiIiIqJrGBwTEREREV3D4JiIiIiI6BoGx0RERERE1zA4JiIiIiK6hsExEREREdE1DI6JiIiIiK5hcExEREREdA2DYyIiIiKiaxgcExERERFdw+CYiIiIiOgaBsdERERERNcwOCYiIiIiuobBMRERERHRNTYbHM+cOROBgYFwdXVFcHAwfvrpp5t+f9myZWjcuDHc3d1Ro0YNvPrqqzh37pxSbomIiIjIFthkcPzFF18gMjISo0aNQlxcHB566CFERETgxIkThX5/+/bt6NatG3r16oXDhw9j5cqV2L17N3r37q2ccyIiIiIqy2wyOJ4yZQp69eqF3r17o379+pg2bRoCAgIwa9asQr+/a9cu1K5dG4MHD0ZgYCAefPBB9O3bF3v27FHOORERERGVZTYXHOfk5GDv3r0IDw+3eD88PBw7d+4s9G9atWqFv//+G2vXroVhGDhz5gxWrVqFJ5544obbyc7ORlpamsWLiIiIiOybzQXHKSkpyM3Nha+vr8X7vr6+SEpKKvRvWrVqhWXLlqFz585wdnaGn58fvL298emnn95wOzExMfDy8jK/AgICrLofRERERFT22FxwnM9kMln83zCMAu/lO3LkCAYPHozRo0dj7969WL9+PRISEtCvX78bpj9ixAikpqaaXydPnrRq/omIiIio7HEs7QwUV9WqVeHg4FCgljg5OblAbXK+mJgYhIaG4q233gIANGrUCBUrVsRDDz2E8ePHo0aNGgX+xsXFBS4uLtbfASIiIiIqs2yu5tjZ2RnBwcHYuHGjxfsbN25Eq1atCv2bixcvokIFy111cHAAcLXGmYiIiIgIUAqOFy1ahO+//978/6ioKHh7e6NVq1b466+/ip3esGHD8Nlnn2H+/Pk4evQohg4dihMnTpi7SYwYMQLdunUzf79Dhw746quvMGvWLMTHx2PHjh0YPHgwmjdvDn9//9vfQSIiIiKyCyrB8cSJE+Hm5gYA+PnnnzF9+nRMmjQJVatWxdChQ4udXufOnTFt2jSMGzcOTZo0wbZt27B27VrUqlULAJCYmGgx53GPHj0wZcoUTJ8+HUFBQXj++edRr149fPXVV9bZQSIiIiKyCyZDoV+Bu7s7/ve//+HOO+/E22+/jcTERCxevBiHDx9GmzZtcPbsWeks3La0tDR4eXkhNTUVnp6eKtt8Py7lpp8Pb1pVJR8lVZT82/o+EhERUdlhjXhNpea4UqVK5qWaN2zYgHbt2gEAXF1dcenSJY0sEBERERHdkspsFY8++ih69+6Npk2b4rfffjMvvnH48GHUrl1bIwtERERERLekUnM8Y8YMtGzZEmfPnsWXX36JKlWqAAD27t2Ll156SSMLRERERES3pFJznJaWhk8++aTAdGpjx47l4hpEREREVGao1BwHBgYiJaXgwKt//vkHgYGBGlkgIiIiIrolleD4RhNiZGRkwNXVVSMLRERERES3JNqtYtiwYQAAk8mE0aNHw93d3fxZbm4ufvnlFzRp0kQyC0RERERERSYaHMfFxQG4WnN88OBBODs7mz9zdnZG48aN8eabb0pmgYiIiIioyESD461btwIAXn31VXz88cdqi2cQEREREZWEymwVCxYs0NgMEREREdFtUQmOMzMz8f7772Pz5s1ITk5GXl6exefx8fEa2SAiIiIiuimV4Lh3796IjY3FK6+8gho1asBkMmlsloiIiIioWFSC43Xr1uH7779HaGioxuaIiIiIiEpEZZ7jypUrw8fHR2NTREREREQlphIcv/feexg9ejQuXryosTkiIiIiohJR6Vbx0Ucf4fjx4/D19UXt2rXh5ORk8fm+ffs0skFEREREdFMqwXHHjh01NkNEREREdFtUguMxY8ZobIaIiIiI6Lao9DkGgAsXLuCzzz7DiBEj8M8//wC42p3i1KlTWlkgIiIiIroplZrjAwcOoF27dvDy8sKff/6JPn36wMfHB6tXr8Zff/2FxYsXa2SDiIiIiOimVGqOhw0bhh49euD333+Hq6ur+f2IiAhs27ZNIwtERERERLekEhzv3r0bffv2LfB+zZo1kZSUpJEFIiIiIqJbUgmOXV1dkZaWVuD9Y8eOoVq1ahpZICIiIiK6JZXg+Omnn8a4ceNw+fJlAIDJZMKJEycwfPhwPPvssxpZICIiIiK6JZXg+MMPP8TZs2dRvXp1XLp0Ca1bt8bdd98NDw8PTJgwQSMLRERERES3pDJbhaenJ7Zv344tW7Zg3759yMvLw/3334927dppbJ6IiIiIqEjEg+MrV67A1dUV+/fvR9u2bdG2bVvpTRIRERERlYh4twpHR0fUqlULubm5Vk135syZCAwMhKurK4KDg/HTTz/d9PvZ2dkYNWoUatWqBRcXF9x1112YP3++VfNERERERLZNpc/xO++8Y7Ey3u364osvEBkZiVGjRiEuLg4PPfQQIiIicOLEiRv+zQsvvIDNmzdj3rx5OHbsGD7//HPce++9VskPEREREdkHlT7Hn3zyCf744w/4+/ujVq1aqFixosXn+/btK1Z6U6ZMQa9evdC7d28AwLRp0/DDDz9g1qxZiImJKfD99evXIzY2FvHx8fDx8QEA1K5du2Q7Q0RERER2SyU47tixo9XSysnJwd69ezF8+HCL98PDw7Fz585C/+bbb79FSEgIJk2ahCVLlqBixYp46qmn8N5778HNza3Qv8nOzkZ2drb5/4XN00xERERE9kVlQB4A9OzZEwEBAbedXkpKCnJzc+Hr62vxvq+v7w1X24uPj8f27dvh6uqK1atXIyUlBQMGDMA///xzw37HMTExiI6Ovu38EhEREZHtUBmQ9+GHH1p9QJ7JZLL4v2EYBd7Ll5eXB5PJhGXLlqF58+Zo3749pkyZgoULF+LSpUuF/s2IESOQmppqfp08edKq+SciIiKiskdlQF5YWBh+/PFHq6RVtWpVODg4FKglTk5OLlCbnK9GjRqoWbMmvLy8zO/Vr18fhmHg77//LvRvXFxc4OnpafEiIiIiIvum0uc4IiICI0aMwKFDhxAcHFxgQN5TTz1V5LScnZ0RHByMjRs34plnnjG/v3HjRjz99NOF/k1oaChWrlyJjIwMVKpUCQDw22+/oUKFCrjjjjtKsEdEREREZI9UguP+/fsDuDrLxPVMJlOxu1wMGzYMr7zyCkJCQtCyZUvMmTMHJ06cQL9+/QBc7RJx6tQpLF68GADQpUsXvPfee3j11VcRHR2NlJQUvPXWW+jZs+cNB+TZkvfjUm742fCmVRVzQlQ23eweAYp3n1gzLSIiKntUguO8vDyrpte5c2ecO3cO48aNQ2JiIoKCgrB27VrUqlULAJCYmGgx53GlSpWwceNGvP766wgJCUGVKlXwwgsvYPz48VbNFxERERHZNpXgWMKAAQMwYMCAQj9buHBhgffuvfdebNy4UThXRERERGTLVILjcePG3fTz0aNHa2SDiIjKAXZ9IaLboRIcr1692uL/ly9fRkJCAhwdHXHXXXcxOCYiIiKiMkElOI6LiyvwXlpaGnr06GEx4wQRERERUWlSmee4MJ6enhg3bhzefffd0soCEREREZGFUguOAeDChQtITU0tzSwQEREREZmpdKv45JNPLP5vGAYSExOxZMkSPP744xpZICIicLAaEdGtqATHU6dOtfh/hQoVUK1aNXTv3h0jRozQyAIRERER0S2pBMcJCQkamyEiIiIiui2ifY5zc3Nx4MABXLp0qcBnly5dwoEDB6y+eh4RERERUUmJBsdLlixBz5494ezsXOAzZ2dn9OzZE8uXL5fMAhERERFRkYl2q5g3bx7efPNNODg4FPjMwcEBUVFRmD59Orp27SqZDaJSx0FQREREtkG05vjYsWNo0aLFDT9v1qwZjh49KpkFIiIiIqIiE605zszMRFpa2g0/T09Px8WLFyWzQESKblZDztpxsga2whCRNNGa47p162Lnzp03/Hz79u2oW7euZBaIiIiIiIpMNDju0qUL3nnnHRw4cKDAZ//9738xevRodOnSRTILRERERERFJtqtYujQoVi3bh2Cg4PRrl073HvvvTCZTDh69Cg2bdqE0NBQDB06VDILRERERERFJhocOzk5YcOGDZg6dSqWL1+Obdu2wTAM3HPPPZgwYQIiIyPh5OQkmQUiIiIioiITXyHPyckJUVFRiIqKkt4UEZUzHJxFRETWJtrnmIiIiIjIljA4JiIiIiK6hsExEREREdE1DI6JiIiIiK5RCY7HjRtX6Ep4ly5dwrhx4zSyQERERER0SyrBcXR0NDIyMgq8f/HiRURHR2tkgYiIiIjolsSncgMAwzBgMpkKvP/f//4XPj4+GlkgIiKyCZyikKh0iQbHlStXhslkgslkwj333GMRIOfm5iIjIwP9+vWTzAIRERERUZGJBsfTpk2DYRjo2bMnoqOj4eXlZf7M2dkZtWvXRsuWLSWzQEQ26ma1Z6w5I7Iu1lYT/X+iwXH37t0BAIGBgWjVqpVVl4qeOXMmJk+ejMTERDRo0ADTpk3DQw89dMu/27FjB1q3bo2goCDs37/favkhIiIiItun0uc4MDAQiYmJN/z8zjvvLFZ6X3zxBSIjIzFz5kyEhobiP//5DyIiInDkyJGbppWamopu3bohLCwMZ86cKdY2iYiKozzUfLO2kYjskUpwXLt27UIH5OXLzc0tVnpTpkxBr1690Lt3bwBXu2/88MMPmDVrFmJiYm74d3379kWXLl3g4OCAr7/+uljbJCIiIioOPkDaJpXgOC4uzuL/ly9fRlxcHKZMmYIJEyYUK62cnBzs3bsXw4cPt3g/PDwcO3fuvOHfLViwAMePH8fSpUsxfvz4W24nOzsb2dnZ5v+npaUVK59ERNZSHmqhiYjKCpXguHHjxgXeCwkJgb+/PyZPnoxOnToVOa2UlBTk5ubC19fX4n1fX18kJSUV+je///47hg8fjp9++gmOjkXb5ZiYGM7BTERERFTOqATHN3LPPfdg9+7dJfrb67tp3Ggu5dzcXHTp0gXR0dG45557ipz+iBEjMGzYMPP/09LSEBAQUKK8EpVVbPIjIiKypBIcX98lwTAMJCYmYuzYsahbt26x0qpatSocHBwK1BInJycXqE0GgPT0dOzZswdxcXEYNGgQACAvLw+GYcDR0REbNmxA27ZtC/ydi4sLXFxcipU3IiIiIrJtKsGxt7d3oTW9AQEBWLFiRbHScnZ2RnBwMDZu3IhnnnnG/P7GjRvx9NNPF/i+p6cnDh48aPHezJkzsWXLFqxatQqBgYHF2j4RERER2S+V4Hjr1q0W/69QoQKqVauGu+++u8h9gP9t2LBheOWVVxASEoKWLVtizpw5OHHihHm1vREjRuDUqVNYvHgxKlSogKCgIIu/r169OlxdXQu8b8+s2XzOpngiIiKyVyrBcevWra2aXufOnXHu3DmMGzcOiYmJCAoKwtq1a1GrVi0AQGJiIk6cOGHVbRKRdfDhioiIyjK1AXnHjh3Dp59+iqNHj8JkMuHee+/FoEGDcO+995YovQEDBmDAgAGFfrZw4cKb/u3YsWMxduzYEm2XiIiIiOxXBY2NrFq1CkFBQdi7dy8aN26MRo0aYd++fWjYsCFWrlypkQUiIiIioltSqTmOiorCiBEjMG7cOIv3x4wZg7fffhvPP/+8RjaIiIiIiG5KJThOSkpCt27dCrzftWtXTJ48WSMLVERciYsKw37CRERUXqh0q2jTpg1++umnAu9v374dDz30kEYWiIiIiIhuSaXm+KmnnsLbb7+NvXv3okWLFgCAXbt2YeXKlYiOjsa3335r8V0iIiIiotKgEhznzyoxc+ZMzJw5s9DPgKtLQufm5mpkiYiIiIioAJXgOC8vT2MzROwzTURERLdFbZ5jIqKyjIMOiYgIUAyON2/ejM2bNyM5OblATfL8+fO1skFEREREdEMqwXF0dDTGjRuHkJAQ1KhRAyaTSWOzRERERETFohIcz549GwsXLsQrr7yisTkiIiIiohJRmec4JycHrVq10tgUEREREVGJqQTHvXv3xvLlyzU2RURERERUYirdKrKysjBnzhxs2rQJjRo1gpOTk8XnU6ZM0cgGEREREdFNqQTHBw4cQJMmTQAAhw4dsviMg/OorOKcyVSe8fon0sOpJMsWleB469atGpshIiIiIrotXASEiIiI1LG2lMoq0eC4U6dORfreV199JZkNIiIiIqIiEQ2Ovby8JJMnIiISx/7XROWLaHC8YMECyeSJiIiIiKyKfY7JJrBvGhGRJZaLRDJUFgEhIiIiIrIFrDkmIiKiMom141QaWHNMRERERHQNa46JiOwAa9hIA2fuoPKANcdERERERNfYbHA8c+ZMBAYGwtXVFcHBwfjpp59u+N2vvvoKjz76KKpVqwZPT0+0bNkSP/zwg2JuiYiIiMgW2GRw/MUXXyAyMhKjRo1CXFwcHnroIURERODEiROFfn/btm149NFHsXbtWuzduxePPPIIOnTogLi4OOWcExEREVFZZpPB8ZQpU9CrVy/07t0b9evXx7Rp0xAQEIBZs2YV+v1p06YhKioKzZo1Q926dTFx4kTUrVsXa9asUc45EREREZVlNhcc5+TkYO/evQgPD7d4Pzw8HDt37ixSGnl5eUhPT4ePj49EFomIiIjIRtncbBUpKSnIzc2Fr6+vxfu+vr5ISkoqUhofffQRMjMz8cILL9zwO9nZ2cjOzjb/Py0trWQZJiIiu8fZQmwDZ9ugorC54DifyWSy+L9hGAXeK8znn3+OsWPH4ptvvkH16tVv+L2YmBhER0ffdj6JiIjyMTgje2RvD4c2FxxXrVoVDg4OBWqJk5OTC9QmX++LL75Ar169sHLlSrRr1+6m3x0xYgSGDRtm/n9aWhoCAgJKnnEiIqJSwIBcBo+r/bK5PsfOzs4IDg7Gxo0bLd7fuHEjWrVqdcO/+/zzz9GjRw8sX74cTzzxxC234+LiAk9PT4sXEREREdk3m6s5BoBhw4bhlVdeQUhICFq2bIk5c+bgxIkT6NevH4Crtb6nTp3C4sWLAVwNjLt164aPP/4YLVq0MNc6u7m5wcvLq9T2g4iIiIjKFpsMjjt37oxz585h3LhxSExMRFBQENauXYtatWoBABITEy3mPP7Pf/6DK1euYODAgRg4cKD5/e7du2PhwoXa2SciIiKiMsomg2MAGDBgAAYMGFDoZ9cHvD/++KN8hoiIiIjI5tlscExEREREcuxtFoqiYnBMRETlUnn94Seim2NwTGSnOM0QEZUXfNAha7K5qdyIiIiIiKQwOCYiIiIiuobBMRERERHRNexzTFSGsJ8wERFR6WJwTERERFbDwXFk6xgcE90G/gjQ7eD1Q0RU9jA4pnLH1gMSW88/ERFRWcYBeURERERE1zA4JiIiIiK6hsExEREREdE1DI6JiIiIiK7hgDwiIiKicobz6t8Yg2MiIip1nIWFiMoKBsdERERE1/BBjdjnmIiIiIjoGtYcExEREdkA9hPWwZpjIiIiIqJrWHNMRESiWNtFRLaENcdERERERNcwOCYiIiIiuobBMRERERHRNQyOiYiIiIiuYXBMRERERHQNg2MiIiIiomtsNjieOXMmAgMD4erqiuDgYPz00083/X5sbCyCg4Ph6uqKOnXqYPbs2Uo5JSIiIiJbYZPB8RdffIHIyEiMGjUKcXFxeOihhxAREYETJ04U+v2EhAS0b98eDz30EOLi4jBy5EgMHjwYX375pXLOiYiIiKgss8ngeMqUKejVqxd69+6N+vXrY9q0aQgICMCsWbMK/f7s2bNx5513Ytq0aahfvz569+6Nnj174sMPP1TOORERERGVZTa3Ql5OTg727t2L4cOHW7wfHh6OnTt3Fvo3P//8M8LDwy3ee+yxxzBv3jxcvnwZTk5OBf4mOzsb2dnZ5v+npqYCANLS0m53F4osKyP9pp+npTnf8ntF+U5ppJX/PWumdavvldW0SmObzH/B7/FYlM42eS5l0iqNbfJcyqRV3G0WVWnkX0N+nGYYRskTMWzMqVOnDADGjh07LN6fMGGCcc899xT6N3Xr1jUmTJhg8d6OHTsMAMbp06cL/ZsxY8YYAPjiiy+++OKLL774srHXyZMnSxxr2lzNcT6TyWTxf8MwCrx3q+8X9n6+ESNGYNiwYeb/5+Xl4Z9//kGVKlVuuh0paWlpCAgIwMmTJ+Hp6am+/fKOx7908fiXHh770sXjX7p4/EtXSY6/YRhIT0+Hv79/ibdrc8Fx1apV4eDggKSkJIv3k5OT4evrW+jf+Pn5Ffp9R0dHVKlSpdC/cXFxgYuLi8V73t7eJc+4lXh6evIGLUU8/qWLx7/08NiXLh7/0sXjX7qKe/y9vLxua3s2NyDP2dkZwcHB2Lhxo8X7GzduRKtWrQr9m5YtWxb4/oYNGxASElJof2MiIiIiKp9sLjgGgGHDhuGzzz7D/PnzcfToUQwdOhQnTpxAv379AFztEtGtWzfz9/v164e//voLw4YNw9GjRzF//nzMmzcPb775ZmntAhERERGVQTbXrQIAOnfujHPnzmHcuHFITExEUFAQ1q5di1q1agEAEhMTLeY8DgwMxNq1azF06FDMmDED/v7++OSTT/Dss8+W1i4Um4uLC8aMGVOgqwfp4PEvXTz+pYfHvnTx+JcuHv/SVVrH32QYtzPXBRERERGR/bDJbhVERERERBIYHBMRERERXcPgmIiIiIjoGgbHRERERETXMDi2ATNnzkRgYCBcXV0RHByMn376qbSzZJe2bduGDh06wN/fHyaTCV9//bXF54ZhYOzYsfD394ebmxvatGmDw4cPl05m7VBMTAyaNWsGDw8PVK9eHR07dsSxY8csvsNzIGfWrFlo1KiRebL9li1bYt26debPeez1xMTEwGQyITIy0vwej7+csWPHwmQyWbz8/PzMn/PYyzt16hS6du2KKlWqwN3dHU2aNMHevXvNn2ufAwbHZdwXX3yByMhIjBo1CnFxcXjooYcQERFhMVUdWUdmZiYaN26M6dOnF/r5pEmTMGXKFEyfPh27d++Gn58fHn30UaSnpyvn1D7FxsZi4MCB2LVrFzZu3IgrV64gPDwcmZmZ5u/wHMi544478P7772PPnj3Ys2cP2rZti6efftr8A8Rjr2P37t2YM2cOGjVqZPE+j7+sBg0aIDEx0fw6ePCg+TMee1nnz59HaGgonJycsG7dOhw5cgQfffSRxarE6ufAoDKtefPmRr9+/Szeu/fee43hw4eXUo7KBwDG6tWrzf/Py8sz/Pz8jPfff9/8XlZWluHl5WXMnj27FHJo/5KTkw0ARmxsrGEYPAeloXLlysZnn33GY68kPT3dqFu3rrFx40ajdevWxpAhQwzD4LUvbcyYMUbjxo0L/YzHXt7bb79tPPjggzf8vDTOAWuOy7CcnBzs3bsX4eHhFu+Hh4dj586dpZSr8ikhIQFJSUkW58LFxQWtW7fmuRCSmpoKAPDx8QHAc6ApNzcXK1asQGZmJlq2bMljr2TgwIF44okn0K5dO4v3efzl/f777/D390dgYCBefPFFxMfHA+Cx1/Dtt98iJCQEzz//PKpXr46mTZti7ty55s9L4xwwOC7DUlJSkJubC19fX4v3fX19kZSUVEq5Kp/yjzfPhQ7DMDBs2DA8+OCDCAoKAsBzoOHgwYOoVKkSXFxc0K9fP6xevRr33Xcfj72CFStWYN++fYiJiSnwGY+/rAceeACLFy/GDz/8gLlz5yIpKQmtWrXCuXPneOwVxMfHY9asWahbty5++OEH9OvXD4MHD8bixYsBlM71b5PLR5c3JpPJ4v+GYRR4j3TwXOgYNGgQDhw4gO3btxf4jOdATr169bB//35cuHABX375Jbp3747Y2Fjz5zz2Mk6ePIkhQ4Zgw4YNcHV1veH3ePxlREREmP/dsGFDtGzZEnfddRcWLVqEFi1aAOCxl5SXl4eQkBBMnDgRANC0aVMcPnwYs2bNQrdu3czf0zwHrDkuw6pWrQoHB4cCT0bJyckFnqBIVv7IZZ4Lea+//jq+/fZbbN26FXfccYf5fZ4Dec7Ozrj77rsREhKCmJgYNG7cGB9//DGPvbC9e/ciOTkZwcHBcHR0hKOjI2JjY/HJJ5/A0dHRfIx5/HVUrFgRDRs2xO+//85rX0GNGjVw3333WbxXv35988QDpXEOGByXYc7OzggODsbGjRst3t+4cSNatWpVSrkqnwIDA+Hn52dxLnJychAbG8tzYSWGYWDQoEH46quvsGXLFgQGBlp8znOgzzAMZGdn89gLCwsLw8GDB7F//37zKyQkBC+//DL279+POnXq8Pgrys7OxtGjR1GjRg1e+wpCQ0MLTNv522+/oVatWgBKqewXGeZHVrNixQrDycnJmDdvnnHkyBEjMjLSqFixovHnn3+WdtbsTnp6uhEXF2fExcUZAIwpU6YYcXFxxl9//WUYhmG8//77hpeXl/HVV18ZBw8eNF566SWjRo0aRlpaWinn3D7079/f8PLyMn788UcjMTHR/Lp48aL5OzwHckaMGGFs27bNSEhIMA4cOGCMHDnSqFChgrFhwwbDMHjstf17tgrD4PGX9MYbbxg//vijER8fb+zatct48sknDQ8PD/PvLI+9rF9//dVwdHQ0JkyYYPz+++/GsmXLDHd3d2Pp0qXm72ifAwbHNmDGjBlGrVq1DGdnZ+P+++83T21F1rV161YDQIFX9+7dDcO4Op3MmDFjDD8/P8PFxcV4+OGHjYMHD5Zupu1IYccegLFgwQLzd3gO5PTs2dNczlSrVs0ICwszB8aGwWOv7frgmMdfTufOnY0aNWoYTk5Ohr+/v9GpUyfj8OHD5s957OWtWbPGCAoKMlxcXIx7773XmDNnjsXn2ufAZBiGIVMnTURERERkW9jnmIiIiIjoGgbHRERERETXMDgmIiIiIrqGwTERERER0TUMjomIiIiIrmFwTERERER0DYNjIiIiIqJrGBwTEREREV3D4JiIiIiI6BrH0s6ArcjLy8Pp06fh4eEBk8lU2tkhIiIiousYhoH09HT4+/ujQoWS1QEzOC6i06dPIyAgoLSzQURERES3cPLkSdxxxx0l+lsGx0Xk4eEB4OrB9vT0LOXcEBEREdH10tLSEBAQYI7bSoLBcRHld6Xw9PRkcExERERUht1OF1gOyCMiIiIiuobBMRERERHRNaUaHG/btg0dOnSAv78/TCYTvv76a4vPDcPA2LFj4e/vDzc3N7Rp0waHDx+2+E52djZef/11VK1aFRUrVsRTTz2Fv//+2+I758+fxyuvvAIvLy94eXnhlVdewYULF4T3joiIiIhsTakGx5mZmWjcuDGmT59e6OeTJk3ClClTMH36dOzevRt+fn549NFHkZ6ebv5OZGQkVq9ejRUrVmD79u3IyMjAk08+idzcXPN3unTpgv3792P9+vVYv3499u/fj1deeUV8/4iIiIjItpgMwzBKOxPA1Y7Tq1evRseOHQFcrTX29/dHZGQk3n77bQBXa4l9fX3xwQcfoG/fvkhNTUW1atWwZMkSdO7cGcD/n3Jt7dq1eOyxx3D06FHcd9992LVrFx544AEAwK5du9CyZUv873//Q7169YqUv7S0NHh5eSE1NZUD8oiIiIjKIGvEa2W2z3FCQgKSkpIQHh5ufs/FxQWtW7fGzp07AQB79+7F5cuXLb7j7++PoKAg83d+/vlneHl5mQNjAGjRogW8vLzM3ylMdnY20tLSLF5UPJmZmVi6dCkyMzNLOytllsYxsofzYA/7oIHH6daSk5PRv39/JCcnl3ZWyjReS7fG8rtobHEfymxwnJSUBADw9fW1eN/X19f8WVJSEpydnVG5cuWbfqd69eoF0q9evbr5O4WJiYkx91H28vLiAiAlsHr1aqSkpBToS24ttnjDXU/6GGltQ/pcrFixAuvWrcMXX3whkr69sIfjJH0tjRkzBomJiRg7dqxI+ho0yj7pa4nld9nYhj1cSxLKbHCc7/p56gzDuOXcddd/p7Dv3yqdESNGIDU11fw6efJkMXNOzzzzDKpVq2buKmNttnjDXU/6GGltQ/pcGIaBChUqoIz0AiuzpI+Txg+pdDAQFRWF9PR0vPXWWyLpa9Ao+6SvJY3AUppG2RoeHo5jx47h0UcfFUlf4zzYYvldZoNjPz8/AChQu5ucnGyuTfbz80NOTg7Onz9/0++cOXOmQPpnz54tUCv9by4uLuYFP7jwR8lUrFgRL7/8MipWrCiSfnZ2NhISEpCdnS2SvgbpY6S1DenC76mnnkKlSpXQoUMHkfTtxUsvvYTHH38cL774okj6Gj+k0sHAjh070KFDh5t2qyvrNIIN6WtJ+jxr0ChbN2zYgHr16mHjxo0i6WsE+LZYfpfZ4DgwMBB+fn4WF0ROTg5iY2PRqlUrAEBwcDCcnJwsvpOYmIhDhw6Zv9OyZUukpqbi119/NX/nl19+QWpqqvk75ZE9NGk5OzvjrrvugrOzs9g2pI+TPZwHAAgLC8Pp06fRtm1bkfSlfyA0aPR1lf6x1gho1qxZg7S0NHz33Xci6WsEA9L3tUawIX0tSZ9nwD7K19DQUKxZs0YsXrGHAF9CqQbHGRkZ2L9/P/bv3w/g6iC8/fv348SJEzCZTIiMjMTEiROxevVqHDp0CD169IC7uzu6dOkCAPDy8kKvXr3wxhtvYPPmzYiLi0PXrl3RsGFDtGvXDgBQv359PP744+jTpw927dqFXbt2oU+fPnjyySeLPFOFPbKHJi3pmg1AvvnSHrqGAFenXfTw8MDkyZNF0reHgMYe+rpqBDQataLSzbvS5astBhvX0zjPixYtwqxZs7Bo0SKxbUiTLls1aJTf1laqwfGePXvQtGlTNG3aFAAwbNgwNG3aFKNHjwZwtW9YZGQkBgwYgJCQEJw6dQobNmyAh4eHOY2pU6eiY8eOeOGFFxAaGgp3d3esWbMGDg4O5u8sW7YMDRs2RHh4OMLDw9GoUSMsWbJEd2fLGHsINuyhu4A9dA0BgOjoaPj7+4sFfhrnWvpBRfoYadC4Xtk15NbsofzWqNz45ZdfcPr0aezevVtsG9Kkyw2t2nVb6m8MlHJw3KZNGxiGUeC1cOFCAFcH0o0dOxaJiYnIyspCbGwsgoKCLNJwdXXFp59+inPnzuHixYtYs2ZNgZklfHx8sHTpUvOUbEuXLoW3t7fSXpZNGsGGPYyylS7ANbqGaKhYsSJatWolej1Jk34Qql69OmbOnFno7DnWIn1PaFyv0mWTRmApXcNuD+W3xj40adIEjo6OaNy4sdg2pEmXGxoPi7bYQlpm+xyT7bOHUbaA7BOvLQ5UKIw9dNOR7jetQfo82Mv1Kl2LZQ+zhthiU3hhTCaTzdVaatI4z7bYQsrgmMTYwyhb9h0sGnv4IZ0wYQIuXLiAmJiY0s5KiUmfB3u4XjUe5Oyha4hGza60w4cPw8PDA0eOHBFJ3x4G/GmcZ1tsIWVwXE7ZQ82Dxk3NvoNFZ+u1Mw0aNEBaWhruu+++0s5KidnDbBXStPZB8n6wh4dRjXJp4sSJeOCBBzBhwgSR9DW6C9jDjEm22OLE4LicsocuCRqka8oyMzOxfft2m15UQWMbGgW4h4cHWrRoYTHg19ZITxenMVuFNI19sIf+uvaw6qV0f12N7gLS15LG74MtlhsMjsspjdoT6ZtOY95Y6Roajem9NM619DY0fkilm8I1AvyRI0fi119/xTvvvCOSvsb0W9LHSWMfpOem1SBdfmsElvbQ7UH6N0jj94F9jslmaDzJ2UNgKV2zqzG9l8a5lt6GVlAmea41amiku4ZoTL9lD4MKo6OjER8fj+joaLFtSJMOmjT6odriLAnXs4fFWNjnmGyGRrAhfVNrBJbSAbjG9F5aiypIbkMjKJM+1xr9RKW7hmg059vDoEJb704GyB8njXtaulyyxaDvevawFLkIg4okNTXVAGCkpqaWdlasIiMjw1i6dKmRkZFR2lkpsYyMDGPJkiWi+3DmzBmjf//+xpkzZ0TS19gHjXMtvQ17ONca7OE8SG9D436wh2vpzJkzRr9+/bgPN2EPv6P2cJ6vZ414jTXH5RQHdBSNPUzADsjXZElfT/ZwrjX6P0qfB3tZMED6ftBoEZImXXOsMWZEeh/s4XdUoyXFFvt+MzgupzQuVg7ouDV7mKsZkP+hM+xgIJjGeZDeB42BZtLnWuM8JCQkICwsDAkJCWLbkCbd51h68Cggvw/28Duq8Ru0aNEizJo1C4sWLRLbhrUxOC6nNH4gpG86jf5e9jAlk0bhJ91f1x4GgtnDg9CkSZPg4eGByZMni6QPyA+Y0zgPvXv3RnJyMvr06SO2DWnSNYpNmzaFm5sbmjRpIpI+IL8P9vA7qvEbtG/fPmRlZSEuLk5sG1ZnlQ4e5QD7HJc9Gn2l7OE4aYiPjzfatm1rxMfHl3ZWSswezrX0Pmic5yVLlhhTp041li5dKpK+RrkRHx9vhIWFiR0neyj77GEshD3215Wg3QeffY6pxNhXqmjsYQlVjaa/LVu2wM/PD1u3bhXbhjTpc20PfY537NiBDh06YOfOnSLpA/I1ZRrN+dWrV0ePHj3E+hxrTGMJyPbN1ihbpbfB/rpFU7FiRbRq1cqmfkcZHJdTGjecdD8je1jK1h7OAyDf/1vjONnDAFLpfdC456Tnm9ZYJlz6XA8aNAi///47Bg4cKJI+YB+LONnDmBHpa0mjbJ0zZw6io6MxZ84csW1YG4Pjckrjh1q6n5E9PLVr9Fn75ZdfcPr0aezevVtsG9L9v+1hlgRDYVCh9D5oLBggXSvq4uKCatWqwcXFRSR9QP5cT506Fa6urvj4449F0gfsYxEn6ftBo/ZbuuJBo2z95ptvkJWVhW+//VZsG9bG4Lic0vihnjhxIh544AFMmDBBJH2NWizpwlWj5qF58+aoWbMmmjVrJraNsLAwnD59Gm3bthVJX+M4Sf8IaQwqlN4HjXJDY3EfadKDCjVqv+1hESd7aNGSplG2zpgxA56enpg+fbrYNqyNwXE5JR3MAPJzfa5atQoHDhzAqlWrRNIH9FaXk/T888+jYcOGeO6558S2MWHCBFy4cAExMTFi25A+ThpsfR80ll62xf6J19No1TKZTDZ/PUnnX7pFyx6Wpwbkz0NsbCy8vb0RGxsruh1rYnBcTmkEM9JP1RrdBaSDAY3+wBpN4dLTMmkcJ3v4Ic3OzkZKSgqysrJE0te4luwh4JCujduzZw/Onz+PPXv2iKQPyJffGgMjpctvjbn2pWncbzt27MBvv/2GHTt2iG3D2hgcl1Mac0xK92Vq0qQJHB0d0bhxY5H0AfkaII35HzUK8Pbt28PFxQUREREi6WscJ+luDxqtEPv378fly5dx4MABkfQ19kH6etWYHx2QrY3LT9tkMoltQ7r8tod5jqUfRgH561WzdVTyerU2BsdllPRTe48ePdC/f390795dJH1AfjUtDw8PtGjRAh4eHiLpA/L9mqX7ZQNAWloa4uPjkZqaKrYN6cUhRo0aBW9vb4wcOVIkfUC+j6VGVybp/uUa3Sqkg4GWLVti3759aNGihUj6gHxtXGhoKOrVqye6UqF02afR3Ut6Hw4fPgwPDw8cOXJEJH1A/p7TKJeaNWsGb29vhISEiG3D2hgcl1EaI0il+xlJd93Q+KGWrnmQ7pcNAN9++y3S09OxZs0asW1ID66xh3mUNboySQccGn1ppe/rgQMHIjU1Fa+//rpI+oB8bZxGYCl9rjXGjEjvg0blhnRXJo1ySWOGGGtjcFxGST/xagTf0iOqNX6opWu/Ndxxxx2oUKECAgICxLYhHeRrdA2Rbq3RmGFA+p7QmCFGeh/atm2L9PR0tGnTRiR9QL6Ljkbfb+l+0xpjRqSvV43Bo9IPWhrlklZXJmticFxGSf9AaEzfIt3tQWMfNJ6qpbVu3RqNGzfGww8/XNpZKdOkm8I1ugFJ3xMaD6TS+3D8+HH4+PggISFBJH1AvouORj/R/O1I0RgzIv0QoVHJJP2gpVGrqzGNpbUxOC6jNJ54pScvl24e1dgHjadqadKD5QD7mO9TOuDQmgZN8p7QqDmW3of69evj0qVLqFevnkj6GjT6iUoHfhpBmXSLk0YFDWD7U0ACtrcPDI7LKOkaGo1gxh72AbD9+UQ1ar+la101muWkg1eNpnDpJXk1ao6lHT16FG5ubjh27JjYNqTPg8Y9LR34adzT0rNJaFTQ2MPUhra4D2U6OL5y5QreeecdBAYGws3NDXXq1MG4ceOQl5dn/o5hGBg7diz8/f3h5uaGNm3a4PDhwxbpZGdn4/XXX0fVqlVRsWJFPPXUU/j777+1d6dYpAsmjeYge9gHjdHI0jSmTJKuddVolpMO/DSawqWX5NWqKZOksWKk9HnQuKcB2do+jXvaHspv6XJD4yFFqxuQVRll2Pjx440qVaoY3333nZGQkGCsXLnSqFSpkjFt2jTzd95//33Dw8PD+PLLL42DBw8anTt3NmrUqGGkpaWZv9OvXz+jZs2axsaNG419+/YZjzzyiNG4cWPjypUrRc5LamqqAcBITU216j6WloyMDGPp0qVGRkZGaWelxOLj4422bdsa8fHxYts4c+aM0b9/f+PMmTNi25Cmca7PnDlj9OvXT+w4ZWRkGEuWLBHdB+njpHEeNO4JadLn2h7Og8Y+LFmyxJg6daqxdOlSkfQ17ml7uR9svVyS/n24njXiNZNhlN1Q/sknn4Svry/mzZtnfu/ZZ5+Fu7s7lixZAsMw4O/vj8jISLz99tsArtYS+/r64oMPPkDfvn2RmpqKatWqYcmSJejcuTMA4PTp0wgICMDatWvx2GOPFSkvaWlp8PLyQmpqKjw9Pa2/s1Rs/fv3R2JiIvz9/TFz5szSzk65tnTpUqSkpKBatWp4+eWXbS59e2EPx0l6HzIzM7F69Wo888wzYs3h8+bNw5YtWxAWFoaePXuKbENaZmYmvv76a3Ts2FHkOGlcq/ZwP9gD7fNgjXitTHerePDBB7F582b89ttvAID//ve/2L59O9q3bw8ASEhIQFJSEsLDw81/4+LigtatW2Pnzp0AgL179+Ly5csW3/H390dQUJD5O4XJzs5GWlqaxcue2MMAqqioKKSnp+Ott94q7ayUmHTfRC3Sze0aA8Gkadxz0sdJYx/soTuWPSwrnJmZie3bt4uda40uOiw3ygZbPA9lOjh+++238dJLL+Hee++Fk5MTmjZtisjISLz00ksAgKSkJACAr6+vxd/5+vqaP0tKSoKzszMqV658w+8UJiYmBl5eXuaX5ByxpUHjB0L6pt6xYwc6dOhw04ecsm7kyJH49ddf8c4774htQ6twlWyE0hjMJk1jUIp0v2mtgTWS15I99JvWuKel+00D8jMY2MOgcOl7zh4G50so08HxF198gaVLl2L58uXYt28fFi1ahA8//BCLFi2y+N7163UbhnHLNbxv9Z0RI0YgNTXV/Dp58mTJd6QM0niSk76pbfFp9HoaA2sWLVqEWbNmFbhvrEn6YcseauLS0tKwe/du0WW8pQM/Q2FgjUYAXoZ7ExaJRuWG9KqXGuWS9P2gda1K3nP2cB4klOng+K233sLw4cPx4osvomHDhnjllVcwdOhQ8/Q1fn5+AFCgBjg5Odlcm+zn54ecnBycP3/+ht8pjIuLCzw9PS1e9kSjJk46oNF4GpV+qtZYBlZjJSpbLPyuJ32uNUbOS08tpTG/rnS5YYvTSl1P436TXv1No1ySvh80HtqlZ/XYsWMHfvvtN+zYsUMkfUBnyjtrK9PB8cWLF1GhgmUWHRwczFO5BQYGws/PzyI4ysnJQWxsrHm53+DgYDg5OVl8JzExEYcOHbLpJYFvl0YtljR7qP1etWoVDhw4gFWrVomkD+isRAXI1sZJz1cKyJ/riRMn4oEHHsCECRNE0geujsMICwsTW/3NHlaM1Kj91rhepWu/pWsU7WExFg3Sfb8dHR1RqVIlODk5iaQPyJdLEsp0cNyhQwdMmDAB33//Pf7880+sXr0aU6ZMwTPPPAPganeKyMhITJw4EatXr8ahQ4fQo0cPuLu7o0uXLgAALy8v9OrVC2+88QY2b96MuLg4dO3aFQ0bNkS7du1Kc/dKlT3M/6gRWEr/kGrUnmisRCUdWO7fvx+XL1/GgQMHRNIH5GuBqlevjpkzZ6J69eoi6QPAq6++imPHjqFXr14i6d999904deoU6tSpI5K+BnuYX1ej9js2NhYHDhzAtm3bRNL/448/ULNmTRw/flwkfQ0aD0FRUVFYv349oqKiRNKfPHkyIiIiMGnSJJH0AflySUKZDo4//fRTPPfccxgwYADq16+PN998E3379sV7771n/k5UVBQiIyMxYMAAhISE4NSpU9iwYQM8PDzM35k6dSo6duyIF154AaGhoXB3d8eaNWvg4OBQGrtVJNKzGIwaNQre3t4YOXKkSPqAfMGhEVhKNyNr1OpqFODSgaU91DJpzEzy+OOP48qVK0WeorK4Dh48iNzcXBw8eFAkfQ3JycmYP3++6HmQLl81Wv5Onz4NZ2dnnD59WiT9vn374syZM3jttddE0gfk7zmNh/ZLly4hJSUFly5dEklfuvsMALRt2xbp6elo06aN2DasrUwHxx4eHpg2bRr++usvXLp0CcePH8f48eMtVnIxmUwYO3YsEhMTkZWVhdjYWAQFBVmk4+rqik8//RTnzp3DxYsXsWbNmjI/+4T0SGGNmR6ka080AqZJkybBw8MDkydPFklfo1ZXo5VAepWluLg4ZGZmYv/+/SLpA/L7oDH6v2rVqnj00UdRtWpVkfSvXLmCrKwsXLlyRSR9Db1790ZycjL69Okjtg3p8lXjnp49ezb8/Pwwa9YskfSHDBmC8+fPIzIyUiR9QP6e01ht0cHBAa6urmLdHjQGdx4/fhw+Pj7sVkG3T3qksMaADunaE41mOenzoFGrq9HXVbqp+s8//0ROTg7+/PNPkfQB+X2QvpYA+ZYO6R9qQP4h5ZNPPkFeXh4+/vhjkfQB+fJV456OjY2Ft7c3YmNjRdLPzc3FlStXzGOIJEjfcxoDqkNDQ1GvXj2xMVIasYDWuBdrYnBcRmn0T5Qe0LFlyxb4+flh69atIukPHToUWVlZojUP0k1OGs1yGs1m0qOR77jjDphMJtSsWVMkfUB+4IvGPS3d0tGoUSM4ODgUaJ2zJukAf9euXWjSpAl++eUXkfQB+ftB456WnsXgmWeeQeXKlW16xg2NWZ969OiB/v37o3v37iLpS5d7+Uwmk01NoVis4PjKlSuIjo62uzl/yyN7WCVq+vTpqFu3LmbMmCGSPiA/YlujWU5j8I70NGiVKlVCjRo1LMYSWJv0wBeNyfala8o0WmukA/yzZ89i8+bNon2Opc+1RvmdH8jcas2AkvL09ESdOnVEp0mVLvs0pnKTftDSWIhKoxLI2ooVHDs6OmLy5MnIzc2Vyg8p0ZgGTbp5VKOZWrr2RLpWALCPhRs0RlTn5OQgIyNDrD+txkOKdO20xkBe6ft6/fr1cHJywg8//CCSPmAfCyA1a9YM3t7eCAkJEUlf+vcB0Cn7bF2DBg2QlpaG++67T2wb5aJbRbt27fDjjz8KZIU0aUyDpjFlknShJ117ojE5+lNPPYVKlSqhQ4cOYtuQHj2v0Yws3bdPo5ZJenS+1pLtkvf1ggULUK9ePcyfP19sG9LnWqP89vDwQIsWLcRaazTKJekuOhoBvnQrhMagcI1tWFuxg+OIiAiMGDECb775Jj7//HN8++23Fi+yDfv27UNWVhbi4uLEtiEd+GmM/pcOmDRorCQoPXpeoxlZuhZf44dU+p7QGLwjvQ+BgYHYtGkTAgMDRdIH5M+1RvktXbmhUS5Jd9HRqACSLvs0yiWNbVidUUwmk+mGrwoVKhQ3OZuRmppqADBSU1NLOytWcebMGaN///7GmTNnxLaRkZFhLFmyxMjIyBBJ/9ChQ8Z9991nHDp0SCR9w7i6D0uXLhXbB+ljlL8NyX0wDMOIj4832rZta8THx9tk+hrs4TycOXPG6Nevn2i5IV022cM9p3E/SJ9rjfvBHq4l6XOtdR6ky41/s0a8Vuya47y8vBu+2BfZdmiMnJd+4tUYkCc9klejH6pG1w3p5naNZYuluyTYw3nQGLwjXTZp3HOAbNcQje4tGi1zksdIYxsa15J07bdGuaTRSmBttzWVm+TcrGT7pAeNREVFIT09HW+99ZZI+oD8D4RGP1QNoaGhWLNmjVj3E41BI9LnWmO2Cunz0LRpU7i5uaFJkyYi6WvQuOfsYUCedPmq0VVK+mFO41qSHqBqD+WShGIHx7m5uXjvvfdQs2ZNVKpUCfHx8QCAd999F/PmzbN6BssrjQtW2rJly7Bp0yYsW7ZMJH2N2hONAFyaxrLF0rUbgPw8mdI/QtLTAgJX9yE+Ph7R0dEi6WvMrqJxvUqTDpo0auK+/vprnDhxQix41ei/rvFQrUGy3NN4SJEulyQUOzieMGECFi5ciEmTJll0rm7YsCE+++wzq2auPNO4YKV98803yMrKEhuoqVG4ao3Ol6TRFC4dWGoslyvdnK8xiEp6OjqNJliN61Wa9AAkjbJPuvwG5LtVSM+4YQ+DbDVaITS6z1hbsYPjxYsXY86cOXj55Zfh4OBgfr9Ro0b43//+Z9XMlWfShZ9GzfTkyZNhGIbY3LQaP9S22Bx0PY3aE+mp1jSWy5W+J/r27YszZ87gtddeE0kfkJ9dRaPcuPvuu3Hq1CnUqVNHJH2NgEZ6mjKNVc1mzJgBT09PTJ8+XSR9jQog6ancNGarGDRoEH7//XcMHDhQJH2NVf7Gjh2LOnXqYMyYMWLbsLZiB8enTp3C3XffXeD9vLw8XL582SqZIvnAT6Ng+s9//gNfX1/MnTtXbBvSpLsLaPxQa8wxKX092cMA0vyuOW+//bZI+gDw/PPPo2HDhnjuuedE0tcoN44ePQo3NzccO3ZMJH2NgEa624PGYDnpZbY1ar+lB/JqVNBMnToVrq6u+Pjjj0XSNxQWStmyZQv8/PywdetWsW1YW7GD4wYNGuCnn34q8P7KlSvRtGlTq2SK5GtoNAom6RogDdJ9jjUmws/OzkZKSoroAFp7aOmQbiV4/PHHceXKFTz22GMi6QPyi0NolBsaq2lJN/NKN1VrrA4qvbCPRu239ABSjf7x0vsgXbsO2ObA82IHx2PGjMGgQYPwwQcfIC8vD1999RX69OmDiRMnYvTo0RJ5LJc0amikfyD++OMP1KxZE8ePHxdJXyNgkn7i1WjS0uivC9j+oBHpVgIXFxd4enqKthJo9Gu29X6iGtNvaQyYkz4P0uWGRt9y6QGkGjX40q1BGtNk2qJiB8cdOnTAF198gbVr18JkMmH06NE4evQo1qxZI9qhu7yRrqHRCDak+4lq/MhJP/FqNGmNGjUK3t7eGDlypNg2pK8njUEj0n37NB5SpO85jXJDujVFukYUkC+/NYIy6XJDayyEZLcH6TIDkK9AsYfpGSWUaJ7jxx57DLGxscjIyMDFixexfft2hIeHWztv5Zr0Ta3RPCo9SMsWm2qup9H/UaO/l3TwqlETJ923T+MhJTMzE8eOHbPp7ljS53rPnj04f/489uzZI5J+PluedhCQn6lHYyyEdLcHjYWopCtQpGumAZ2ufdZW7OC4R48e2LZtm0ReSJl0s5zWSlSSpAfMafS706idlq7d0Jg1RLoGReMh5dVXX8WxY8fQq1cvsW1IlxvSAfiVK1eQlZUlNt0dIF/DLl3xAMg/8GoETNI17Brz4Eu3pGhUPGg9kFpTsYPj9PR0hIeHo27dupg4cSJOnTolkS8SptWnWTIo05jpQbpmV6PfnUbttHQtvsYiI9L9EzUeUp566im4u7uL/ZDaw3Lnrq6uqFq1Ktzc3ETSB+QDS43yWzpo0uhmJB28asyDL30eNFqD8ss8k8kktg1rK3Zw/OWXX+LUqVMYNGgQVq5cidq1ayMiIgKrVq3iVG42ROOGkA7KNGZ6kP6h1ujvpTHdkHQtkNZKhZKBq8ZDSteuXREWFoaXX35ZJH2NAF/a5MmTERERITb/OiDfkqLRB196Gxpzl69btw5ZWVlYt26dSPoav6PS29D4fWjWrBm8vb0REhIitg1rK1Gf4ypVqmDIkCGIi4vDr7/+irvvvhuvvPIK/P39MXToUPz+++/WzidZmcYNAcgGGxrNQdI0+ntpkK4F0uiSoNEULn3PSQdlGgG+NI0uCdIPERpln/S1pDF3uT3M3qL1Wy1Jo3+5tZUoOM6XmJiIDRs2YMOGDXBwcED79u1x+PBh3HfffZg6daq18kg2SjrY0Hhql6YxlZsG6cFmGjWW9nA9aRwn6WBAeopGjS4J0g8RWkv+Sl5LGlNx2sPsLfZAowuktRU7OL58+TK+/PJLPPnkk6hVqxZWrlyJoUOHIjExEYsWLcKGDRuwZMkSjBs3TiK/ZEPsoTlIugC3h2ZqQL7vnVaNpa2fB+njpBEM2MO0gNJlk0bNsXS3NY1rSbp2WuNasge22OJU7OC4Ro0a6NOnD2rVqoVff/0Ve/bsQb9+/SwmbH/sscfg7e1tzXySlWk8tUv/QGjsg/QAJI1+0xo0B3VIkT7X9nDPaQQD0teS9CqCGjTuN+lWLY19kL7nNFr+NMoNDbZW8VDs4Hjq1Kk4ffo0ZsyYccNBRJUrV0ZCQsLt5o0E2UNzkD3MuGEP/aYB+aDMHs61PUxtqHW9Sv6QavRD1SAdbEjfDxotf9LlhkbLn/Q+aATfthhvFDs4fuWVV+Dq6iqRl0KdOnUKXbt2RZUqVeDu7o4mTZpg79695s8Nw8DYsWPh7+8PNzc3tGnTBocPH7ZIIzs7G6+//jqqVq2KihUr4qmnnsLff/+ttg9lkT30r7SHGTfs4TwA8gWsPZxre+hCo3EepH9INWZJYL/pskH6etU4RvawWq4tdj8p0YC83bt3IyoqCi+++CI6depk8bKm8+fPIzQ0FE5OTli3bh2OHDmCjz76yKLLxqRJkzBlyhRMnz4du3fvhp+fHx599FGkp6ebvxMZGYnVq1djxYoV2L59OzIyMvDkk08iNzfXqvm1JunCVeOpXXp1InuYccMezgOgV7MrSXpBFnvpQiN9HqR/SDVmq7CHftOA7LnW6i4gvQ/SizjZQ1cpW2whLXZwvGLFCoSGhuLIkSNYvXo1Ll++jCNHjmDLli3w8vKyauY++OADBAQEYMGCBWjevDlq166NsLAw3HXXXQCuXvTTpk3DqFGj0KlTJwQFBWHRokW4ePEili9fDgBITU3FvHnz8NFHH6Fdu3Zo2rQpli5dioMHD2LTpk1Wza812UMTrMYCF9LsoR+qxnmQrt3QuB+kV9Oyh/6JGudB+odUYx+k7weNa0n6OGmcB+mHFHv4jdO4lspFzfHEiRMxdepUfPfdd3B2dsbHH3+Mo0eP4oUXXsCdd95p1cx9++23CAkJwfPPP4/q1aujadOmmDt3rvnzhIQEJCUlITw83Pyei4sLWrdubR41v3fvXly+fNniO/7+/ggKCrrpyPrs7GykpaVZvDTZQxPs3XffjVOnTqFOnTqlnZUSs4d+qA0aNEBaWhruu+8+sW0AsjU00ivwAUB0dDT8/f3FgmONfZAOBjT2QTqw1NgH6RpFrSn7JLeRlpaG3bt3IzU1VSR9QP5a0ljESfqBV+N+sMUpS4sdHB8/fhxPPPEEgKuBaGZmJkwmE4YOHYo5c+ZYNXPx8fGYNWsW6tatix9++AH9+vXD4MGDsXjxYgBAUlISAMDX19fi73x9fc2fJSUlwdnZGZUrV77hdwoTExMDLy8v8ysgIMCau3ZL0n2ZNGos//jjD9SsWRPHjx8XSV9jH6TPg0bB5OHhgRYtWljMKGNtixYtwqxZs7Bo0SKR9DXmyZRubtfYB+kaGo19kA4sNfYhKioK69evR1RUlEj6Gn1dpbsBaSwfDcg+tEsvOQ/IV6Bo3A8aD0LWVuzg2MfHx9yft2bNmjh06BAA4MKFC7h48aJVM5eXl4f7778fEydORNOmTdG3b1/06dMHs2bNsvje9et1G4ZxyzW8b/WdESNGIDU11fw6efJkyXekBOxh9L/0wBeNWlfp86BRMGn8kErPAKDRX1f6ntA4D9JdEjT2QbqpWmMf7GHVNOlrSXrhIMA+Vr2UrsHXuB/279+Py5cv48CBA2LbsLZiB8cPPfSQ+WZ54YUXMGTIEPTp0wcvvfQSwsLCrJq5GjVqFGgKrl+/Pk6cOAEA8PPzA4ACNcDJycnm2mQ/Pz/k5OTg/PnzN/xOYVxcXODp6Wnxsicao86lJ2C3h64nGgWTRgEu/UOnMaDDHhatsYd9kG6q1tiHqKgouLu746233hLbhjTpa0l64SDAPmYDkv6N0Lgfmjdvjpo1a6JZs2Zi27C2YgfH06dPN5+kESNG4M0338SZM2fQqVMnzJs3z6qZCw0NxbFjxyze++2331CrVi0AQGBgIPz8/Cx+MHNychAbG4tWrVoBAIKDg+Hk5GTxncTERBw6dMj8nfJKOqiUniVBozZRo+uGrS/HCwBbtmyBn58ftm7dKpJ+aGgo1qxZI3rPSv9IaMwaAtjeZPvXa9++PVxcXBARESGSvsZ5mD59OurWrYsZM2aIpK9xT0t3b7GHe9oezoMG6XtaQom6Vfj7+1/94woVEBUVhW+//RbvvfceDh48aNXMDR06FLt27cLEiRPxxx9/YPny5ZgzZw4GDhwI4Gp3isjISEycOBGrV6/GoUOH0KNHD7i7u6NLly4AAC8vL/Tq1QtvvPEGNm/ejLi4OHTt2hUNGzZEu3btrJpfW6LRrUJ69L9GbaL0cdLoGqKxDela/EmTJsHDwwOTJ08WSV+D9P0A2MeCARMmTMCFCxcQExMjkr7GeZAe3GkP5bfGPW0Ps7doXK/SbLL8Nqxk//79RoUKFayVnNmaNWuMoKAgw8XFxbj33nuNOXPmWHyel5dnjBkzxvDz8zNcXFyMhx9+2Dh48KDFdy5dumQMGjTI8PHxMdzc3Iwnn3zSOHHiRLHykZqaagAwUlNTb3ufyoKMjAxj6dKlRkZGhtg2zpw5Y/Tv3984c+aMWPr9+vUTS98w5I/T3Llzja5duxqfffaZSPpa25A+F9LXkgaNfZA+D5999pnRpUsXY968eSLpG4ZhTJ8+3QgNDTVmzJghkr69XEvSZZ/0cTp06JBx3333GYcOHRJJ3zAMY8mSJcbUqVONpUuXiqQ/Y8YMo2XLlsbMmTNF0jcM+7leNffBGvFaiRYB0fTkk0/i4MGDyMrKwtGjR9GnTx+Lz00mE8aOHYvExERkZWUhNjYWQUFBFt9xdXXFp59+inPnzuHixYtYs2aN+uwTxWUPi4AAsk289rCUbVhYGE6fPo22bdva9DY0zoXkedCgsfjEqlWrcODAAaxatUokfY3ZVZ5//nk0bNgQzz33nEj6GudBuuuGxv0mfZwmT56MixcvitYm2sPsLdLnQaObkcY9Z21lPjgur+yheVS6OUhjsIV0s5lGc5PGNqR/hOyhaVGjCVZ61hCNYEA6wNfokiA944bGogrSx0njYVd6fl2NAdX2sJCJxj1nbQyOyyh7WHGsZ8+e2LNnD1599VWR9DVqv6VrygYNGoTff//d3I9eQlRUFNLT00VHzkv/CGnsg3QNikatq/SsIRqDYHfs2IHffvsNO3bsEElfI7CUXnhHY1EF6QFzkydPRkREBCZNmiSSPiA/v67GYDnpWEBjsS5bnDWkyMHxt99+e9PXjz/+KJjN8kc68NP4oR4+fDgqV66MESNGiKSvUfstXVM2depUuLq64uOPPxZJH9CZMskQHpAnPRsGYB+109LnWiMoc3R0RKVKleDk5CSSvkaXBBcXF1SrVg0uLi4i6Uvfb4D8wEjpqT4B+fl1NcoM6Vjg6NGjcHNzKzAzmDVpdeO0piIHxx07drzpa+jQoZL5JCvTaB795JNPkJeXJxb4adR+SzebaSw/qlFTJl2jqBEMaCwfnZKSgqysLJH0AftYelm6RlHjfpAuXzVq8KXLJo3KDen5dTVatKTVr18fly5dQr169cS2oXGura3IwXFeXt4tX7m5uZJ5LVekLyaNwjU2Nhbe3t6IjY0VSV9rSUrJgEx68BEg34cTkK9R1BhUKD1oZM+ePTh//jz27Nkjkn4+yetVI8AHZPdBo/ZbunzVuKel56bV6IcqvQ/r1q1DVlYW1q1bJ5I+IN/dS6PmeNGiRZg1axYWLVoktg1rY5/jMkq64NBoWvzll19w+vRp7N69WyT9w4cPw8PDA0eOHBFJH+B5KCrpml2NQYXSLRHS3QUA+X3QuOekBwhptEJI39fSAy8B+W4VGv1QpcsNjbJVuuuGxup1GtertTE4LqOkCw6NpkXp5pqJEyfigQcewIQJE0TSB+SPk8YPhEazmXRNmXSXB0C+y4DGksLSgZ/0gD9AfjCbRquZdLmhUfZpDNSSnrFCutxo0qQJHB0d0bhxY5H0AfmuGxqtlxrlhrUxOC6jpDuwa9RY/vHHH6hZsyaOHz8ukr7GgA7p46QxUEH6PADyTdUa51q6n6j0ksKAfB95jcGdHh4eaNGiBTw8PETS1yj77OF+kC43NLpVSB8n6WsVkL/nNO4HjQHV1sbguJzSqLHUqN2QZotT0FxP4zxoNFVLs4fab+mHLY37QTrA19gHe7gfpMsNjfMgPXZHY55j6ePE+6FwDI7LKI3RndIXqi2uinM96XksNc6zxnnQaKqWJj3ISaO2T/p60pjXVTrA19gH6aBJa/S/5G+ERquZxoxGGgGf9Dak07fF34cSB8c5OTn4+++/ceLECYsXWYd0k5NGk5Z0waSx7KX0YAh7OA+Azuh5aRqDa6RJjwrXmNdVOvDT2gfJANweVvnTID2jkUbZag+xgMYMMdZW7OD4999/x0MPPQQ3NzfUqlULgYGBCAwMRO3atREYGCiRx3JJuqlDevUjQL4pRaPwll7BTqNJS2PKO+nRyBo1ZRoDF6VJB/ga87pKBxwaq1JKl00a5Yb0wEiNe1p6+kSNeb/tYXC+xnGytmIHxz169ECFChXw3XffYe/evdi3bx/27duHuLg47Nu3TyKP5ZYtr34EyDctaoymll7BTqNpUWN+3aFDhyIrKwuRkZEi6WvU0EjP96kRDEiPnteY11X6YU5jYKRGYCndNUR6lT/NLg8mk0kkfc77XTQai45ZW7GD4/379+M///kPIiIi0KRJEzRu3NjiRdYh3dShsTKbdOB38OBB5Obm4uDBgyLpAzrHSZrG/LqTJ0/GxYsXxeYT1RjQIT3fp0bzpXRAo9H1RHouZY2BkdKzGERFRWH9+vWIiooSSR+QD2g07ulmzZrB29sbISEhIulrVDxIlxsa50Fj4KK1FTs4vu+++5CSkiKRF/oX6aYUjbkNpWkEffZwnMaOHYs6depgzJgxYtvIyclBRkYGrly5IpK+RuEqfa41msKlAxqNeV2l50TVGBipsZy6NOl90LinpR9Srly5gqysLLFyD5AvN7QCV1uaqQIoQXD8wQcfICoqCj/++CPOnTuHtLQ0ixdZh3Stq0ZTivSAOY2gT/o4JSQkICwsDAkJCSLpAzpzTIaGhqJevXpifdiTk5Mxf/580cGX0vN9ajSFSy+zLV0zDchfrxoDeaXLDY2yT2OQra3PknDlyhVcvnwZubm5IukD9jF7S7lYPrpdu3bYtWsXwsLCUL16dVSuXBmVK1eGt7c3KleuLJHHckm6f6JGB3npUeEa/R+lj1Pv3r2RnJyMPn36iKQP6AzIa9++PVxcXBARESGSvsZxkh6YojGAVHq5XI0+lhorFUp3SZBuqtYo+6S70NjDLAmJiYlwcHBAUlKSSPqAfCwwdOhQrFy5EsOGDRNJH7DNmYCKHRxv3boVW7duxZYtWyxe+e+RdUgXHBod5KX79mnccNLH6ZNPPkFeXp7YgD9Avg8nID/A87PPPoOvry/mzp0rkj4gX3Os0X9deiYGjWtJo6+rNOkaS+nZYQD5Pvj2MEvCU089BQ8PD9H5e6VjgRMnTsAwDNGpeDW6Y1lbsYPj1q1b3/RF1iHdz0hjUm7pvn3ShTcg30wdFxeHPn36YP/+/SLpAzor5EmPzq9evTp69Ogh2k9U+sdao/+69EwM0v2BAfk+kPbQHUvjPPTo0QP9+/dH9+7dRdLX6LYh/aDl6emJOnXqwNPTUyR9QD4WmD17Nvz8/DBr1iyR9AGd7ljWVqJFQH766Sd07doVrVq1wqlTpwAAS5Yswfbt262aufJMup+Rxnrq0s1B0oU3IN9MrVF7ojEASXrgi8a0T9L3hMY9Jz0P8Y4dO9ChQwfs3LlTJH1AvuzT6IMv3a1C4zwAsrXsGrXftjhLwvWk7weNChpbVOzg+Msvv8Rjjz0GNzc37Nu3z9xckZ6ejokTJ1o9gyRDIyiTbg7SmCNYOtjQCJg0BiBJ/whp9JuWXhhHY7YK6cBPYx+kaVxL0veDxnmQHkSlUfsN2N4sCdeTrmTSiAXKxTzH48ePx+zZszF37lyLKbRatWrFRUBsiMZsFdI3ncaiCtI1NBoFk8ZAMOkHlf379+Py5cs4cOCASPqAzsI40j/UGnOWSu+D9H2t0W9aegYAjYoB6TEdGrXfGi1O0qQrmTS6t9hiDX6xg+Njx47h4YcfLvC+p6cnLly4YI08EeR/IDR+RKUDcI2CTzp41ag5lu4PrEFjaWfpAXMa16v0j5DGPkjXWGr0wdd4IJUmPYjKHgbkadSISh8nje4tGtPFWVuxg+MaNWrgjz/+KPD+9u3bRZfxLW+knxY1BuRJB+Aa09FJB/gaPxDS/YE1SC/tDMj3YddozpeuUdS456R/rCtWrIhWrVqJ1rrawwOp9CAqjYoBaRo1osuWLcOmTZuwbNkykfT5sFi4YgfHffv2xZAhQ/DLL7/AZDLh9OnTWLZsGd58800MGDBAIo/lknSfMo2CSbrg0Hhqlw7wNc6DxoOQNI2ZSaQDS63mfMkWJ417TvrHWqP22x4eSKXPtT2sGKnRveWrr77CuXPnxCrKNAZs2+LDYrGD46ioKHTs2BGPPPIIMjIy8PDDD6N3797o27cvBg0aJJFHs5iYGJhMJkRGRprfMwwDY8eOhb+/P9zc3NCmTRscPnzY4u+ys7Px+uuvo2rVqqhYsSKeeuop/P3336J5tQbJLg9aA2tsfb32li1bYt++fWjRooVI+hrnQaN/uTSNadA0JtvPysqyKL+szR5anKRrdjVqv6WPk8Z4C+l90Agspafi1PDEE0/A2dkZ7du3F0lf41oqN1O5TZgwASkpKfj111+xa9cunD17Fu+9956182Zh9+7dmDNnDho1amTx/qRJkzBlyhRMnz4du3fvhp+fHx599FGkp6ebvxMZGYnVq1djxYoV2L59OzIyMvDkk0+KLvl4u6RrNzT6AEn/UGvsQ79+/ZCYmIj+/fuLbUN6gJNGMCA9I4bGoBHp63XixIk4c+aM6IA/6Rk3NB607GEQlfRxmjNnDqKjozFnzhyR9AH5e04jKJMeZKsVWHp6eorVftvD/SahRMExALi7uyMkJATNmzdHpUqVrJmnAjIyMvDyyy9j7ty5FktUG4aBadOmYdSoUejUqROCgoKwaNEiXLx4EcuXLwcApKamYt68efjoo4/Qrl07NG3aFEuXLsXBgwexadMm0XzfDunmfOmlnQH5WlGNfQgICICDgwMCAgJE0tcomDSawqXPxY4dO/Dbb79hx44dIukD8terxkpU0sGA1mwYktuwh+5Y33zzDbKysvDtt9+KpA/Iz1YhPfASkB9kq7EEtnR3LI172hYVKTju1KlTkV8SBg4ciCeeeALt2rWzeD8hIQFJSUkIDw83v+fi4oLWrVubp4fZu3cvLl++bPEdf39/BAUF3XQKmezsbKSlpVm8NEl3GZCevxeQr9mVXp4auBpsNGrUCOPHjxdJX6NWV6NpUfpc5BfcJpNJJP3rtyNBYyWqu+++G6dOnRIbHK3RlUl6Gxr3g3SXhAULFqBevXqYP3++SPqA/GwVGg+80t2xNLrFSc8HrXE/2O08x15eXuaXp6cnNm/ejD179pg/37t3LzZv3gwvLy+rZ3DFihXYt29foTUhSUlJAABfX1+L9319fc2fJSUlwdnZ2aLG+frvFCYmJsZiv6VqDm/EHlaJkq5N1Bh1rrUSlSSN+Xulz0VoaCjq1asn1l0AkK/F11iJSnpWD41+otIP1dKrXgLy3SoCAwOxadMmBAYGiqQPyPcTdXR0RKVKlSzWS7A2jW5A0jWu0r/VGveDLQ4KL1JwvGDBAvPL19cXL7zwAhISEvDVV1/hq6++Qnx8PF588UVUrVrVqpk7efIkhgwZgqVLl8LV1fWG37u+NskwjFvWMN3qOyNGjEBqaqr5dfLkyeJlvozTaEqRrp3WaNLSmGpNmnTTIiB/LjQG5EnX4mvUMknP6qHRx1L6oVqjxUlj2j5p0rV9Y8eORZ06dTBmzBiR9AH53zmNbnHS+6DRimyL0/YVu8/x/Pnz8eabb8LBwcH8noODA4YNG2b1Jp69e/ciOTkZwcHBcHR0hKOjI2JjY/HJJ5/A0dHRXGN8fQ1wcnKy+TM/Pz/k5OTg/PnzN/xOYfI7wf/7ZU80mkela101gg3pm1qjuUl6/l7APhZLsYdpn6TPtcYDqXTwqjF1lcaKjtKkfyM0WuWk90GjkskezoMtVjIVOzi+cuUKjh49WuD9o0ePIi8vzyqZyhcWFoaDBw9i//795ldISAhefvll7N+/H3Xq1IGfn5/FD2ZOTg5iY2PNza/BwcFwcnKy+E5iYiIOHTok2kRb1mn8UEsHr/awDxoPKRqzekgHrxoPQrbY9Hc96XtC4zxId9HRqP3WmJdbmr1cS5L7oNFfV3ofNALXclFz/Oqrr6Jnz5748MMPsX37dmzfvh0ffvghevfujVdffdWqmfPw8EBQUJDFq2LFiqhSpQqCgoLMcx5PnDgRq1evxqFDh9CjRw+4u7ujS5cuAK72l+7VqxfeeOMNbN68GXFxcejatSsaNmxYYIBfWSJdgGv8QEgHZRr7AMj2KdMI8IcOHYqVK1di2LBhYtuQLmA1jpN0/8SEhASEhYUhISFBJH1A556Q7mMp3VStMUuCdDcgrbJPUnJyMubPny82/SMgf5w0xnNIk16BD5CfYlJCsYPjDz/8EMOHD8fUqVPx8MMP4+GHH8bUqVMRFRUl2qH7RqKiohAZGYkBAwYgJCQEp06dwoYNGyxWJpo6dSo6duyIF154AaGhoXB3d8eaNWssuoaUNdLNlxrNo9JLRmr099I4TtISEhJw6dIl/Pnnn2LbkK4Z0AgGpPsc9+7dG8nJyejTp49I+oD89apxz0mfB+nlqQH5+0GjXJKeu9we7geNld+kyz6NaQE1Bv1ZW7GD4woVKiAqKgqnTp3ChQsXcOHCBZw6dQpRUVEqweaPP/6IadOmmf9vMpkwduxYJCYmIisrC7GxsQgKCrL4G1dXV3z66ac4d+4cLl68iDVr1qjPPlFc0k1OGk9y0gPBNKZBkz5OGkHfc889h5o1a+LZZ58V24b0cbKHh5T3338f58+fF61lkq7B17jnpEkvTw3I3w8aTeHSlRuffPIJ8vLy8PHHH4ukD8gfJ41lwqVbOmbMmAFPT09Mnz5dJH1AZ9CftZV4ERAAdjlQrayQbkbWeJKTHhykMZhN+jhpBH19+/bFmDFj8Nprr4ltQ/o4aS13Lmn+/PkICQnBggULxLZhDwNIpbehMSBP+n7Q6MMpXbmhMbWhdFcpjTEj0i0dGufBFqdELVJw3LRpU9x///1FepFt0JjOCJDtn6gxgEr6iVcr6JPuJyp9nDT6HEvTqD2Rrimzh0VApLsLAPZRbkhXbmjUfmu0dEiXrdItHRrnwRYrN4oUHHfs2BFPP/00nn76aTz22GM4fvw4XFxc0KZNG7Rp0waurq44fvw4HnvsMen8khVJ39TSzUEaE7xrLJZi6+cBANatW4esrCysW7dObBvSsrOzkZKSgqysLJH0NY6RPSx6IL2NqKgorF+/HlFRUWLbkD7XGg+L0gPmNK5V6VYIjbIVkL0fNM4DoFNuWFORguMxY8aYX2fPnsXgwYPx888/Y8qUKZgyZQp27tyJyMhInDlzRjq/ZCXS/ckA+eYgjTkmpSfz1+hWoTEASWMpWGmHDx+Gh4cHjhw5IpJ+bGwsDhw4gG3btomkD8jfExrXq/SgP40fael7TmPmE+kBcxrlt3TrokbZKv0wp1G7botjRord53jlypXo1q1bgfe7du2KL7/80iqZInkaq6ZJNwdpNPFKB0waAyM1BiBpLAUrTfo4nT59Gs7Ozjh9+rRI+oD8vKsazaPSP9aTJ09GREQEJk2aJJI+IH8tacz0ID1gTqNbnHTfbI2yVfphTmMcQblYBMTNzQ3bt28v8P727dtvusQzlS0aq6ZpDHyx9f5eGgMjNc6DxlKw0qSP0+LFi9G0aVPRJljpeVc1mvM1ViqUXGQEkL+WNGZ6kB6opTGoUGMhKulrSfphTqOSqVwsAhIZGYn+/ftj0KBBWLp0KZYuXYpBgwZh4MCBGDp0qEQeSYA9DHDSmHNV+kdOa2CkNOnRyPaw6EFgYCA2bdqEwMBAsW1ozLsqTbpGUaPckKYxw4B0YKlRmyj9O6fRXUD6N0gjFigXNcfDhw/H4sWLERcXh8GDB2Pw4MGIi4vDwoULMXz4cIk8ko2SDmjsYc5VwPYGKhRG+odUI6CxhwDcxcUF1apVg4uLi0j6GjM9SNcyafR1laYRbEivcKpRmyh9T2t0M7KHcmnVqlU4cOAAVq1aVdpZKbISzXP8wgsvYMeOHfjnn3/wzz//YMeOHXjhhResnTcSpHHD2UMNjTSNkfMapGsfpAdGAvK1QBqBpTSNgbzSAYdGX1fp8lUjsJQumzQCS43fIFtfTl0jFvjll19w+vRp7N69W2wb1nZbi4CQ7dJoDpKu2dUYSCBNo/ZKo/CT3ob0wEhA/sdaI7CUvic0BvICtj91lXT5qlFzbOtT9uWnb+uzt0jvg8YDRPPmzVGzZk00a9ZMbBvWVqTg2MfHBykpKQCAypUrw8fH54Yvsg0asyRI/1DbQw2QxkA2e5h+a9SoUfD29sbIkSNF0gfka7/vvvtunDp1CnXq1BFJH5AfXPP888+jYcOGeO6550TSB3SmcpPuViEdvGoE+NJlk0a51LJlS+zbtw8tWrQQSV+j9lv6d06je6LGBADW5liUL02dOtW8dvi0adMk80NK/j1LwsyZM0W28dJLL8HNzU2s4Ph30+LLL78sso0VK1Zgy5YtyMnJQc+ePa2e/r8HskkN1HrmmWfw9ddfixbg0gGHxnHKzMzE6tWr8cwzz4gEyH/88Qdq1qyJ48ePWz3tfPkBvhSNe076WpIulwDL4FWi3NAI8KXvOY1yafDgwahQoQKGDBmCTZs2iWxDuvZb+p7TaIGVLpdEGFQkqampBgAjNTW1tLNiFWfOnDH69+9vnDlzprSzUmIZGRnG0qVLjYyMDLFtzJ071+jatavx2WefiaSvsQ8ZGRnGkiVLxLchuR/x8fFG27Ztjfj4eJH0DcMwlixZYkydOtVYunSpSPq854rmzJkzRr9+/Wz6ONlDuaFxz0mLj483wsLCxPbhs88+M7p06WLMmzdPJH3DkD/XGteSNmvEa0Xuc5yWllakF9kGjblvNfq6GsJP7dLN1BrT6GgNjJQ8FxrzQUs3hfOeKxqt5WwlacwdK30epOfM1iA9faLGQGHp3wh7mNZVQpGDY29vb1SuXPmGr/zPifJJrzuv0WfNHgoOjQJcOgDXmA9aegYAjdkqpO8JjXtO43rVIBm8apwHjT7ytk5joLA9zHJji4ocHG/duhVbtmzBli1bsHnzZri4uGDJkiXm9/I/J8onPX2LxmAIe7B//35cvnwZBw4cENuGIdwHUqPWVbrmeMyYMUhMTBQN8KX3QeOe07hepUkHrxrn4ejRo3Bzc8OxY8fEtmHrNAYKa5QbVFCRg+PWrVubX23atIGDgwNatGhh8X7r1q0l80o2pkmTJnB0dETjxo1F0reHWl2NZnDp8wDINyPbw+ITGrXf0l0SNO456etV457TWLaY56H0bdmyBX5+fti6davYNqKiopCeno633npLJH2NstUWzzXnOS6jbPFiKozJZLLplaikz4NGf2DpVdMA+R9rjTmCpWtdK1asiFatWokGNNI1+Bo/pNLXq0aXBOnV5TTYw3lISEhAWFgYEhISRNKXvt8A+QBco2y1xQXBGByXUdIFh0bwbQ/No9I3tb0sgS19PTVo0ABpaWm47777RNIH5GuONX4gpGvw7WFFR40uCfZwnKSn+NKYa793795ITk5Gnz59RNIPCwvD6dOn0bZtW5H0AfnfCI2y1RZ/524rODaZTNbKB11HugDXeGqvX78+Ll26hHr16oltQ5pGzYA90LiepFshpGuOta4lyfTt5T6w5XOgRTrw05gN47PPPoOvry/mzp0rkr49zOjh4eGBFi1amNeykGCLq9kWOTju1KmTxSsrKwv9+vUr8D5Zh3QztUbtycGDB5Gbm4uDBw+KbUOaxpRM9kD6etJohZCuOdaoZZKunZ48eTIiIiIwadIkkfQ1aNTga6x8KS06Ohrx8fGIjo4WSV9jKXLpqdw0al2laZRLGqvZWluRg2MvLy+LV9euXeHv71/gfbId0rUbV65cQVZWFq5cuSK6HWm2XguUnZ2NlJQUZGVliW5H8jhpDCqUDvClgw1AvvlSY9YQaWfPnsXmzZtF+03/e3U5CRrd4nJycpCRkSFWfrdv3x4uLi6IiIgQSV+DxngOaRrlki3OXV6k5aMBYMGCBZL5IGX/bgaXWtbR1dUVVatWhZubm0j6GhYtWoRly5YhLS0NAwYMsHr6Gs1NGnNxSl9PGk1/gGyALx1sALbZfHk96X34/vvvkZOTg3Xr1mH48OEi25BeGll6WXvgap/gEydOiPUJ/neXhDlz5ohsQ3pJeI37TXobGuWSLXZP5IC8ckqjW4U9NC3u27cPWVlZiIuLE0lfo9uGxlyc0teTxnGS7jfdqFEjODg4ICgoSCR9wD66AUk3wT7xxBNwdnZG+/btRdIH5LvFaQQbzz//PBo2bIjnnntOJH2NbhXS97TG/SZ9PzRr1gze3t4ICQkRSR+wzXKJwXE5pTFPpj00LWoEltJP09LnIZ/kfmhMjSUd4Gv0wZe+rzXuOekm2KpVq+LRRx9F1apVRdIH5I+TRh9O6T74bdq0wfnz50XXR7CH+aal7wetriG2VGsMlPHgOCYmBs2aNYOHhweqV6+Ojh07FlitxzAMjB07Fv7+/nBzc0ObNm1w+PBhi+9kZ2fj9ddfR9WqVVGxYkU89dRT+PvvvzV3pVyyhxk3pANLjcFBGq0E0udCY5Uo6R86R0dHVKpUCU5OTiLpA/LzEGvcc9K1ovbQCqHRh1N69pbBgwejQoUKGDJkiEj6gH0sFCV9P2h0DVm0aBFmzZqFRYsWiW3D2sp0cBwbG4uBAwdi165d2LhxI65cuYLw8HCLp/FJkyZhypQpmD59Onbv3g0/Pz88+uijSE9PN38nMjISq1evxooVK7B9+3ZkZGTgySefRG5ubmnsVpHYyyIgkk+LGkEfp/cqGulzIb1KlAaNbkbSE/przE2rEbza+v2gMW+sdM3xJ598gry8PHz88cci6WvQ+J2Wvh80WiF++eUXnD59Grt37xbbhrWV6eB4/fr16NGjBxo0aIDGjRtjwYIFOHHiBPbu3QvgagE3bdo0jBo1Cp06dUJQUBAWLVqEixcvYvny5QCA1NRUzJs3Dx999BHatWuHpk2bYunSpTh48CA2bdpUmrt3U9I1DxorXWnVMkmSrqHRCAS0ntpteZUoDRr7ID211KRJk+Dh4YHJkyeLpA9cLZvmz59v07XfgO01I19POsCPjY2Ft7c3YmNjRdLXoNHyJ92lTPohCNCZbcjaynRwfL3U1FQAgI+PD4CrS0MmJSUhPDzc/B0XFxe0bt3a3Ay+d+9eXL582eI7/v7+CAoKumlTeXZ2NtLS0ixemqQLJo1maulaV3to4tVo9tN4apc+Fxo17NK1QBq1fdKzekRHR8Pf31+03JBe1Uy6XALk7wetWUkk7zfpwc6A/D2tUS5JtwZptMBqzTZkTTYTHBuGgWHDhuHBBx80j/ZOSkoCAPj6+lp819fX1/xZUlISnJ2dUbly5Rt+pzAxMTEW8zcHBARYc3duSTpo0mimln4i1bipbXHy8us1b94cNWvWRLNmzcS2IR1waNSwa9QCSZM+ThUrVkSrVq1EH+akm9s1asqkyyaNckk6wJ84cSIeeOABTJgwQSR9QP6e1jgPGrN6SLdycLYKQYMGDcKBAwfw+eefF/js+mWsDcO45dLWt/rOiBEjkJqaan6dPHmyZBkvozRmMNCooZG+qaV/SDX6rPXo0QP9+/dH9+7dxbahcZykZ6uwh4Ev0g/VGg8Qu3btQpMmTfDLL7+IpK/xUA3Ilk32EOADOst4S97TGudBuvy2h253EmwiOH799dfx7bffYuvWrbjjjjvM7/v5+QFAgRrg5ORkc22yn58fcnJycP78+Rt+pzAuLi7w9PS0eNkTjYJPur+uRrcKe5hxA7D9AUga3YDsYeCLNI1mZK1tSJK+rzXKb+kHLenuAoD8PW0P58Eeut1JKNPBsWEYGDRoEL766its2bKlwProgYGB8PPzs3hqy8nJQWxsrHk0dXBwMJycnCy+k5iYiEOHDomOuLYFtv7UrlEzDdj+jBsaBZN0za5GX1fpHyGN6bfsYX5d6W1o3A/20GomTaO7gHS5ZA9TxWkMltP6rbamMh0cDxw4EEuXLsXy5cvh4eGBpKQkJCUl4dKlSwCudqeIjIzExIkTsXr1ahw6dAg9evSAu7s7unTpAgDw8vJCr1698MYbb2Dz5s2Ii4tD165d0bBhQ7Rr1640d69UafxASD+1r1q1CgcOHMCqVatE0gfkj5NG4aoRgEdFRWH9+vWIiooSSb969eqYOXMmqlevLpI+YB+Dd6SbSDWake1hrIJ02WQP/eM1untJl0sasz5J0xgsp1ExYG1lOjieNWsWUlNT0aZNG9SoUcP8+neBEBUVhcjISAwYMAAhISE4deoUNmzYYHGip06dio4dO+KFF15AaGgo3N3dsWbNGjg4OJTGbpUJ9tDvTmO0s/Rx0uhzrBGAS9diaRwne1hqVrqJVKPcsIe+rtJlkz3M3mIP5ZJG1xBpGuWSxkw91lamg2PDMAp99ejRw/wdk8mEsWPHIjExEVlZWYiNjTXPZpHP1dUVn376Kc6dO4eLFy9izZo16rNPlEXSBYd07YbGaGdA9jhpzZMpHVhOnjwZERERmDRpkkj6Gi0d0gtcaAQDGk2k0uWGPQwqlF523h5mq9AQFRUFd3d3sVmZtLqGSJff0ve01tSD1lSmg2OSo/EDIV27odHULn2cNGqANH7kpKf40qhNlF7gQuNHTrqJ1B4e5jTuOekFX+xhtgqN+2H69OmoW7cuZsyYIZK+RtcQ6XtO4/fBFgcjMzgup9LS0rB7927zwioSpJtrNPp7STcHaRQaGoMhtB62JEnP/a1xjKTvuVOnTuG7777DqVOnRNIH5H+sNe456XJD456WHsymEZRpzOcvXS5JX0saS8JrPMxZG4Pjcurw4cPw8PDAkSNHxLYh3TyqMb2XdHOQPQxwAuQLcI0fUum5vzVqLKXvuWXLluHy5ctYtmyZSPqAfI2lxv0gXW5oDHCSLl81WoOk72mNckn6WtJYEl5rjJNVGVQkqampBgAjNTW1tLNiFWfOnDH69+9vnDlzprSzUmIa+5CRkWEsXbrUyMjIEEn/zJkzRr9+/Wx6HwzDMObOnWt07drV+Oyzz0TS19gH6W1o7IO06Ohow9PT0xg3blxpZ6XE7OFakr7fDMN+fiMky1d7uJbs4TxfzxrxmskwbHyyRCVpaWnw8vJCamqq3S0IQjeWmZmJ1atX45lnnhGpjVu6dClSUlJQrVo1vPzyy1ZPX0tmZia+/vprdOzY0Wbn/JQ+1/bgs88+Q2xsLNq0aYNevXqVdnZKxB7Osz3cbxrmzZuHLVu2ICwsDD179izt7JASa8Rr7FZBdBPS/URtsrmpENLN+RqDd+xh7lhpGtM+SdNaFMfWp0HTuOekGcJdmezhGFHhGBwT3YR04WoPKyxp0AhopM+1BnsIyqTZy6qU0uxhH+xhtUV7CMBtcR8YHJdTGjM9SN8QGjecdOGqsQ8a25C+njQCGulzrXHPSdd+28O1pBHga8wAIE16RgyN+0F68KXGrCHSq17awwJLEhgcl1PSy2oC8jeERjP4smXLsGnTJrHR+Rr7oLENjZWipGt0pZf81ThG0lM02su1JO2dd97Bf//7X7z77rsi6WsENNIzYmj8Bkk/VEuXGcDVGTd+++037NixQyR9e1hgSQKD43Lq0qVLSElJwaVLl8S2IV0wnT17Fps3bxatefjmm2+QlZWFb7/9ViR9jWU1NbbRoEEDpKWl4b777hNJX7r2BJBf8lf6GAHA/v37cfnyZRw4cEAkfY1r6e6778apU6dQp04dkfQ1AssTJ07AMAycOHFCJH2N+0G6m5HGbxAg+1AtXWYAwJUrV5CVlYUrV66IpK/RKif9sCiBwXE55eDgAFdXVzg5OYltQ7r58vvvv0dOTg7WrVsnkj4ALFiwAPXq1cP8+fNF0rfFZTUL4+LigmrVqsHFxUUk/V9++QWnT5/G7t27RdIHdJYjN5lMoj/WzZs3R82aNdGsWTOR9DWu16NHj8LNzQ3Hjh0TSV+jpmzKlClwc3PDRx99JJK+RlAm3c3I1dUVVatWhZubm0j6gPy5Hjp0KLKyshAZGSmSPiB/nKQXewGAhIQEXLp0CX/++afYNqyNwXE51axZM3h7eyMkJERsGxr9UKtWrSr6xFu9enX06NFDbIlqW1xWszQ0adIEjo6OaNy4sdg2pJfAlq7VBYDnn38eDRs2xHPPPSeSvsZsFfXr18elS5dQr149kfQ1+onOnz8fISEhWLBggUj6GkGZdLeKyZMnIyIiApMmTRJJH5A/15MnT8bFixdFF9CQPk4a3Zgef/xxODo6Ijw8XGwb1sbguJzy8PBAixYt4OHhIbYN6RWWunbtirCwMNH5gaVrHuxhtS4NGter9LmWrtUF5AMajcFsf/zxB2rWrInjx4+LpK9xz0kvWzx9+nTUrVsXM2bMEEkfkO9WUb16dcycOVOs4gGQvx80ZraRPk5NmzaFm5sbmjRpIpI+ABw/fhw+Pj5ISEgQ24a1MTguo6T7xWnUAEn/QGj8yEnXPNjDLAyAfAAeFhaG06dPo23btiLpA/LnQrpWF7CPeV2lu7doDA6SXrZYumwF5H8jNGarkL4fxo4dizp16mDMmDEi6WvQKJc0KgasjcFxGSVdi6VRA7Rlyxb4+flh69atIulrBJbSNQ+AfO2DxkOEdAA+YcIEXLhwATExMSLp55M8F/ZwHjT660rXlEVHRyM+Ph7R0dEi6QPyD9XSwXc+yftBumURkL8fpH/jNGiUSz169ED//v3RvXt3sW1YG4PjMko68NOoAZIe2a4R4Evvg8bUWBp9LKUfIjSa/qQDP40aS3uY11W6bMrJyUFGRobY6H/APs6DdNkUHR0Nf39/0eBYeqo1jYWDpO8HjWvJFhcPYnBcRklfTBo1QPbQ11WaRuGqUTMgvR8aNQ/SweukSZPg4eEhOnhH+qFa41qSLpsaNWoEBwcHBAUFiaQP2Md5sIc+x9JzBGt0T7SHcS+2iMFxOaXxtKhRcNg6jWNkD+dao+ZBuuuGRk0ZINsUrtGVSXob0lPF5ZM8D/ZwT2vIPwcmk0kkfY1yyR7GvdgiBsfllEZfWltsStGmcYy0agY0Rm5Lku66IT1VHKDTIqRxnm15LmhAfpEOjXtaumzS6NoXGhqKevXq2dTKbNezhxlobBGD43JKY3U5aQkJCQgLC7Op6WGupzFiW6Ovq3QwoPFD2qZNG5w/fx6tW7cWSV8jcG3atCnmzp0rFuBr7MOnn36KN998E9OnTxdJX2N0vnRzvj30/dZY5a99+/ZwcXFBRESESPoa5bf0kvAaZavGcbI2BsdllPQFq7G6nLRu3bohLi7OpkbAXk9jAnaNmR6kgwGNgYuDBw9GhQoVMGTIEJH0NQKagQMHIjU1Fa+//rpI+hoPWgsWLEBaWprYAhqaM9BINedr7IP0Paex6qV02adRfu/Zswfnz5/Hnj17RNLXKFs1ZiaxNgbHZZR0DY3G6nLSAX5AQAAcHBwQEBAgkj4gP6hQYxaGBg0aIC0tDffdd5/YNhwdHVGpUiWx5cg1Bi5+8sknyMvLw8cffyySvkZA89RTT8Hd3V1s6iqNQYU9evSAp6cnevToIZK+9Aw0gHxzvsb9IL0NjVUvpctXjbJV+kFL41rSGm9hTQyOyyjpTvIaq8tJP5GOGjUK1atXx8iRI0XSB+QHpWg08bq4uKBatWpwcXER20ZUVBTc3d3FFiXQGBy0a9cuNGnSBL/88otI+hpBmfR9rbH4RK9evfDMM8+gZ8+eYtuQJn1fayyKI33Paax6Kd2tQqNsbdasGby9vRESEiKSvsYiUZmZmTh27Jho1w1rY3BcRkl3ktcY0CEdDEydOhWurq5iNX2AfZwHDVrL2UrSqEGRJn09aSw+oVHDLk36PGjU4EuXfRoBvvRx0piuVPohQnouaADo3bs3kpOT0adPH7FtWBuD43LKHqZv0WjSkqbRh1OjAJduNtPoFyddg6JxHqTva+kBf4D8Q4o9nAeNGnzpbnEaYyGkj5NGi5b0NjT6fkt3WZNQroLjmTNnIjAwEK6urggODsZPP/1U2lkqVdI1ZBo/QiaTyaZr+jRqgDSazaSnKdPokiBd26fxQ5qZmYnt27eLBTTSgxYB+etV4zxI17quW7cOWVlZogOqpce9aFRuaLR0cGrDW5Pusiah3ATHX3zxBSIjIzFq1CjExcXhoYceQkREBE6cOFHaWSsVGlMySf8IHT58GB4eHjhy5IhI+ho0aoA0mqmlryeNBy3p2SQ05hOVHj3/2WefwdfXF3PnzhVJH9Bb/c2W7du3D1lZWYiLixPbhvT9oNFfV7oGX6NFS7ps1Rj3olG5YW3lJjieMmUKevXqhd69e6N+/fqYNm0aAgICMGvWrNLOWqnQaM6XDgZGjRoFb29v0QF50k2LW7ZsgZ+fH7Zu3SqSPqDTl1b6etKo7ZN+iNCYT1S6Ni4wMBCbNm1CYGCgSPqAfFCmEdBIGzp0KLKyshAZGSm2DXsZDyFZ7mkEfdIBvkafY43KDWsrF8FxTk4O9u7di/DwcIv3w8PDb9jckp2djbS0NIuXPYmOjkZ8fDyio6NLOyslptG0KP1DKj3BO6DTrUKj/6B0bZ/0udBorQFkuxppBPjSP9a2WIt1vcmTJ+PixYui3bGkg7Ls7GykpKQgKytLJH1AZ6ERDZJln0afY43fIGsrF8FxSkoKcnNz4evra/G+r68vkpKSCv2bmJgYeHl5mV+Sc+mWBltvVgR0mhala101uoZo1ABJzyeqEVhKnwuNQbDS+6BxHqTva1usxbqeRvkt3fKnUfbZw7Ukfc9p9Dm2yVYIoxw4deqUAcDYuXOnxfvjx4836tWrV+jfZGVlGampqebXyZMnDQBGamqqRpbFnTlzxujfv79x5syZ0s5KiWnsQ0ZGhrF06VIjIyNDJH172AeNbWjsA++JW7OH86CxD9J4rZaNbbBsLTvb+LfU1NTbjtdMhmEHVYi3kJOTA3d3d6xcuRLPPPOM+f0hQ4Zg//79iI2NvWUaaWlp8PLyQmpqKjw9PSWzS0REREQlYI14rVx0q3B2dkZwcHCBKv2NGzeKDkgjIiIiItviWNoZ0DJs2DC88sorCAkJQcuWLTFnzhycOHEC/fr1K+2sEREREVEZUW6C486dO+PcuXMYN24cEhMTERQUhLVr16JWrVqlnTUiIiIiKiPKRZ9ja2CfYyIiIqKyjX2OiYiIiIisqNx0q7hd+RXs9rYYCBEREZG9yI/TbqdjBIPjIkpPTwcAu1sMhIiIiMjepKenw8vLq0R/yz7HRZSXl4fTp0/Dw8MD6enpCAgIwMmTJ9n/2M6lpaXxXJcTPNflA89z+cFzXT5cf54Nw0B6ejr8/f1RoULJeg+z5riIKlSogDvuuAMAYDKZAACenp684coJnuvyg+e6fOB5Lj94rsuHf5/nktYY5+OAPCIiIiKiaxgcExERERFdw+C4BFxcXDBmzBi4uLiUdlZIGM91+cFzXT7wPJcfPNflg8R55oA8IiIiIqJrWHNMRERERHQNg2MiIiIiomsYHBMRERERXcPgmIiIiIjoGgbHJTBz5kwEBgbC1dUVwcHB+Omnn0o7S2RlY8eOhclksnj5+fmVdrboNm3btg0dOnSAv78/TCYTvv76a4vPDcPA2LFj4e/vDzc3N7Rp0waHDx8unczSbbnVue7Ro0eBe7xFixalk1kqsZiYGDRr1gweHh6oXr06OnbsiGPHjll8h/e17SvKebbmPc3guJi++OILREZGYtSoUYiLi8NDDz2EiIgInDhxorSzRlbWoEEDJCYmml8HDx4s7SzRbcrMzETjxo0xffr0Qj+fNGkSpkyZgunTp2P37t3w8/PDo48+ivT0dOWc0u261bkGgMcff9ziHl+7dq1iDskaYmNjMXDgQOzatQsbN27ElStXEB4ejszMTPN3eF/bvqKcZ8CK97RBxdK8eXOjX79+Fu/de++9xvDhw0spRyRhzJgxRuPGjUs7GyQIgLF69Wrz//Py8gw/Pz/j/fffN7+XlZVleHl5GbNnzy6FHJK1XH+uDcMwunfvbjz99NOlkh+Sk5ycbAAwYmNjDcPgfW2vrj/PhmHde5o1x8WQk5ODvXv3Ijw83OL98PBw7Ny5s5RyRVJ+//13+Pv7IzAwEC+++CLi4+NLO0skKCEhAUlJSRb3t4uLC1q3bs372079+OOPqF69Ou655x706dMHycnJpZ0luk2pqakAAB8fHwC8r+3V9ec5n7XuaQbHxZCSkoLc3Fz4+vpavO/r64ukpKRSyhVJeOCBB7B48WL88MMPmDt3LpKSktCqVSucO3eutLNGQvLvYd7f5UNERASWLVuGLVu24KOPPsLu3bvRtm1bZGdnl3bWqIQMw8CwYcPw4IMPIigoCADva3tU2HkGrHtPO1ozw+WFyWSy+L9hGAXeI9sWERFh/nfDhg3RsmVL3HXXXVi0aBGGDRtWijkjaby/y4fOnTub/x0UFISQkBDUqlUL33//PTp16lSKOaOSGjRoEA4cOIDt27cX+Iz3tf240Xm25j3NmuNiqFq1KhwcHAo8bSYnJxd4KiX7UrFiRTRs2BC///57aWeFhOTPRsL7u3yqUaMGatWqxXvcRr3++uv49ttvsXXrVtxxxx3m93lf25cbnefC3M49zeC4GJydnREcHIyNGzdavL9x40a0atWqlHJFGrKzs3H06FHUqFGjtLNCQgIDA+Hn52dxf+fk5CA2Npb3dzlw7tw5nDx5kve4jTEMA4MGDcJXX32FLVu2IDAw0OJz3tf24VbnuTC3c0+zW0UxDRs2DK+88gpCQkLQsmVLzJkzBydOnEC/fv1KO2tkRW+++SY6dOiAO++8E8nJyRg/fjzS0tLQvXv30s4a3YaMjAz88ccf5v8nJCRg//798PHxwZ133onIyEhMnDgRdevWRd26dTFx4kS4u7ujS5cupZhrKombnWsfHx+MHTsWzz77LGrUqIE///wTI0eORNWqVfHMM8+UYq6puAYOHIjly5fjm2++gYeHh7mG2MvLC25ubjCZTLyv7cCtznNGRoZ172mrzHlRzsyYMcOoVauW4ezsbNx///0WU4mQfejcubNRo0YNw8nJyfD39zc6depkHD58uLSzRbdp69atBoACr+7duxuGcXXapzFjxhh+fn6Gi4uL8fDDDxsHDx4s3UxTidzsXF+8eNEIDw83qlWrZjg5ORl33nmn0b17d+PEiROlnW0qpsLOMQBjwYIF5u/wvrZ9tzrP1r6nTdc2SkRERERU7rHPMRERERHRNQyOiYiIiIiuua3geN++fTh48KD5/9988w06duyIkSNHIicn57YzR0RERESk6baC4759++K3334DAMTHx+PFF1+Eu7s7Vq5ciaioKKtkkIiIiIhIy20Fx7/99huaNGkCAFi5ciUefvhhLF++HAsXLsSXX35pjfwREREREam5reDYMAzk5eUBADZt2oT27dsDAAICApCSknL7uSMiIiIiUnRbwXFISAjGjx+PJUuWIDY2Fk888QSAq5Otc1lGIiIiIrI1txUcT5s2Dfv27cOgQYMwatQo3H333QCAVatWcVlGIiIiIrI5IouAZGVlwcHBAU5OTtZOmoiIiIhIjOPtJnDhwgWsWrUKx48fx1tvvQUfHx8cOXIEvr6+qFmzpjXyWCbk5eXh9OnT8PDwgMlkKu3sEBEREdF1DMNAeno6/P39UaFCyTpI3FZwfODAAYSFhcHb2xt//vkn+vTpAx8fH6xevRp//fUXFi9efDvJlymnT59GQEBAaWeDiIiIiG7h5MmTuOOOO0r0t7cVHA8bNgyvvvoqJk2aBA8PD/P7ERER6NKly+0kXebk79/Jkyfh6elZyrkhIiIiouulpaUhICDAIi4trtsKjnfv3o3//Oc/Bd6vWbMmkpKSbifpMie/K4WnpyeDYyIiIqIy7Ha6wN7WbBWurq5IS0sr8P6xY8dQrVq120maiIiIiEjdbQXHTz/9NMaNG4fLly8DuBqlnzhxAsOHD8ezzz5rlQwSEREREWm5reD4ww8/xNmzZ1G9enVcunQJrVu3xt133w0PDw9MmDDBWnkkIiIiIlJxW32OPT09sX37dmzZsgX79u1DXl4e7r//frRr1w4C0ycTEREREYm6rUVAYmJiMGLEiALv5+bmomvXrvj8889vK3NlSVpaGry8vJCamsoBeURERERlkDXitdtePnrOnDkW7+Xm5uLFF1/E/v37bydpIiIiIiJ1t9WtYu3atWjXrh28vb3xwgsv4PLly+jcuTP+97//YevWrdbKIxERERGRitsKjoODg7F69Wo8/fTTcHFxwbx583D8+HFs3boVvr6+1sojEREREZGK2+pWAQBt2rTBkiVL8Nxzz+HPP/9EbGwsA2MiIiIisknFrjnu1KlToe9Xq1YN3t7eeO2118zvffXVVyXPGRERERGRsmIHx15eXoW+/9hjj912ZoiIiIiISlOxg+MFCxZI5KNQJpMJq1evRseOHYv0/R9//BGPPPIIzp8/D29vb9G8EREREZH9ue0+x5ISExMRERFh1TTHjh2LJk2aWDVNIiIiIrIPxQ6O77//fpw/fx4A0LRpU9x///03fN2OnJwc+Pn5wcXF5bbSsUWZmZkwmUwwmUzIzMy06e1o7YsGezovWtuxp/MP2Ncx476UTfZ0zOzpvAD2dczsbTvWVuxuFfnTtgEocneHomjTpg2CgoLg7OyMxYsXo0GDBti2bZtFt4qdO3diwIAB+N///oegoCC88847eOaZZxAXF2dRG7x37168/fbbOHLkCJo0aYIFCxagXr16WLhwIaKjowFc7bIBXO0m0qNHD6vtBxERERHZrmIHx2PGjCn039awaNEi9O/fHzt27IBhGKhfv775s/T0dHTo0AHt27fH8uXL8ddffyEyMrLQdEaNGoWPPvoI1apVQ79+/dCzZ0/s2LEDnTt3xqFDh7B+/Xps2rQJwI0HGBIRERFR+XNbi4Dk27NnD44ePQqTyYT69esjODi4ROncfffdmDRpUqGfLVu2DCaTCXPnzoWrqyvuu+8+nDp1Cn369Cnw3QkTJqB169YAgOHDh+OJJ55AVlYW3NzcUKlSJTg6OsLPz++mecnOzkZ2drb5/2lpaSXaJyIiIiKyHbcVHP/999946aWXsGPHDvPsEBcuXECrVq3w+eefIyAgoFjphYSE3PCzY8eOoVGjRnB1dTW/17x580K/26hRI/O/a9SoAQBITk7GnXfeWeS8xMTEmLtgEBEREVH5cFuzVfTs2ROXL1/G0aP/r717j635/uM4/jpoj6r2uJXTlmrdasGQMHeji2olBJ0REq2ISYZZTLaYucRclo6Rn2YRizQqLtsyyrKFids2YlNzC41glXW0Gqo9ynZo+/39oT1Rrdv5np7Tnj4fyUnO9/s559NX+8nn5J1vP+fzzVZhYaEKCwuVnZ0twzA0c+bMV+4vODj4mW2GYbjWCT95riYBAQGu55XvKS8vf6UsixYtUnFxseuRm5v7Su8HAABA/WPqyvGvv/6qEydOKDY21nUuNjZWGzdu1JAhQ0yHe1L37t21fft2OZ1O1xcCs7KyXrmfwMBAlZWVvfB1Vqu1Qe6UAQAA0JCZunIcFRWlR48eVTtfWlqqyMhIM11XM3XqVJWXl+vdd99Vdna2Dhw4oLVr10pStSvKzxMdHa2cnBydPXtWt2/frrKuGAAAAA2bqeI4NTVV8+bNU1ZWlmuJQ1ZWlubPn+8qXD0lNDRUP/zwg86ePas+ffpo8eLFWrp0qSRVWYf8IklJSUpISNDIkSMVFhamnTt3ejSnJwQHB8swDBmG8dylJvXh53jrd/EGfxoXb/0cfxp/yb/+ZvwudZM//c38aVwk//qb+dvP8TSL8ayFuy+hZcuWevDggUpLS9WkyeMVGpXPn/4jFBYWmktag+3bt2vGjBkqLi5WUFCQx/t/ksPhkM1mU3FxsUJDQ2v1ZwEAAODVeaJeM7XmeMOGDWbe/soyMjLUqVMnRUZG6ty5c/r444/1zjvv1HphDAAAgIbBVHGcnJzsqRwvJT8/X0uXLlV+fr7Cw8M1adIkrVq1yqsZAAAA4L/cWlZRXl6u8vJy11IKSbp165Y2bdqk+/fva9y4cRo6dKhHg/oayyoAAADqNp8tq5g5c6YCAgK0efNmSY9v7dy/f3/9999/Cg8P1/r167V3716NGTPGrVAAAACAL7i1W8Xx48f19ttvu44zMjJUWlqqK1eu6Ny5c1qwYIG++OILj4UEAAAAvMGt4vjGjRvq2rWr6/jQoUNKSkqSzWaT9Hgt8sWLFz2TEAAAAPASt4rjpk2b6t9//3Udnzx5UgMHDqzSXlJSYj4dAAAA4EVuFce9e/fWtm3bJD2+hfStW7cUFxfnar927ZoiIiI8kxAAAADwEre+kLdkyRKNGTNG3377rfLy8pSSkqLw8HBX+549ezRkyBCPhQQAAAC8wa3ieOTIkTp9+rQOHjwou92uSZMmVWnv06eP3njjDY8EBAAAALzF1O2jGxL2OQYAAKjbPFGvubXmGAAAAPBHFMcAAABABYpjAAAAoALFMQAAAFDBVHHcqVMn3blzp9r5oqIiderUyUzXAAAAgNeZKo6vX7+usrKyauedTqdu3LhhpmsAAADA69za53jfvn2u5wcOHJDNZnMdl5WV6dChQ4qOjjYdDgAAAPAmt4rj8ePHS5IsFouSk5OrtAUEBCg6Olrr1q0zHQ4AAADwJreK4/LycklSTEyMTp06pTZt2ng0FAAAAOALbhXHlXJycjyVAwAAAPA5U8XxihUrntu+dOlSM90DAAAAXmWqON6zZ0+V40ePHiknJ0dNmjRR586dKY4BAABQr5gqjs+cOVPtnMPhUEpKiiZMmGCmawAAAMDrPH6HvNDQUK1YsUJLlizxdNcAAABAraqV20cXFRWpuLi4NroGAAAAao2pZRX/+9//qhwbhqG8vDxt27ZNCQkJpoIBAAAA3maqOF6/fn2V40aNGiksLEzJyclatGiRqWAAAACAt9W7fY5HjBihPn36aMOGDS/1+szMTC1cuFA5OTmaN2/eS78PAAAADY+p4vhJubm5slgsat++vae69IjZs2drxowZev/99xUSEqKUlBQVFRUpMzPT19EAAABQx5j6Ql5paamWLFkim82m6OhodezYUTabTZ9++qkePXrkqYxuKykpUUFBgUaPHq2IiAiFhIT4OhIAAADqMFPF8dy5c7V582alpqbqzJkzOnPmjFJTU7VlyxbNmzfPUxmf6eHDh/roo48UGRmp4OBgDRgwQEePHpUkHT161FUMx8XFyWKxaMSIEdq6dav27t0ri8Uii8Xiej0AAABgalnFzp07tWvXLiUmJrrOvf7664qKitKUKVO0adMm0wGfZ8aMGbp+/bp27dqliIgI7dmzRwkJCbpw4YIGDx6sy5cvKzY2Vt9//70GDx6sZs2aadasWXI4HEpPT5cktWrVqsa+nU6nnE6n69jhcNTq7wIAAADfM3XluGnTpoqOjq52Pjo6WoGBgWa6fqFr165p586d+u677zRs2DB17txZCxcu1NChQ5Wenq7AwEC1bdtW0uMC2G63KzQ0VEFBQbJarbLb7bLb7c/MuWbNGtlsNtejQ4cOtfr7AAAAwPdMFcdz5szRZ599VuUKq9Pp1KpVqzR37lzT4Z7nzz//lGEY6tatm5o3b+56HDt2TNeuXTPd/6JFi1RcXOx65ObmeiA1AAAA6jJTyyrOnDmjQ4cOqX379urdu7ck6dy5c3r48KHeeustTZw40fXa3bt3m0v6lPLycjVu3FinT59W48aNq7Q1b97cdP9Wq1VWq9V0PwAAAKg/TBXHLVq0UFJSUpVz3lp+0LdvX5WVlamgoEDDhg176fcFBgaqrKysFpMBAACgvjJVHFd+qc0XunXrpmnTpmn69Olat26d+vbtq9u3b+vw4cPq1auXxowZU+P7oqOjdeDAAV2+fFmtW7eWzWZTQECAl9MDAACgLjK15jguLk5FRUXVzjscDsXFxZnp+qWkp6dr+vTp+vDDDxUbG6tx48bp999/f+7V61mzZik2Nlb9+vVTWFiYjh8/Xus5AQAAUD9YDMMw3H1zo0aNlJ+f79oVolJBQYEiIyPrxI1APMXhcMhms6m4uFihoaG+jgMAAICneKJec2tZxfnz513PL126pPz8fNdxWVmZ9u/fr8jISLcCAQAAAL7iVnHcp08f1x3malo+ERQUpI0bN5oOBwAAAHiTW8VxTk6ODMNQp06d9McffygsLMzVVnnzjae3VwMAAADqOreK444dO0p6vNcwAAAA4C9MbeWWkZHx3Pbp06eb6R4AAADwKlO7VbRs2bLK8aNHj/TgwQMFBgaqWbNmKiwsNB2wrmC3CgAAgLrNE/WaqX2O7969W+VRUlKiy5cva+jQodq5c6eZrgEAAACvM1Uc16Rr1676/PPPNX/+fE93DQAAANQqjxfHktS4cWPdvHmzNroGAAAAao2pL+Tt27evyrFhGMrLy1NaWpqGDBliKhgAAADgbaaK4/Hjx1c5tlgsCgsLU1xcnNatW2emawAAAMDrTBXH7HMMAAAAf2KqOK50+/ZtWSwWtW7d2hPd1UmVO945HA4fJwEAAEBNKus0EzsVu18cFxUVafHixfrmm2909+5dSY/3PZ4yZYpWrlypFi1auB2qLrp3754kqUOHDj5OAgAAgOe5d++ebDabW+916yYghYWFGjRokG7cuKFp06bptddek2EYys7O1o4dO9ShQwedOHGi2k1C6rPy8nLdvHlTISEhslgsvo6DGjgcDnXo0EG5ubncqKUBYvwbNsa/YWP8G7Ynxz8kJET37t1TRESEGjVyb1M2t64cr1ixQoGBgbp27ZratWtXrS0+Pl4rVqzQ+vXr3QpVFzVq1Ejt27f3dQy8hNDQUD4cGzDGv2Fj/Bs2xr9hqxx/d68YV3KrpM7MzNTatWurFcaSZLfblZqaqj179pgKBgAAAHibW8VxXl6eevTo8cz2nj17Kj8/3+1QAAAAgC+4VRy3adNG169ff2Z7Tk6OX+9cgbrJarVq2bJlslqtvo4CH2D8GzbGv2Fj/Bs2T4+/W1/Imzlzpq5evaqDBw8qMDCwSpvT6dTo0aPVuXNnbdmyxSMhAQAAAG9wqzj+559/1K9fP1mtVs2ZM0fdu3eXJF26dElfffWVnE6nsrKy2PYMAAAA9YpbxbH0eOnEe++9p59//tm10bLFYtGoUaOUlpamLl26eDQoAAAAUNvcLo4r3b17V1euXJEkdenSRa1atfJIMAAAAMDbTBfHAAAAgL9w79YhQB2xfPlyWSyWKg+73e7rWKhFv/zyi8aOHauIiAhZLBZlZmZWaTcMQ8uXL1dERISCgoI0YsQIXbx40Tdh4XEvGv+UlJRqnwkDBw70TVh41Jo1a9S/f3+FhISobdu2Gj9+vC5fvlzlNcx///Uy4++p+U9xjHqvR48eysvLcz0uXLjg60ioRffv31fv3r2VlpZWY3tqaqq+/PJLpaWl6dSpU7Lb7Ro1apTu3bvn5aSoDS8af0lKSEio8pnw008/eTEhasuxY8c0Z84cnTx5UgcPHlRpaani4+N1//5912uY//7rZcZf8sz8d+v20UBd0qRJE64WNyCJiYlKTEyssc0wDG3YsEGLFy/WxIkTJUlbt25Vu3bttGPHDs2ePdubUVELnjf+laxWK58Jfmj//v1VjtPT09W2bVudPn1aw4cPZ/77uReNfyVPzH+uHKPeu3LliiIiIhQTE6MpU6bor7/+8nUk+EhOTo7y8/MVHx/vOme1WvXmm2/qxIkTPkwGbzp69Kjatm2rbt26adasWSooKPB1JNSC4uJiSXJtBMD8b1ieHv9Knpj/FMeo1wYMGKCMjAwdOHBAX3/9tfLz8zV48GDduXPH19HgA5W3rW/Xrl2V8+3ateOW9g1EYmKitm/frsOHD2vdunU6deqU4uLi5HQ6fR0NHmQYhhYsWKChQ4eqZ8+ekpj/DUlN4y95bv6zrAL12pP/Xu3Vq5cGDRqkzp07a+vWrVqwYIEPk8GXLBZLlWPDMKqdg3+aPHmy63nPnj3Vr18/dezYUT/++KPrX+2o/+bOnavz58/rt99+q9bG/Pd/zxp/T81/rhzDrwQHB6tXr16uvbfRsFSuM3v6KlFBQUG1q0loGMLDw9WxY0c+E/zIvHnztG/fPh05ckTt27d3nWf+NwzPGv+auDv/KY7hV5xOp7KzsxUeHu7rKPCBmJgY2e12HTx40HXu4cOHOnbsmAYPHuzDZPCVO3fuKDc3l88EP2AYhubOnavdu3fr8OHDiomJqdLO/PdvLxr/mrg7/1lWgXpt4cKFGjt2rKKiolRQUKCVK1fK4XAoOTnZ19FQS0pKSnT16lXXcU5Ojs6ePatWrVopKipKH3zwgVavXq2uXbuqa9euWr16tZo1a6apU6f6MDU85Xnj36pVKy1fvlxJSUkKDw/X9evX9cknn6hNmzaaMGGCD1PDE+bMmaMdO3Zo7969CgkJcV0httlsCgoKksViYf77sReNf0lJiefmvwHUY5MnTzbCw8ONgIAAIyIiwpg4caJx8eJFX8dCLTpy5IghqdojOTnZMAzDKC8vN5YtW2bY7XbDarUaw4cPNy5cuODb0PCY543/gwcPjPj4eCMsLMwICAgwoqKijOTkZOPvv//2dWx4QE3jLslIT093vYb5779eNP6enP/cPhoAAACowJpjAAAAoALFMQAAAFCB4hgAAACoQHEMAAAAVKA4BgAAACpQHAMAAAAVKI4BAACAChTHAAAAQAWKYwAAAKACxTEAAABQgeIYAAAAqEBxDAAAAFT4P+SGudPq5CpXAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 800x700 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "num_steps = 25\n",
    "\n",
    "for samples, labels in test_loader:\n",
    "    print(len(samples[2]))\n",
    "    output = net(samples[2])\n",
    "    print(len(net.spkIn))\n",
    "    print(len(output))\n",
    "    print(labels[2])\n",
    "    #spk_In = torch.stack(net.spkIn)\n",
    "    spk_rec1 = torch.stack(net.spk1)\n",
    "    spk_rec2 = torch.stack(net.spk2)\n",
    "    plot_snn_spikes_modified(net.spkIn[0], spk_rec1, spk_rec2, num_steps, \"Fully Connected Spiking Neural Network\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b0e7839e-d98b-469c-96ca-3eea48f0ef2d",
   "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": "3.8.18"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}